Structural and Thermal Performance of Cold-formed Steel ...eprints.qut.edu.au/46654/1/Prakash_Kolarkar_Thesis.pdf · Structural and Thermal Performance of Cold ... Commonly Used Cold-formed

School of Urban Development

Queensland University of Technology

Structural and Thermal Performance of Cold-formed Steel

Stud Wall Systems under Fire Conditions

By

Prakash Nagaraj Kolarkar

BE (Civil) (Govt. College of Engineering Pune, India)

ME (Structures) (Govt. College of Engineering Pune, India)

A Thesis Submitted to the School of Urban Development, Queensland University of Technology in Partial Fulfillment of the Requirements for

the Degree of DOCTOR of PHILOSOPHY

September 2010

ABSTRACT

Cold-formed steel stud walls are a major component of Light Steel Framing (LSF)

building systems used in commercial, industrial and residential buildings. In the

conventional LSF stud wall systems, thin steel studs are protected from fire by placing

one or two layers of plasterboard on both sides with or without cavity insulation.

However, there is very limited data about the structural and thermal performance of

stud wall systems while past research showed contradicting results, for example,

about the benefits of cavity insulation. This research was therefore conducted to

improve the knowledge and understanding of the structural and thermal performance

of cold-formed steel stud wall systems (both load bearing and non-load bearing) under

fire conditions and to develop new improved stud wall systems including reliable and

simple methods to predict their fire resistance rating.

Full scale fire tests of cold-formed steel stud wall systems formed the basis of this

research. This research proposed an innovative LSF stud wall system in which a

composite panel made of two plasterboards with insulation between them was used to

improve the fire rating. Hence fire tests included both conventional steel stud walls

with and without the use of cavity insulation and the new composite panel system.

A propane fired gas furnace was specially designed and constructed first. The furnace

was designed to deliver heat in accordance with the standard time temperature curve

as proposed by AS 1530.4 (SA, 2005). A compression loading frame capable of

loading the individual studs of a full scale steel stud wall system was also designed

and built for the load-bearing tests. Fire tests included comprehensive time-

temperature measurements across the thickness and along the length of all the

specimens using K type thermocouples. They also included the measurements of load-

deformation characteristics of stud walls until failure.

The first phase of fire tests included 15 small scale fire tests of gypsum plasterboards,

and composite panels using different types of insulating material of varying thickness

and density. Fire performance of single and multiple layers of gypsum plasterboards

was assessed including the effect of interfaces between adjacent plasterboards on the

thermal performance. Effects of insulations such as glass fibre, rock fibre and

cellulose fibre were also determined while the tests provided important data relating

to the temperature at which the fall off of external plasterboards occurred.

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions i

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions ii

In the second phase, nine small scale non-load bearing wall specimens were tested to

investigate the thermal performance of conventional and innovative steel stud wall

systems. Effects of single and multiple layers of plasterboards with and without

vertical joints were investigated. The new composite panels were seen to offer greater

thermal protection to the studs in comparison to the conventional panels.

In the third phase of fire tests, nine full scale load bearing wall specimens were tested

to study the thermal and structural performance of the load bearing wall assemblies. A

full scale test was also conducted at ambient temperature. These tests showed that the

use of cavity insulation led to inferior fire performance of walls, and provided good

explanations and supporting research data to overcome the incorrect industry

assumptions about cavity insulation. They demonstrated that the use of insulation

externally in a composite panel enhanced the thermal and structural performance of

stud walls and increased their fire resistance rating significantly. Hence this research

recommends the use of the new composite panel system for cold-formed LSF walls.

This research also included steady state tensile tests at ambient and elevated

temperatures to address the lack of reliable mechanical properties for high grade cold-

formed steels at elevated temperatures. Suitable predictive equations were developed

for calculating the yield strength and elastic modulus at elevated temperatures.

In summary, this research has developed comprehensive experimental thermal and

structural performance data for both the conventional and the proposed non-load

bearing and load bearing stud wall systems under fire conditions. Idealized hot flange

temperature profiles have been developed for non-insulated, cavity insulated and

externally insulated load bearing wall models along with suitable equations for

predicting their failure times. A graphical method has also been proposed to predict

the failure times (fire rating) of non-load bearing and load bearing walls under

different load ratios. The results from this research are useful to both fire researchers

and engineers working in this field. Most importantly, this research has significantly

improved the knowledge and understanding of cold-formed LSF walls under fire

conditions, and developed an innovative LSF wall system with increased fire rating. It

has clearly demonstrated the detrimental effects of using cavity insulation, and has

paved the way for Australian building industries to develop new wall panels with

increased fire rating for commercial applications worldwide.

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions iii

TABLE OF CONTENTS

Abstract i

Table of Contents iii

List of Figures vi

List of Tables xx

Statement of Original Authorship xxii

Acknowledgements xxiii

Chapter 1.0: Introduction 01-13

1.1: Cold-formed Steel Members 01

1.2: Need for Fire Resistant Structures 04

1.3: Fire Resistance of LSF Stud Wall Systems 06

1.4: Problem Definition 08

1.5: Aims of this Research 09

1.6: Research Method 11

1.7: Contents of Thesis 13

Chapter 2.0: Literature Review 14-51

2.1: Experimental Research 14

2.2: Analytical Research 30

2.3: Mechanical and Thermo Physical Properties of Steel Stud Wall Assembly Components at Elevated Temperatures

39

2.4: Literature Review Findings Relevant to this Research 49

Chapter 3.0: Experimental Work to Determine the Mechanical Properties of G500 Cold-Formed Steel at Elevated Temperatures.

52-74

3.1: Introduction 52

3.2: Experimental Investigation 54

3.3: Comparison of Reduction Factors with Results as Obtained by Other Researchers and as Recommended by Steel Design Codes

66

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions iv

3.4: Conclusion 74

Chapter 4.0: Thermal Performance of Gypsum Plasterboards and Composite Panels

75-130


4.2: Test Setup and Procedure 76

4.3: Test Specimens 78

4.4: Conclusion 128

Chapter 5.0: Thermal Performance of Non-Load Bearing Wall Systems 131-188



5.3: Construction Details of Test Specimens 134

5.4 Test Set-up and Procedure 142

5.5 Observations, Results and Discussion 144

Chapter 6.0: Structural and Thermal Performance of Load Bearing Wall Systems

189-332



6.3: Construction Details of Test Specimens 195

6.4: Test Set-up and Procedure 205

6.5: Observations and Results 218

6.6: Discussions

332

Chapter 7: Discussions and Recommendations 333-376

7.1: Discussions 333

7.2: Simplified Method for the Determination of Failure Times of Wall Specimens 357

7.3 Essential Points to Consider for Thermal Modeling 371

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions v

7.4: Conclusion: 376

Chapter 8: Summary 377-381

8.1: Main Research Outcomes 379

8.2: Recommendations to the Construction Industry 380

8.3: Future Research 381

References 382-388

List of Figures Page No.

Chapter 1 1-13

Figure 1.1: Commonly Used Cold-formed Steel Structural Shapes 01

Figure 1.2: Strength of Steel at Elevated Temperature Relative to Yield Strength at Ambient Temperature

02

Figure 1.3: Applications of Cold-formed Steel Products 03

Figure 1.4: Transportable Houses 03

Figure 1.5: House Frames 03

Figure 1.6: Steel Stud Wall System 04

Figure 1.7: Wall Panel Showing Steel Channels Sections and Plasterboards 05

Figure 1.8: Fire Ratings of Some Exterior Wall Systems of Boral 07

Figure 1.9: New LSF Stud Wall System using a Composite Panel 10

Chapter 2 14-51

Figure 2.1: Construction of Assemblies 18

Figure 2.2: Typical Steel Frame Fabrication Layout for Wall Specimens 22

Figure 2.3: Location of Temperature Measurements and Simulation Boundaries

22

Figure 2.4: Structural Failure Modes 27

Figure 2.5: Total Horizontal Deflection for Load-bearing Systems 31

Figure 2.6: Thermal Bowing and Secondary Deflection 34

Figure 2.7: Stud End Conditions 34

Figure 2.8: Gypsum Plasterboard 39

Figure 2.9: Thermal Conductivity of Gypsum Plasterboard 43

Figure 2.10: Specific Volumetric Enthalpy of Gypsum Plasterboard 43

Figure 2.11: Specific Heat of Type X Gypsum Board 44

Figure 2.12: Thermal Conductivity of Type X Gypsum Board 45

Figure 2.13: Density Variation of Type X Gypsum Plasterboard on Heating 45

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions vi

Figure 2.14: Mass Loss in Gypsum Plasterboard on Heating 47

Chapter 3 52-74

Figure 3.1: Tensile Test Specimen 54

Figure 3.2: Furnace Details 56

Figure 3.3: Details of Test Rig and its Components 57-58

Figure 3.4 EDCAR (Experimental Data Collection and Recorder) 59

Figure 3.5 Strain Measurement using LSE 60

Figure 3.6: Typical Speckle Output for Strain Measurements 61

Figure 3.7: Comparison of Stress-Strain Curves using Strain Gauges and Laser Speckle Extensometer

62

Figure 3.8: Determination of (a) Yield strength and (b) Elastic modulus. 63

Figure 3.9: Stress-Strain Graphs at Different Temperatures 65

Figure 3.10: Graph Showing Strength Reduction Factors associated with Various Percentages of Yield Strength as Obtained from Tests

66

Figure 3.11: Yield Strength Reduction Factors 67

Figure 3.12: Modulus of Elasticity Reduction Factors 67

Figure 3.13: Variation of Yield Strength Reduction Factors with Temperature

68-69

Figure 3.14: Comparison of 0.2% Strength Reduction Factors with AS 4100 (SA, 1998) Recommendations

70

Figure 3.15: Comparison of Modulus of Elasticity Reduction Factors with AS 4100 (SA, 1998) Recommendations

70

Figure 3.16: Comparison of Yield Strength Reduction Factors with Test Results and Predictive Equation

72

Figure 3.17: Comparison of Elastic Modulus Reduction Factors with Test Results and Predictive Equation

72

Figure 3.18: Comparison between Ranawaka (2009) Equation and Predictive Equation in the determination of Yield Strength Reduction Factors

73

Figure 3.19: Comparison between Ranawaka (2009) Equation and Predictive Equation in the determination of Elastic Modulus Reduction Factors

73

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions vii

Chapter 4 75-130

Figure 4-1: View Showing Adapter Attached to Large Furnace for Carrying Out Fire Testing of Small Scale Specimens

77

Figure 4-2: Adapter Details 77

Figure 4-3: View Showing Plasterboard Specimen Installed for Fire Testing 78

Figure 4-4: Pressure Transducer used for Determining Furnace Chamber Pressure during Testing

78

Figure 4-5: Thermocouples on the Ambient Side of the Specimen 80

Figure 4-6: Instrumentation for Test Specimen 1 81

Figure 4-7: Fire Testing of Test Specimen 1 83

Figure 4-8: Time-Temperature Profile of Test Specimen 1 84

Figure 4-9: Temperature-Depth Profiles of Test Specimen 1 84
















P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions viii

Figure 4-25: Construction of Test Specimen 6 99

Figure 4-26: Instrumentation for Test Specimens from 6 to 15 99







Figure 4-33: Test Specimen 7 Installed in the Furnace for Testing 104








Figure 4- 41: Temperature-Depth Profiles of Test Specimen 9 108







Figure 4-48: Construction of Test Specimens 12, 13 and 14 116

Figure 4-49: Fire Testing of Test Specimen 11 119-120



P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions ix









Figure 4-60: Time-Temperature profiles for interface Ins-Pb2 of Test Specimens 6 to 15

128

Figure 4-61: Average Time-Temperature profile for interface Ins-Pb2 of Test Specimens 6 to 9 compared with Time-Temperature profile of Pb1-Pb2 interface temperature of Test Specimen 4

130

Chapter 5 131-188

Figure 5-1: Typical steel wall frame used to build NLB test wall specimens 134


Figure 5-3: Thermocouple Locations for Test Specimen 1 135

Figure 5-4: Test Specimen 2 with a Joint in the Exposed Plasterboard over the Central Stud and Thermocouple Locations

136

Figure 5-5: Thermocouple Locations for Test Specimen 3 136

Figure 5-6: Construction and Placement of Test Specimen 4 in the Furnace 137

Figure 5-7: Thermocouple Locations for Test Specimens 4, 5 and 6 138



Figure 5-10: Thermocouple Locations for Test Specimens 7, 8 and 9 140


Figure 5-12: Test Specimen placed in the specially built adapter in the large furnace

142

Figure 5-13: Test Specimen subjected to fire on one side 143

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions x

Figure 5-14: Test Specimen 1 after the fire test 145

Figure 5-15: Test Specimen 2 after the fire test 145

Figure 5-16: Time-Temperature Profile for Test Specimen 1 (No joints in plasterboard)

146

Figure 5-17: Time-Temperature Profile for Test Specimen 2 (With a joint in the exposed plasterboard over the central stud)

146

Figure 5-18: Time –Temperature Profiles of the Flanges in Stud No.1 of Test Specimens 1 and 2

148


148


149

Figure 5-21: Average Unexposed Surface Temperature of Test Specimens 1 and 2

150

Figure 5-22: Time-Temperature Profiles of Cavity facing surfaces of Specimens 1 and 2

151

Figure 5-23: Lateral Deflections of the Central Studs in Test Specimens 1 and 2

152

Figure 5-24: Test Specimen 3 after the fire test (no cavity insulation) 154

Figure 5-25: Test Specimen 4 after the fire test (glass fibre cavity insulation) 154

Figure 5-26: Test Specimen 5 after the fire test (rock fibre as cavity insulation)

155

Figure 5-27: Test Specimen 6 after the fire test (cellulose as cavity insulation)

156

Figure 5-28: Time-Temperature Profiles of Plasterboard surfaces in Test Specimen 3 (No Cavity Insulation)

159

Figure 5-29: Time-Temperature Profiles of Plasterboard surfaces in Test Specimen 4 (Cavity Insulation – Glass Fibre)

159

Figure 5-30: Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 5 (Cavity Insulation-Rock Fibre)

160

Figure 5-31: Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 6 (Cavity Insulation – Cellulose Fibre)

160

Figure 5-32: Time-Temperature Profiles across Studs in Test Specimen 3 (No Cavity Insulation)

164

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xi

Figure 5-33: Time-Temperature Profiles across Studs in Test Specimen 4 (Cavity Insulation – Glass Fibre)

164

Figure 5-34: Time-Temperature Profiles across Studs in Test Specimen 5 (Cavity Insulation-Rock Fibre)

165

Figure 5-35: Time-Temperature Profiles across Studs in Test Specimen 6 (Cavity Insulation – Cellulose Fibre)

165

Figure 5-36: Time-Temperature Profiles across the Cross-section of Test Specimen 3 (No Cavity Insulation)

167

Figure 5-37: Time-Temperature Profiles across the Cross-section of Test Specimen 4 (Cavity Insulation – Glass Fibre)

167

Figure 5-38: Time-Temperature Profiles across the Cross-section of Test Specimen 5 (Cavity Insulation – Rock Fibre)

168

Figure 5-39: Time-Temperature Profiles across the Cross-section of Test Specimen 6 (Cavity Insulation – Cellulose Fibre)

168

Figure 5-40: Deflection-Time Profiles of Test Specimen 3 (No Cavity Insulation)

170

Figure 5-41: Deflection-Time Profiles of Test Specimen 4 (Cavity Insulation – Glass Fibre)

170

Figure 5-42: Deflection-Time Profiles of Test Specimen 5 (Cavity Insulation – Rock Fibre)

171

Figure 5-43: Deflection-Time Profiles of Test Specimen 6 (Cavity Insulation – Cellulose Fibre)

171

Figure 5-44: Test Specimen 7 after the fire test (Glass fibre as external insulation)

174

Figure 5-45: Test Specimen 8 after the fire test (Rock fibre as external insulation)

174

Figure 5-46: Test Specimen 9 after the fire test (Cellulose fibre as external insulation)

175

Figure 5-47: Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 7 (External Insulation-Glass Fibre)

176

Figure 5-48: Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 8 (External Insulation-Rock Wool)

177

Figure 5-49: Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 9 (External Insulation-Cellulose Fibre)

177

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xii

Figure 5-50: Time-Temperature Profiles across Studs in Test Specimen 7 (External Insulation-Glass Fibre)

181

Figure 5-51: Time-Temperature Profiles across Studs in Test Specimen 8 (External Insulation-Rock Fibre)

182

Figure 5-52: Time-Temperature Profiles across Studs in Test Specimen 9 (External Insulation-Cellulose Fibre)

182

Figure 5-53: Time-Temperature Profiles over the Entire Cross-section of Test Specimen 7 (External Insulation-Glass Fibre)

184

Figure 5-54: Time-Temperature Profiles over the Entire Cross-section of Test Specimen 8 (External Insulation-Rock Fibre)

184

Figure 5-55: Time-Temperature Profiles over the Entire Cross-section of Test Specimen 9 (External Insulation-Cellulose Fibre)

185

Figure 5-56: Lateral Deflection -Time Profiles of Test Specimen 7

(External Insulation-Glass Fibre)

186

Figure 5-57: Lateral Deflection -Time Profiles of Test Specimen 8 (External Insulation-Rock Fibre)

186

Figure 5-58: Lateral Deflection -Time Profiles of Test Specimen 9 (External Insulation-Cellulose Fibre)

187

Chapter 6 189-332

Figure 6-1 (a): Basic Local Failure Modes 189

Figure 6-1 (b): Basic Global Failure Modes 190

Figure 6-2: Test Wall Frame 191

Figure 6-3: Stud to Plasterboard Connections 192

Figure 6-4: Protection of Joints 193

Figure 6-5: Stud to Track Connection at the Top 195


Figure 6-7: Fixing of Face Plasterboard on the Ambient Side of Test Specimen 3

197

Figure 6-8: Construction of Test Specimen 4 Using Glass Fibre as Cavity Insulation

198

Figure 6-9: Construction of Test Specimen 5 using Rock Fibre as Cavity Insulation

199-200

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xiii

Figure 6-10: Construction of Test Specimen 6 using Cellulose Fibres as Cavity Insulation

201

Figure 6-11: Construction of Test Specimen 7 using Glass Fibres as External Insulation

202

Figure 6-12: Construction of Test Specimen 8 using Rock Fibres as External Insulation

203

Figure 6-13: Construction of Test Specimen 9 using Cellulose Fibres as External Insulation

204

Figures 6-14: Details of Furnace Operation and Components 206-207

Figure 6-15: Loading Frame 209

Figure 6-16: Loading Arrangement 209-210

Figure 6-17: Hydraulic Pump and its Connections 211

Figure 6-18: Test Set-up for Ambient Temperature Test 212

Figure 6-19: LVDTs Used in the Measurement of Axial Shortening and Out-of-plane Deflection of Test Specimen Wall

213

Figure 6-20: Thermocouple Locations for Load Bearing Wall Specimens 215

Figure 6-21: Infrared Gun Used for the Measurement of Ambient Side Temperatures

217

Figure 6-22: Test Specimen Complete with all its Instrumentation Ready for Fire Test

217

Figure 6-23: Failure of Test Specimen 1 218

Figure 6-24: Load Vs Axial Deformation - Profiles of Test Specimen 1 at Ambient Temperature

219

Figure 6-25: Fire Performance Test of Specimen 2 220

Figure 6-26: Detachment and Opening of Plasterboard joints Caused by Shrinkage

221

Figure 6-27: Test Specimen 2 after Removing the Exposed Plasterboard Layer

222

Figure 6-28: Stud Failure Initiated by Plasterboard Fall-off 223

Figure 6-29: Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 2

225

Figure 6-30: Time-Temperature Profiles across Studs 1 to 4 of Test Specimen 2 226-227

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xiv

Figure 6-31: Axial Deformation Plots for Studs of Test Specimen 2 228-229

Figure 6-32: Lateral Deflection -Time Profiles of Test Specimen 2 at Mid-Height

229

Figure 6-33: Close up of Test Specimen 3 after Removing the Exposed Plasterboards

234-235

Figure 6-34: Studs of Test Specimen 3 after the Fire Test 235

Figure 6-35: Time-Temperature Plots of Plasterboard Surfaces in Test Specimen 3

237-238

Figure 6-36: Time-Temperature Plots of Flange and Web Surfaces of Central Studs in Test Specimen 3

239-240

Figure 6-37: Time-Temperature Profiles across Central Studs at Mid-height in Test Specimen 3

241


Figure 6-39: Lateral Deflection-Time Plots of Test Specimen 3 243-244

Figure 6-40: Axial Load -Time Profile of Test Specimen 3 during Fire Test 245

Figure 6-41: Test Specimen 4 after the Fire Test 248-250


252-253


254-255

Figure 6-44: Time-Temperature Profiles across Central Studs at Mid-Height in Test Specimen 4

255

Figure 6-45: Outward Lateral Deflection of Test Specimen 4 at Failure 257

Figure 6-46: Axial Deformation Plots for Studs of Test Specimen 4 258

6-47: Lateral Deflection-Time Plots of Test Specimen 4 259-260




267-268

Figure 6-51: Time-Temperature Plots of Flanges and Web Surfaces of Central Studs in Test Specimen 5

269

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xv


270

Figure 6-53: Specimen Behaviour during the Test 272-273

Figure 6-54: Ambient Side of Wall after Fire Test 273-274


Figure 6-56: Lateral Deflection-Time Plots of Test Specimen 5 275




283


284-285


286

Figure 6-62: Inward Thermal Bowing of Test Specimen in the Initial Period of the Test

287






297


298-299


299





P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xvi


310


311-312


313

Figure 6-77: View of Loading Arrangement 314-315






324-325


326-327


327




Chapter 7 333-376

Figure 7-1: Time-temperature Profiles for the Central Stud in Non-Load Bearing Wall Test Specimens 4 and 7 (Glass fibre insulation)

334

Figure 7-2: Time-temperature Profiles for the Central Stud in Non-load Bearing Wall Test Specimens 5 and 8 (Rock fibre insulation)

335

Figure 7-3: Time-temperature Profiles for the Central Stud in Non-Load Bearing Wall Test Specimens 6 and 9 (Cellulose fibre insulation)

335

Figure 7-4: Time-temperature Profiles for the Central Stud Hot Flanges in Non-Load Bearing Wall Test Specimens 4 to 9

336

Figure 7-5: Hot Flange Temperatures of the Central Stud in NLB Wall Test Specimens 3 to 9

337

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xvii

Figure 7-6: Temperature Difference across the Central Studs C/S in NLB Wall Test Specimens 3 to 9

338-339

Figure 7-7: Ambient Side Temperature of External Plasterboard 2 in Test Specimens 3 to 9

339-340

Figure 7-8: Temperature on the Ambient Face of the Non-load Bearing Wall Test Specimens 3 to 9

340-341

Figure 7-9: Fall off Times of Plasterboard 2 in Non-Load Bearing Wall Test Specimens 4 to 9

341

Figure 7-10: Ambient Side Time-temperature Profiles of Test Specimens 4 to 9

342

Figure 7-11: Average Time-temperature Profiles for the Central Studs in Load Bearing Wall Test Specimens 5 and 8

343


344

Figure 7- 13: Average Hot Flange Temperatures of the Central Studs of Load Bearing Wall Test Specimens 3 to 9

345

Figure 7-14 : Average Temperature Difference across the Central Studs for Load Bearing Wall Test Specimens 3 to 9

346

Figure 7-15: Average Temperature Difference across Central Studs and their Lateral Deformations versus Time for Test Specimens 5 and 8

347


347

Figure 7- 17:Average Lateral Deformations of the Central Studs in Load Bearing Wall Test Specimens

348

Figure 7-18: Average Time-temperature Profiles of Hot Flanges for the Central Studs in Test Specimens 4 to 9

349

Figure 7-19: Average Time-temperature Profiles of Cold Flanges for the Central Studs in Test Specimens 4 to 9

350

Figure 7-20: Temperature Difference across the Central Studs in Cavity Insulated and Externally Insulated Specimens

350

Figure 7-21: Average Time-temperature Profiles of Pb2-Cav Surface in Test Specimens 3 to 9

352

Figure 7- 22: Average Time-Temperature Profiles on the Ambient Side of the Exposed Base Layer Plasterboard 2 in LBW Test Specimens 3 to 9

353

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xviii

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xix

Figure 7- 23: Ambient Side Temperatures of LBW Test Specimens 3 - 9 354

Figure 7-24: Variation of Yield Strength Reduction Factor of 1.15 mm G500 Steel with respect to Temperature

357

Figure 7-25: Idealized Hot Flange Temperatures of Load Bearing Test Specimens 2 to 9

360

Figure 7-26: Development of Hot Flange Failure Times for a Given Load Ratio

360

Figure 7-27: Determination of Hot Flange Failure Times using Load Ratio 362

Figure 7-28: Load Ratio Vs Critical Hot Flange Temperatures at Stud Failure 365

Figure 7-29: Load Ratio Vs Stud Failure Times for Test Specimen 2 366







Figure 7-36: Load Ratio Vs Stud Failure Times for all Test Specimens Using Predictive Equations

369

List of Tables

Chapter 2 14-51

Table 2.1: Fire Resistance of Typical Floors, Walls and Partitions (From SCI, 1993)

16

Table 2.2: Small Scale Assembly Parameters and Fire Test Results (Sultan & Lougheed, 1994)

17

Table 2.3: Full Scale Fire Test Specimens (Used by Gerlich, 1995) 19

Table 2.4: Summary of Fire Resistance Tests on Load-Bearing LSF Walls by Kodur et al. (1999)

25

Table 2.5 Summary of Fire Resistance Tests on Load-Bearing LSF Walls by Alfawakhiri (2001)

26

Table 2.6: Mechanical Properties of Australian Manufactured Plasterboards Goncalves et al., (1996)

42

Chapter 3 52-74

Table 3.1: Mechanical Properties of 1.15 mm G500 CFS at Ambient and Elevated Temperatures.

64

Table 3.2: Reduction Factors for Yield Strength and Modulus of Elasticity of 1.15 mm G500 Steel

65

Chapter 4 75-130

Table 4-1: Details of Plasterboard and Composite Panel Test Specimens 79

Table 4-2: Time–Temperature profile of the ambient side of the insulation (Ins-Pb2 interface) in Test Specimens 6, 7, 8 and 9 using glass fibre as insulation material

109

Table 4-3: Time–Temperature Profile of the Ambient Side of the Insulation (Ins-Pb2 interface) in Test Specimens 10 and 11 using Rock Fibre as Insulation Material

115

Table 4-4: Time–Temperature profile of the ambient side of the insulation (Ins-Pb2 interface) in Test Specimens 12, 13 and 14 using cellulose fibre as insulation material

124

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xx

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xxi

Chapter 5 131-188

Table 5-1: Details of non-load bearing wall specimens 133

Table 5-2: Central Stud Temperatures of Test Specimens 1 and 2 150

Table 5.3: Hot Flange Temperature versus Time for the Central Stud 166

Table 5-4: Failure Times of Wall Components in minutes 172

Table 5-5: Hot Flange Temperature versus Time for the Central Stud 183

Table 5-6: Failure times of Wall Components in Minutes 187

Chapter 6 189-332

Table 6-1: Details of Test Specimen Configuration 194

Chapter 7 333-376

Table 7-1: Failure Times of Test Specimens 355

Table 7-2: Stud Reversal Times for Cavity Insulated and Externally Insulated Specimens along with the Corresponding Temperatures

356

Table 7-3: Comparison of Predicted Hot Flange (HF) Failure Times of Load Bearing Wall Specimens with Actual Local Buckling of HF (minutes) at a Load Ratio of 0.2

361

Table 7-4: Comparison of Predicted Failure Times of Non-load Bearing Wall Specimens with their Actual Failure Times.

363

Table 7-5: Predictive Equations for Obtaining Stud Failure Times for Different Wall Models

370

Table 7-6: Table Comparing the Actual Stud Reversal Times and Wall Failure Times with the Predicted Values

371

Statement of Original Authorship

The work contained in this thesis has not been previously submitted for a degree or

diploma at any other higher education institution. To the best of my knowledge and

belief, the thesis contains no material previously published or written by another person

except where due reference is made.

Prakash Nagaraj Kolarkar

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xxii

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions xxiii

Acknowledgements

The research described in this report was carried out in the School of Urban

Development, Queensland University of Technology, Australia.

I would like to thank my supervisor Prof. Mahen Mahendran of the Queensland

University of Technology for his inspiration, guidance and enthusiasm. Thanks also to

my fellow researchers Banduka Heva and Gunalan for their help during my experimental

work and to the technical staff in the laboratory, Arthur, Brian and Jim Hazelman for

their help in the testing and data acquisition work.

I would also like to thank my parents, especially my mother who has been a constant

source of encouragement, my wife and son who exhibited immense patience and

willingness to work around my schedule, and my sister who was happy to manage all my

commitments back in my home country. Their support has been of immense help in the

completion of this thesis.

Chapter 1: Introduction

1.1: Cold-formed Steel Members

Steel members are widely used in buildings due to their advantages of high strength,

good ductility and fast fabrication and erection. Two types of structural steel are used

in the building industry, i.e. hot-rolled and cold-formed steels. In cold-formed steel

products, the strength comes from the material and how it is shaped. The load bearing

capacity of a thin flat sheet of steel can be greatly increased if it is formed into an

efficient multi-sided cross-section. The strength to weight ratio of cold-formed steel

products is very favourable in comparison to the thicker hot-rolled steel products.

Figure 1.1 shows some of the commonly used cold-formed steel structural shapes.

The depth of the members generally ranges from 50 mm to 300 mm while their

thicknesses are in the range of 0.75 to 3 mm.

Plain C-section Lipped C-section Swage Beam Hat Section

Z – Section Back to Back C-Sections Box C-Section Figure 1.1: Commonly Used Cold-formed Steel Structural Shapes

There are many differences between hot-rolled and cold-formed steel sections.

1) Many complex structural shapes are possible with cold-formed steel. There is a

gain in material strength and hardness due to cold working effects.

P.N.Kolarkar: Structural and Thermal Performance of Cold-formed Steel Stud Wall Systems under Fire Conditions 1

2) The high strength to weight ratio of cold-formed steel makes it much easier and

economical to mass produce, transport and install cold-formed products.

3) The thicker hot-rolled sections generally prevent the occurrence of local

buckling before yielding. In contrast, local buckling may become a concern for

cold-formed steel sections as it can occur at stresses well below the yield point.

4) Cold-formed steel loses strength more rapidly than hot-rolled steel when

exposed to increasing temperatures. According to Sidey and Teague (1988) hot-

rolled steel retains its full strength up to 4000C, beyond which the strength

quickly decreases. The loss of strength in the case of cold-formed steel is 10 –

20% more than that of hot-rolled steel as shown in Figure 1.2

Figure 1.2: Strength of Steel at Elevated Temperature Relative to Yield

Strength at Ambient Temperature (NAHB Research Centre, 2002)

5) Cold-formed steel exhibits superior corrosion resistance than hot-rolled steel due

to the improved galvanising and other coating technology. The protective

coating system is not damaged during the cold-forming process (Davies 2000).

6) Cold-formed steel products exhibit more accurate complex shapes of precise

lengths due to the recent progress in rolling and forming technologies.

The main applications of cold-formed steel products have been in elements such as

purlins and sheet rails, cladding and decking, pallet racking and shelving. Their

strength, lightweight, versatility, non-combustibility, ease of prefabrication and

handling has made cold-formed steel members very popular in the building industry.

Trusses, wall Frames, posts and beams made of cold-formed steel as shown in Figure

1.3 are being regularly used. Cold-formed steel is an attractive alternative over other

materials for constructing the entire buildings as shown in Figures 1.4 and 1.5.


a: Trusses b:Wall Frames

c: Posts d: Beams

Fig.1.3: Applications of Cold-formed Steel Products (Steelbuilt Kit Homes, 2005)

Figure 1.4: Transportable Houses (Steelbuilt Kit Homes, 2005)

Figure 1.5: House Frames (Steelbuilt Kit Homes, 2005)


1.2: Need for Fire Resistant Structures

Fire resistant barriers play an important role in maintaining building integrity and

reducing the spread of fire from the room of origin to adjacent compartments. The

traditional method of stud wall construction is with light timber framing and sheet

material linings. However, there has been an increasing demand for prefabricated light

steel frame systems (LSF). Because of its high strength and yet good forming

properties, the material generally used is galvanised mild steel. Steel track and stud

are seen as an environmentally friendly, recyclable alternative to timber stud system.

The replacement of timber with steel becomes more prevalent in regions where timber

resources are scarce and also in commercial or community applications where other

advantages such as speed of assembly and fire retardance are more important.

Figure 1.6: Steel Stud Wall System (Gyprock, 2005)

Partition wall panels composed of a cold-formed steel frame lined with one- or two-

side sheathing (for example plasterboards) have been widely used in building

constructions since 1940s. The panels are typically constructed by first connecting


studs and tracks with rivets to form the frame, and then connecting sheathing boards

to the frame with self-drilling screws (see Figure 1.6). These panels can be easily

assembled to fabricate load-bearing as well as non-load-bearing partition walls.

Plasterboard

Stud

Plasterboard

Figure 1.7: Wall Panel Showing Steel Channel Sections and Plasterboards

Cold-formed thin-walled (CF-TW) steel channels are the predominant sections used

as load bearing wall studs in light-weight steel construction. Under fire conditions,

because of their thinness, CF-TW steel sections heat up quickly resulting in fast

reduction in their strength and stiffness. However, if gypsum boards are combined

with thin-walled steel channels as shown in Figure 1.7 to form steel stud walls, their

fire resistant performance will improve since the gypsum boards can protect the steel

studs from fire exposure, thus delaying temperature rises in the steel studs.

When the walls are used as part of a fire resistant construction, they should satisfy

three fire resistant requirements, namely stability, insulation and integrity.

a) Load-bearing Capacity (Stability)

For load-bearing elements of structure, the test specimen shall not collapse in such a

way that it no longer performs the load-bearing function for which it was constructed.

The purpose of stability requirements in fire is two-fold. Internally, to maintain the

viability of escape routes for a sufficient period to allow safe evacuation and search

and rescue; and externally, to prevent toppling of the building which would endanger

people in the vicinity. The stability requirements to ensure safe egress are independent

of storey level but the external risk from overbalancing of the building depends both

on the height of the structure and the level on which the fire occurs.


b) Insulation

For elements of structure such as walls and floors which have a function of separating

two parts of a building.

a) The average temperature of the unexposed face of the specimen shall not increase

above the initial temperature by more than 140°C.

b) The maximum temperature at any point of this face shall not exceed the initial

temperature by more than 180°C.

c) Integrity

Initial integrity failure shall be deemed to have occurred when a cotton pad is ignited

or when sustained flaming, having duration of at least 10 s, appears on the unexposed

face of the test specimen. It is required to maintain structural integrity during a fire to

avoid structural collapse and to prevent spread of flame and smoke into adjacent

areas. Ultimate integrity failure shall be deemed to have occurred when collapse of

the specimen occurs or at an earlier time based upon integrity and insulation criteria.

When LSF stud wall panels are used as load-bearing walls, sufficient fire resistance is

essential to

prevent or delay the spread of fire within the building or to another building

prevent sudden collapse of building components for the safety of the people and

the fire fighting personnel and assure integrity over a specific interval of time to

facilitate the safe evacuation of the people and allow the fire fighters to operate

safely is a major issue.

The extent to which these walls can withstand fire conditions without losing on

integrity, insulation and stability is known as fire resistance rating.

1.3: Fire Resistance of LSF Stud wall Systems

Non-load bearing LSF stud wall systems have an established history of use, mainly in

light industrial and commercial partitioning. Advantages over timber framing include:

Light-weight nature and dimensional stability of the frame, Speed and ease of frame


erection (often friction fit connections of studs to top and bottom tracks), No lining

delays due to high moisture content of framing, Aesthetic quality of finished wall and

Demountability. These advantages have resulted in a ready acceptance of non-load

bearing LSF stud wall systems as ‘infill’ partitioning in buildings, which have a

conventional structural shell, such as reinforced concrete or masonry construction. In

response to a market demand for fire separations in this area of light industrial and

commercial partitioning, lining manufacturers have developed, tested and published a

range of fire resistance ratings. In Australia the details of tested non-load bearing LSF

stud wall systems are published by Boral and Gyprock (see Figure 1.8). They have

prescribed steel stud walls with plasterboard linings achieving fire resistance ratings

ranging from 60 to 120 minutes. These systems are based on full scale fire resistance

tests against the standard IS0 834 fire curve.

Figure 1.8: Fire Ratings of Some Exterior Wall Systems of Boral (Gyprock, 2005)


Load bearing LSF stud wall systems are less likely to be used as ‘infill’ commercial

partitioning, and will more likely form part of a total LSF construction system. With

the developing use of LSF in load bearing applications the demand for fire resistance

ratings has increased.

1.4: Problem Definition

Cold-formed thin-walled (CF-TW) steel channels are the predominant sections used

as load bearing wall studs in light-weight steel construction. Under fire conditions,

CF-TW steel sections (high section factor) heat up quickly resulting in a rapid

reduction in their strength and stiffness. The use of high strength steels is also

becoming popular in load bearing LSF stud wall construction. The structural

behaviour of high strength steel stud walls is yet to be researched. Also, in Australia

there is no data available on the fire ratings of load-bearing steel stud wall systems.

In parallel with the growing interest in LSF stud wall systems, the understanding and

application of specific Fire Engineering Design is used increasingly for the fire safety

design of buildings. To more accurately apply Fire Engineering Design, a better

understanding of the fire performance of components constituting the LSF stud walls

systems is required. With increasing demand for higher fire ratings of LSF stud wall

systems, the current practice is to prescribe more than two layers of gypsum

plasterboard lining on either side of the cold-formed steel frame making the entire

construction process more labour oriented and expensive. Therefore there is an urgent

need to develop innovative LSF stud wall systems made with improved plasterboard

and insulation systems and verify their improved fire performance.

Several researchers have carried out investigations to determine the impact of

different types of cavity insulations on the thermal performance of stud wall systems.

Sultan and Lougheed (1994) noted that rock and cellulose fibres when used as cavity

insulation improved the fire ratings of the wall systems whereas glass fibres hardly

contributed to any improvement in the thermal performance of the stud wall system.

Feng et al. (2003) reported that the thermal performance of non-load bearing wall

specimens improved with the use of cavity insulations. However, Sultan (1995)

observed that glass fibre cavity insulation has no impact on the thermal performance

of stud walls whereas cellulose fibre cavity insulation actually reduces the fire


resistance. Hence, as very limited data is available about the thermal performance of

non-load bearing and load bearing wall systems and past research has only provided

contradicting results about the benefits of cavity insulation to the fire rating of stud

wall systems, it is necessary to conduct further research by undertaking fire tests on

both non-load bearing and load bearing wall models with and without the use of

cavity insulations to increase the knowledge in this field and provide definitive

methods of improving the fire ratings of the stud wall systems.

The current practice of constructing non-load bearing and load bearing walls is not

very favourable when considering their role as fire resistant barriers. The use of glass

fibres and mineral wool as cavity insulation has only resulted in decreasing the

stability of load-bearing walls due to increased temperature gradients across the wall

and thus promoting larger lateral deflections leading to an early collapse of the wall.

By undertaking a detailed investigation into the structural behaviour of high strength

cold-formed steel studs in load bearing walls under simulated fire conditions and also

studying the fire performance of non-load bearing and load bearing wall panels with

and without insulation using both small scale and large scale fire tests it is proposed

we can fully understand their thermal and structural performances and hence develop

simple design rules that will contribute to the improvement of fire safety design.

1.5: Aims of This Research

Overall aim of this research is to improve the knowledge and understanding of the

structural and thermal performance of both conventional and innovative cold-formed

high strength steel stud wall systems (load bearing and non-load bearing) under

simulated fire conditions and develop reliable and simple methods to predict their fire

resistance rating.

Specific tasks of this research are:

1) To design and build a custom made gas furnace suitable for the standard fire

tests of both small and large scale LSF stud wall systems.


2) To design and build a compression loading frame capable of loading the

individual studs of a large scale LSF stud wall specimen.

3) To conduct both small scale and large scale fire tests to determine the fire

performance and thermal deformations of non-load bearing and load-bearing

stud wall systems using the developed fire test rig.

4) To investigate the structural and thermal performance of currently used cold-

formed high strength steel (LSF) stud wall systems (both load bearing and non-

load bearing) with and without cavity insulations under simulated fire

conditions with temperatures of up to 10000C using the developed fire test rig.

5) To develop new cold-formed steel stud wall systems (both load bearing and

non-load bearing) with improved fire performance based on a new composite

panel in which insulation is located externally between two plasterboards (see

Figure 1.9), and investigate their structural and thermal performance

Composite panel with insulation between two layers of plasterboard

Figure 1.9: New LSF Stud Wall System using a Composite Panel

6) To determine the fire resistance rating (failure times) of load bearing and non-

load bearing cold-formed steel stud wall systems under fire attack from one

side based on full scale fire tests.

7) To determine experimentally the fire performance of Gypsum plasterboards

8) To determine the fire performance of different types of insulations such as

glass fibre, rock fibre and cellulose fibre including the effect of insulation

density and thickness on the fire performance of stud wall systems.


9) To determine experimentally the temperature effects on the mechanical

properties of cold-formed steel and develop empirical equations to predict the

yield strength and modulus of elasticity at elevated temperatures.

10) To identify deficiencies in the conventional stud wall systems and the

improvements provided by the new stud wall systems assembled with the

composite panels shown in Figure 1.9.

11) To develop idealised time-temperature profiles for existing and proposed cold-

formed steel stud wall systems exposed to standard (cellulosic) fire curve.

12) To provide accurate structural and thermal performance data for the numerical

modelling of cold-formed steel stud wall systems under fire conditions.

13) To provide simple methods to determine the fire resistance rating of LSF load

bearing walls under different loading conditions.

1.6: Research Method

The research method essentially consisted of the following steps.

Step 1 - Literature Review: A comprehensive literature review on previous

experimental and analytical works, mechanical and thermo-physical properties of

steel stud wall assembly components at elevated temperatures, cold-formed steel, and

cavity insulations.

Step 2 - Establish a fire research laboratory at QUT: A propane fired gas furnace was

specially designed and constructed. The furnace was designed to deliver heat in

accordance with the standard time-temperature curve as proposed by AS 1530.4 (SA,

2005) or ISO 834-1 (ISO, 1999). A compression loading frame capable of loading the

individual studs of a full scale steel stud wall system was also designed and built for

conducting load bearing tests.

Step 3 - Conduct Fire Tests: To determine the fire resistance rating, standard fire tests

were carried out by exposing one side of the wall specimens to heat in the specially

designed furnace with controlled fuel input to achieve the specified time-temperature


curve and simulate the specimen’s exposure to heat in a real fire. The test specimens

were representative of the construction elements subject to loadings and end

constraints similar to the conditions of actual components. Such standardised furnace

tests provided good comparative data for systems tested under identical conditions.

Following tests were conducted to improve the knowledge and understanding of the

structural and thermal performance of cold-formed steel stud wall systems.

1) The mechanical properties of 1.15 mm G500 cold-formed steel at elevated

temperatures using steady state tests in an electric furnace. Tensile tests were carried

out in the elevated temperature range of 1000C to 8000C at intervals of 1000C.

2) The thermal properties of Gypsum Plasterboards (FireSTOP, Boral Industries) was

studied in a temperature range from 200C to 11000C using the gas furnace.

3) Fire tests on different types (Based on number of plasterboards, type of insulation,

number of joints) of small scale non-load bearing wall specimens with and without

cavity insulations. An adapter for the large propane fired gas furnace was designed

and constructed for these tests. The adapter requires the use of only a single burner for

the small scale tests as against the six burners for the full scale tests.

4) Fire tests of specially designed composite panels to improve the fire resistance

rating of wall systems.

5) Ultimate load test on large scale steel stud wall specimen at ambient temperature. A

loading frame was specially designed and built to load the individual studs.

6) Fire tests in accordance with AS 1530.4 (SA, 2005) were carried out on large scale

load bearing wall specimens to study their thermal and structural response at elevated

temperatures. A total of 16 transducers (to measure the axial shortening and lateral

displacement of the wall) and 45 to 57 K type thermocouples (to measure the

temperature at various locations across the wall) were used.

Step 4 - Develop new improved wall systems: New cold-formed LSF wall systems

with increased fire resistance rating and lower lateral deformations than the

conventional cavity insulated systems was achieved by the use of external insulation.

The improvement was validated by conducting several fire tests of both small scale

and large scale wall systems.

Step 5: -Develop simple methods to predict fire performance of LSF stud wall

systems: Simple predictive models for the mechanical properties of high strength



steels at elevated temperatures, the idealised time-temperature profiles of studs in LSF

walls under fire conditions and their fire resistance rating as a function of varying

arrangements of plasterboards and insulation were developed using simple methods to

predict the failure times of non-load bearing and load bearing wall models.

1.7: Contents of Thesis

Chapter 2: Past research on thermal performance of steel stud wall systems at elevated

temperatures is presented covering a range of research papers, reports and thesis.

Chapter 3: Presents detailed experimental work carried out to determine the

deterioration of mechanical properties of high grade (G500) cold-formed steel with

increasing temperatures. The chapter focuses on high grade steel as it is fast gaining

popularity in the construction industry.

Chapter 4: Deals with a series of small scale experiments performed to determine the

thermal performance of gypsum plasterboard specimens and their composite panels

using different types of insulating material of varying thickness and density.

Chapter 5: This chapter examines and compares the thermal performance of nine

small scale non-load bearing wall specimens built using cold-formed steel frame,

gypsum plasterboards and various types of insulating material.

Chapter 6: This chapter examines and compares the structural and thermal

performance of nine large scale load bearing wall specimens built in a manner similar

to the small scale non-load bearing wall specimens.

Chapter 7: This chapter presents the outcomes of the tests performed on non-load

bearing and load bearing wall specimens and compares the structural and thermal

performance of the conventional wall models with the new models proposed in this

thesis. It then develops idealised time-temperature profiles and simple fire design

methods for various LSF stud wall systems considered in this study

Chapter 8: Presents the main findings and recommendations.

Chapter 2: Literature Review

This chapter presents a detailed literature review that covers a range of research

papers, reports and theses in the field of fire performance of light gauge steel frame

systems.

2.1: Experimental Research

Son and Shoub (1973) carried out fire endurance tests on two load-bearing stud wall

assemblies with glass fibre batt cavity insulation. Each assembly consisted of double

module walls of gypsum board and steel studs. The outer plasterboards were type X

Gypsum boards 15.9 mm thick while the inner ones facing the cavity between the

walls were 12.7 mm in thickness. Studs used were lipped channel sections (76.2 x

44.5 x 12.7 x 1.21mm). The glass fibre insulation used in assembly two was thicker

than the one used in assembly one. A uniformly distributed load of 15 kN/m was

applied to each wall. On exposure to fire from one side, the structural failure of the

fire exposed wall in assembly 1 occurred in 42 minutes as compared to 67 minutes in

assembly two. In both assemblies, the structural failure occurred only after the

collapse of the exposed plasterboard. It was also observed, that as compared with

assembly one the heat penetration in the second assembly was much slower. This was

attributed to the thicker insulation used in assembly two.

The investigators recommended the use of two layers of plasterboard with staggered

board joints to eliminate the direct passage of heat onto the steel studs when the joints

open out in the fire.

Klippstein (1978) carried out tests on ten wall panels exposed to ASTM E119 fire.

The first seven of these tests were sponsored by American Iron and Steel Institute

(AISI) and conducted at Underwriters’ Laboratory (UL). The other tests were

sponsored by U. S. Steel Corporation (USSC). The objective of these tests was to

empirically determine the variation of the stud temperature and the lateral deflection

of the stud during the test up to the failure of the wall, which would serve as inputs in

predicting the structural behaviour of the studs when exposed to ASTM E119 or

similar fires.


All panels consisted of C–shaped steel studs of varying thickness and dimensions,

spaced at 600 mm centres. One to three layers of (12.7 mm or 15.9 mm) gypsum

boards were attached on the fire side. One gypsum board was attached to the cold side

of the panels. Out of the ten wall panels, four wall panels had fibreglass insulation

placed between studs and claddings as cavity insulation. The average load per stud

ranged from 15.12 kN to 44.7 kN.

The steel studs closer to the wall ends were seen to be at lower temperatures than the

central ones, possibly due to the flow of cold air from outside into the furnace

chamber caused by a negative pressure inside the furnace. The central studs being at a

higher temperature than the studs at the wall ends, expanded more and consequently

attracted more load during the initial phase of the fire test. In the later phase of the test

the load was redistributed to the studs farther away from the central ones and the

failure times of the wall panels varied from 37 minutes to 127 minutes, with the

higher failure times generally seen associated with greater number of gypsum boards

on the fire side and lower wall loads.

Kwon et al. (1998) carried out four fire tests on two types of load bearing exterior

walls (Wall-1 and Wall-3) at the fire Insurers Laboratories of Korea (FILK). The wall

specimens were 3 m long and 2.4 m high. They consisted of C-shaped lipped steel

studs 140 X 40 X 1 mm spaced at 450 mm on centres. Rockwool insulation was

placed between the studs and the claddings as cavity insulation. Another 10 mm thick

unspecified insulation material was used as the base lining layer on the exterior side

of the wall specimens. Two layers of 12.5 mm thick type X gypsum plasterboards

made in Korea were attached on both sides of the framing for the Wall-1 specimens.

Out of the two Wall-1 specimens, one specimen was exposed to fire from outside and

the other specimen was exposed to fire from inside. The exhibited structural failure

times were 28 minutes and 40 minutes, respectively.

Wall-3 specimens were lined on either side with one layer of 15.9 mm thick type X

Gypsum boards made in Canada. When tested for fire endurance from outside and

inside, the structural failure times observed were 25 minutes and 30 minutes,

respectively. The researchers observed that the fire resistant properties of the steel

stud walls depended mainly upon the fire resistant properties of the Gypsum boards


and to have a fire rating of one hour, at least two layers of 12.5 mm thick type X

gypsum boards would be required on either side of the framing.

SCI Publication (1993)

The publication in its section dealing with “Building Design using Cold-Formed Steel

Sections: Fire Protection” presents the fire ratings of cold-formed steel sections using

planar protection with respect to various parameters such as the number of

plasterboards, protection thickness, type of plasterboard and insulation as reproduced

in table 2.1 below.

Table 2-1: Fire Resistance of Typical Floors, Walls and Partitions Comprising Cold-Formed Steel Sections and Planar Board Protection, and Heated from One

Side Only (From SCI, 1993) Fire Resistance (hours) Form of

construction

Number

of layers

of board

Protection

thickness

(mm)

Plasterboard Fire

resistant

board†

Notes

1 12.5 - 0.5 - Floors with

ceiling

protection

2

2

12.5

15

0.5

-

1.0

1.5

+ 60 mm glass

wool mat**

-

1

1

1

12.5

12.5

15

0.5

0.5

0.5

0.5

1.0

1.0

-

+ 25 mm glass

wool mat*

-

Non-load

bearing

walls

(partitions)

(number of

layers per

face)

2

2

2

12.5

12.5

15

1.0

1.0

1.5

1.5

2.0

2.0

-

Boxed section

depth > 60 mm

-

1 12.5 - 0.5 - Load

bearing

walls

2

2

12.5

15

0.5

-

1.0

1.5

-

-

† ‘Fireline’ or ‘Firecheck’ board or similar

* Glass wool mat is required for insulation purposes for more than 30 minutes fire resistance

** For floors, the glass wool mat is only necessary for fire resistant suspended ceilings


Sultan and Lougheed (1994) performed several small scale fire resistance tests on

gypsum board clad steel stud wall assemblies using glass fibres, rock fibres and

cellulose fibres as cavity insulation. The test specimens were 914 mm in height and

914 mm in width with depth depending upon the number of layers of gypsum board

used. The small scale wall assemblies were constructed using two types of gypsum

boards (regular and Type X). Details of the test specimens are as shown in Table 2.2

and Figure 2.1

Table 2.2: Small Scale Assembly Parameters and Fire Test Results (Sultan & Lougheed, 1994)

Assembly Number

Gypsum Board Layers (Exp/Unexp)

Gypsum Board

Thickness (mm)

Gypsum Board Type

Insulation Type

Insulation Thickness

(mm)

Point Failure (min)

Average Failure (min)

S - 09 1 X 1 12.7 X None - 46 46 S – 22 1 X 1 12.7 X GF 90 46 48 S – 14 1 X 1 12.7 X RF 40 69 72 S – 15 1 X 1 12.7 X CF 90 69 71

S – 10 1 X 2 12.7 X None - 86 86 S – 23 1 X 2 12.7 X GF 90 88 93 S – 26 1 X 2 12.7 X RF 90 114 117 S – 18 1 X 2 12.7 X CF 90 134 135

S – 12 2 X 2 12.7 X None - 129 129 S – 25 2 X 2 12.7 X GF 90 139 139 S – 27 2 X 2 12.7 X RF 90 160 162 S – 21 2 X 2 12.7 X CF 90 157 163

S – 01 2 X 2 12.7 RL None - 82 84 S – 32 2 X 2 12.7 RL GF 90 74 76 S – 33 2 X 2 12.7 RL RF 90 98 101 S – 34 2 X 2 12.7 RL CF 90 102 ***

X – Type X Gypsum Board 12.7 mm thick (7.83 kg/m2)

RL – Regular lightweight gypsum board with glass fibre in gypsum core (7.35 kg/m2)

GF – Glass Fibre Insulation RF – Rock Fibre Insulation

CF – Cellulose Fibre Insulation


Fire Exposed Side Fire Exposed Side

Unexposed Side Unexposed Side

1 x1 Assembly 1x2 Assembly (a) (b)

Fire Exposed Side

Unexposed Side

(c)

Figure 2.1: Construction of Assemblies

The authors observed that compared to uninsulated wall assemblies, the cavity side of

the exposed gypsum board of insulated wall assemblies heated up more rapidly

reaching temperature levels of 7000C far earlier as compared to the temperature rise of

the exposed gypsum board in an uninsulated wall assembly. Compared to the

uninsulated assemblies, the assemblies with cavity insulation recorded much higher

temperatures on the exposed side of the cavity just after the calcination of the exposed

board.

The authors observed that, in the case of type X gypsum board, the temperature

increase was primarily due to the burning of combustible material used in the

insulation, whereas with regular gypsum boards the temperatures on the exposed side

of the cavity were comparable to the furnace temperatures implying a rapid and

extensive failure of the gypsum board. The advantage gained in the use of cavity

insulations was that, the board on the ambient side remained at a much lower

temperature for a longer time as compared to the board in the uninsulated wall

assembly.

After the failure of the gypsum board on the exposed side, the cavity insulation helped

in providing some initial protection against fire to the gypsum board on the ambient

side. This protection offered was around 5 – 10 minutes with glass fibres, 10 – 15


minutes with rock fibres and 25 - 30 minutes with cellulose fibre insulation. The

increase in temperature of the unexposed gypsum after the initial protection period

was found to be most rapid in case of assembly with glass fibre insulation in the

cavity. The temperature in the cavity was seen to exceed even those measured in the

uninsulated assembly, thus giving a neutral effect on the fire resistance of assemblies

constructed with type X gypsum board. For regular gypsum board assemblies, this

increased temperature led to an earlier failure of the boards, thus in fact, lowering the

fire resistance rating of the assembly below that of the uninsulated assembly.

The authors remarked that the Rock and Cellulose fibre cavity insulations, gave

approximately a 30 minute improved fire resistance when compared with uninsulated

wall assemblies.

Gerlich (1995) conducted tests on LSF load bearing walls lined with Gypsum plaster

boards at elevated temperatures at the fire testing laboratory of BTL (Building

Technology Limited) pertaining to BRANZ (Building Research Association of New

Zealand). Gerlich carried out three full-scale furnace tests on specimens as detailed in

Table 2.3

Table 2.3: Full Scale Fire Test Specimens (Used by Gerlich, 1995)

Fire Test Number FR2020 FR2028 FR2031 Wall Height (mm) 2850 3600 3600 Steel Grade (MPa) 300 450 450 Framing Type C-section Lipped C-section Lipped C-section Stud size (mm) 76 X 32 X 1.15 102 X 51 X 1.0 102 X 51 X 1.0 Stud spacing (mm) 600 600 600 Nog spacing One row central One row central One row central Frame connections Welded Welded Welded No. of load-bearing studs 4 4 4 Load (kN/stud) 6 16 12 Lining exposed (mm) 16.0 12.5 12.5 Lining unexposed (mm) 16.0 12.5 9.5 Fire curve ISO 834 ISO 834 ‘real’

The three wall frames were actually made of six cold-formed studs welded to the top

and bottom channels. The flanges of the top channel were cut in the end bays to

minimise load transfer to the cooler edge studs, and thus have only central 4 effective

load bearing studs. The steel frames were lined on both the faces by a single layer of

glass-fibre reinforced gypsum plasterboard. The sheets were fixed vertically to all the


studs with long self-drilling drywall screws spaced at 300 mm c/c. The vertical joints

formed over studs were tape reinforced and plaster stopped. The sheet length covered

the full height so that horizontal joints were not required.

The two wall specimens named FR 2020 and FR 2028 were exposed to the standard

ISO 834 time-temperature curve whereas the third specimen FR2031 was exposed to

a much severe time-temperature curve to simulate a real fire condition. Gerlich

observed that in the actual fire test, the furnace temperatures were considerably

different from those required as he experienced some difficulties in driving the

furnace.

Tests were carried out in accordance with AS 1530: Part 4 (SA, 1990). They were

well instrumented giving detailed information regarding temperatures and deflections.

In all the tests vertical thermal expansion of the steel-framing members was allowed

to take place freely, so that additional axial loads would not build up when the studs

expand. Vertical displacement of the loading platen indicated the thermal expansion

of the studs.

Horizontal displacements in all the tests were observed to be towards the furnace. No

evidence was found of significant double curvature along the length of the studs in

any of the test specimens implying that thermal deformations override any rotational

end restraint, justifying the assumption of hinges at the stud ends in a typical fire test

set up.

The failure of all the tested walls was governed by the structural collapse of load

bearing studs through buckling of the compression flange on the ambient side of the

wall assembly near midspan. The reversal of the vertical displacement of the loading

platen indicated the failure of the test specimens. Lateral buckling about the minor

axis and flexural torsional buckling was prevented by the lateral support provided by

the unexposed linings in tests FR2020 and FR2028. In test FR2031 torsional buckling

was noticed as the thinner unexposed lining at elevated temperature significantly

degraded and could not provide sufficient lateral support and prevent this buckling

mode. The structural failure times were observed at 72, 44 and 32 minutes for the

specimens FR2020, FR2028 and FR2031, respectively. In all the experiments it was


the structural failure, which led to integrity failure. All the wall specimens satisfied

the insulation criterion right up to structural failure.

It was observed that walls with low levels of axial load may perform better in fire

tests than in actual fire situation because frictional restraints (between specimen and

specimen holder) and redistribution of load can enhance the test results.

In the work done by Gerlich (1995)

1) Information regarding furnace pressures is not documented adequately.

2) Actual furnace temp-time control was not satisfactory.

3) Vertical thermal expansions of the studs were allowed at elevated

temperatures to occur freely, which is unlikely in real life conditions.

4) It would have been better to have horizontal joints instead of vertical joints of

the plasterboard on the exposed face, because when the boards shrink due to

loss of water the joints will open out and expose the steel stud underneath over

the entire length to direct fire.

5) Since too many parameters affecting the performance of the wall have been

altered between the specimens it is difficult to draw any correlation between

the samples tested.

Kodur et al. (1999) conducted three fire resistance tests on load bearing LSF walls at

FRM of IRC/NRC. They studied the behaviour of the wall assemblies (designated as

W1, W2 and W3) exposed to standard ISO 834 fire conditions and having Glass fibre,

Rock fibre and cellulose as cavity insulation material, respectively.

The wall assemblies tested were 3048 mm high, 3658 mm long and 157 mm deep

with each assembly consisting of a single row of galvanized cold-formed steel studs

(minimum yield strength 230 MPa) protected with two layers of fire resistant gypsum

boards on each side. The studs used were lipped C – sections 92.1 mm deep by 41.3

mm wide with 12.7 mm stiffening lips. They were spaced at 406 mm, 610 mm and

406 mm in the three specimens W1, W2 and W3, respectively.


Figure 2.2: Typical Steel Frame Fabrication Layout for Wall Specimens (Kodur et al.’s (1999) Work Reproduced By Alfawakhiri, 2001)

Figure 2.3: Location of Temperature Measurements and Simulation Boundaries (Kodur et al.’s (1999) Work Reproduced By Alfawakhiri, 2001)

The lateral stability of each wall specimen in its plane was provided by top and

bottom channel tracks, diagonal cross bracings, two rows of bridging and solid

blocking in four locations as shown in Figure 2.2. Nine resilient channels spaced at

406 mm c/c were attached perpendicular to the studs on the fire exposed side of all the

frames.

The gypsum boards on the fire exposed side were oriented horizontally and attached

to the resilient channels (giving horizontal joints). On the unexposed side, the boards

were oriented vertically and attached to the steel studs (giving vertical joints). All the


gypsum board joints were taped and covered with two applications of joint compound.

The testing of the wall specimens was carried out in a propane fired wall furnace of

NRC.

Total vertical load applied to the specimens W1 and W3 was 73.3 kN and to the

specimen W2 was 46.8 kN. This load was achieved by incremental loading and then it

was held constant for 45 minutes before furnace ignition and then throughout the fire

endurance test. The number of load bearing studs was 10, 07 and 10 in the assemblies

W1, W2 and W3, respectively. The test set up was very well instrumented giving

detailed information pertaining to measured temperatures and deflections. A slight

negative furnace pressure around –30 Pa at the bottom probe and neutral at the top

probe was maintained throughout the tests.

In all the tests, it was observed that the structural failure was rather abrupt resulting in

the overall out of plane buckling of the walls in the direction away from the furnace

with compression failure (local buckling) of the hot flange. In all the tests the lateral

deflections were initially positive (towards the furnace) due to thermal bowing effects.

After reaching the critical temperature the lateral deflections for the lower ¼ portion

of the wall showed reversal of displacement (i.e. away from the furnace) due to loss of

stiffness of the hot flange leading to local compression buckling near the first web

perforation in the web of the stud at 0.2 H level. Average hot flange temperatures just

prior to structural failure were about 550oC, 800oC and 650oC for specimens W1, W2

and W3, respectively. Lateral deflections of the end studs though not measured in any

of the tests were observed to be less than the central studs. This was because the end

studs were heated much slower and experienced smaller temperature gradients than

the central studs. Vertical displacements of the loading beam indicated thermal

expansion of the studs. Gradual downward movement was observed just prior to

structural failure.

The structural failure in all the tests occurred before heat penetration failure of the

unexposed side could take place. Gypsum boards on the fire-exposed side of the

specimen were observed to fall off several minutes prior to structural failure. Visual

inspection of cavity insulation material after the tests revealed that glass fibre

insulation had undergone only limited damage in certain areas, rock fibre insulation

was in good condition, but cellulose insulation was totally burnt out. Structural failure


before integrity failure in all the tests indicated that much higher fire resistance ratings

was possible for similar non load-bearing LSF assemblies.

The time-temperature curves indicated that the two layers of board on the exposed

face provided about 40 minutes of delay in the temperature rise of the hot flange. This

duration was insensitive to type of insulation used in the stud walls. The temperature

rise in steel studs after this initial period was the fastest in wall W1 with glass fibre

insulation and the slowest in wall W3 with cellulose fibre insulation. This Kodur et al.

(1999) observed was probably due to the lower bulk density and associated lower

thermal capacity of the glass fibre insulation and the higher bulk density of the

cellulose fibre insulation.

The structural failure times of wall specimens W1, W2 and W3 were observed to be

55, 73 and 70 minutes, respectively. Kodur et al (1999) also observed that the higher

fire resistance of wall W2 as compared to W3 was probably due to greater stud

spacing and therefore lower total load on specimen W2 as compared to specimen W3.

The structural failure of all specimens clearly highlighted the significant role played

by loading in determining the failure times at elevated temperatures.

Following points could be noted in the work of Kodur et al. (1999)

1) Failure of studs in all the test specimens was due to the local compression

failure of the hot flange as opposed to the compressive failure of the cold

flange near stud mid height with the studs buckling towards the furnace as

observed by Gerlich (1995).

2) The negative pressure within the furnace, sucked in cool outside air

leading to a drop in temperature of the edge studs as compared to the

central ones leading to unequal thermal expansions of the studs and thus

contributed to the building up of internal stresses within the framework.

3) Vertical thermal expansions of the studs were allowed.

4) Magnification of the thermal bowing deflections due to axial compression

was ignored.


Alfawakhiri (2001) conducted three fire resistance tests on load bearing LSF walls

with plasterboard linings to study the effects of cavity insulation, stud spacing and

effect of resilient channels on failure times. His three specimens were based entirely

upon the previous work carried out by Kodur et al (1999) on specimens W1, W2 and

W3. Alfawakhiri named his three specimens as W4, W5 and W6.

He suggested:

1) Test W4 as a duplicate of test W1, but without cavity insulation to

establish effect of cavity insulation on the fire resistance of load bearing

LSF walls.

2) Test W5 as a duplicate of test W2, but with 406 mm c/c stud spacing to

establish the effect of stud spacing on the fire resistance of load bearing

LSF walls.

3) Test W6 as a duplicate of test W4, but without resilient channels to note its

effect on the fire resistance of load bearing LSF walls.

Tables 2.4 and 2.5 help in comparing the test set-ups and results of both Kodur et al.

(1999) and Alfawakhiri (2001).

Table 2.4: Summary of Fire Resistance Tests on Load-Bearing LSF Walls by Kodur et al. (1999)

Fall of Time of Gypsum board

*

(min)

Temperature rise +

(oC)

Sp. No.

Stud Spacing

(mm)

Insulation Type

(Fibre)

Resilient Channels*

Load **

(kN/m)

Face layer

Base layer

Structural

Failure Time

(Min) Max Ave.

W1 406 Glass Yes 21.5 50 In place

55 52 36

W2 610 Rock Yes 14.3 57 67 73 50 42

W3 406 Cellulose Yes 21.5 57 In place

70 42 37

* On exposed side ** Including Self Weight + On unexposed side, under pads at failure time.


Table 2.5 Summary of Fire Resistance Tests on Load-Bearing LSF Walls by Alfawakhiri (2001)

Fall of Time of Gypsum board on

Exposed side (min)

Temperature rise on unexposed

side, under pads at failure time

(oC)

Sp. No.

Stud Spacing

(mm)

Insulation Type

(Fibre)

Resilient Channels

on exposed

side

Load Including

Self-weight

(KN/m)

Face layer

Base layer

Structural Failure Time

(Min)

Maximu Avera

W4 406 - Yes 21.5 58 In place

76 64 60

W5 406 Rock Yes 21.5 53 In place

59 37 26

W6 406 - No 21.5 In plac

In place

83 76 69

Alfawakhiri (2001) used the same method and material as Kodur et al (1999) to

construct his wall specimens. The test set-up and instrumentation of specimens W4,

W5 and W6 was done in a manner similar to specimens of Kodur et al. so that

comparisons could be drawn accurately.

He observed that

1) All the three wall specimens exhibited structural failure before any significant heat

penetration to the unexposed side could occur indicating the possibility of much

higher fire ratings for non-load bearing LSF wall assemblies.

2) The delay of approximately 40 min in the temperature rise of the hot flanges of the

studs in all the tests due to the protection offered by dual layers of fire resistant

gypsum boards indicated that the delay could be regarded as a stable property of the

boards as it remained unaffected by the parameters of the test series.

3) The comparison of the tests suggested that the insulation placed in the wall cavity

reduced the fire resistance of load bearing LSF walls. The insulation restricted the

passage of heat through the cavity causing an accelerated rise in the temperature of

the hot flange and a delayed temperature rise in the cold flange on the ambient side.

Hence in insulated walls there was a high temperature gradient across the steel section

as compared to a very low temperature gradient (almost uniform) across the steel

cross section of uninsulated walls.


4) Structural buckling of the uninsulated wall frames was due to the overall buckling

of the studs towards the furnace, finally leading to local compression failure of the

cold flange near the mid height. In case of insulated walls the lateral deflections were

seen to be initially positive i.e. towards the furnace followed by an early reversal in

the direction of deflection at a height of 0.25H from base leading to structural failure

by local compression failure of the hot flange near that level as shown in Figure 2.4

He concluded that

1) A comparison of W2 and W5 showed that wider stud spacing improved the fire

resistance of load bearing LSF walls in standard tests.

2) A comparison of W4 and W6 showed that use of resilient channels reduced the

fire resistance of load bearing LSF walls because it reduced the ability of fire exposed

gypsum boards to remain in place.

Figure 2.4: Structural Failure Modes: (a) Uninsulated Walls (b) Insulated Walls (Alfawakhiri, 2001)

Following points could be noted in the work of Alfawakhiri (2001)

1) Test W5 was intended to be a duplicate of Test W2 with the objective of

testing the effect of stud spacing on fire resistance of LSF wall assemblies. To

achieve this Alfawakiri changed the stud spacing from 610 mm c/c (in W2) to

406 mm c/c in specimen W5. He also changed the loading from 14.3 kN/m (in


W2) to 21.5 kN/m (in W5), which actually is detrimental for the fire resistant

property of the wall. Hence the early failure of wall W5 as compared to wall

W2 cannot be entirely attributed to the decreased spacing of the studs.

2) Vertical thermal expansions of the studs were allowed.

3) The end studs remained significantly cooler than the central ones and

exhibited smaller thermal bowing than the studs in the central part developing

undesirable internal stresses within the framework.

Feng et al. (2005) conducted eight tests on loaded full-scale steel stud walls, two at

ambient temperature and six exposed to the standard fire condition on one side. The

tests were carried out in the UMIST fire-testing laboratory. Each wall panel was of

size 2200 X 2000 mm. Studs used were lipped channel sections (either 100 X 54 X 15

X 1.2 mm or 100 X 56 X 15 X 2 mm ) with a minimum yield strength of 350 MPa

spaced at 750 mm centres. Two elongated service holes were provided at the centre of

the webs, one at 300 mm from the top and the other at 300 mm from the bottom.

Unlipped C - sections of size 100 X 56 X 2 mm were used as top and bottom tracks.

Lateral bracing was provided by 4 flat bars fixed horizontally, two on either side of

the framing, one placed at 650 mm from the top and the other at 650 mm from the

bottom. The frame assembly was lined with one layer of 12.5 mm thick Fireline

Gyprock board on either side. Isowool 1000 was used as cavity insulation. The

Gyprock plasterboards were fixed horizontally to the steel studs by screws at 300 mm

centres. Three different load levels, being 0.2, 0.4 and 0.6 times the load carrying

capacity of the same panel tested at ambient temperature, were applied during the six

fire tests on the three panels using each type of lipped channels respectively. The

instrumentation consisted of 6 control thermocouples, placed inside the furnace to

regulate the average furnace temperature in accordance with BS 476 fire time-

temperature relationship, 30 thermocouples placed within the wall specimen to obtain

the time-temperature history of the assembly, 9 displacement transducers to record the

vertical in plane movement of the panel and nine displacement transducers placed at

three locations along each of the three studs to measure the out of plane movements of

the test panel.


Load was applied using three hydraulic jacks, located directly above the three lipped

channels of the test panel. The load was transmitted onto the panel through a moving

steel beam of the reaction frame. The load was kept unchanged during the fire test and

the specimen was considered to have failed when the applied loads could not be

maintained.

The investigators concluded that the failure mode was generally observed to be due to

global buckling of the studs about the major axis, with the flexural buckling of the

steel channels being restrained by the unexposed gypsum board. The investigators

recommended the use of non-combustible cavity insulation as it was observed that in

some of the specimens that the panel failure took place shortly after the burning out of

the insulation material. The investigators also observed that the fall of the gypsum

boards on the fire side was an effect and not a cause of the panel structural failure. It

was observed that under the same load ratio, panels using thinner channels had lower

fire ratings than panels using thicker channels. At higher load ratios (0.4 and 0.7)

panels using the thinner (1.2 mm) channels did not achieve even a 30-minute standard

fire resistance.

Sultan (2010) conducted 41 full-scale wall fire resistance tests at the National

Research Council of Canada, in accordance with ULC-S101 standard fire exposure, to

determine the gypsum board fall temperatures from the wall panels. The tests used

assemblies with wood and steel studs protected with either one or two layers of Type

X gypsum board and with or without insulation in the wall cavity. The temperature

criterion recorded for the fall off of the plasterboards was based on the sudden

temperature rise measured on the back side of the fire exposed gypsum board caused

by its failure. The parameters studied included resilient channels, spaced either 406

mm o.c. or 610 mm o.c. installed between the gypsum board and framing for sound

reduction purposes. The insulations used were glass and rock fibre batts and

cellulosic fibre insulation either spayed wet or dry blown in the wall cavity.

Plasterboards used were of Type X gypsum board 12.7 mm or 15.9 mm thick. The fall

off temperatures for assemblies with a single layer of gypsum board, with and without

insulation in wall cavity, and with different screw spacing was observed to be in the

range of 7550C to 7850C. The fall off temperatures for both single and two layers of

gypsum board was observed to have very little difference.


2.2: Analytical Research

Gerlich (1995) proposed a temperature sensitive structural analysis spreadsheet in

accordance with the AISI design manual (AISI, 1991) with strength and stiffness

expressions suitably modified to take into account the temperature effects. He

assumed the studs were free to rotate and expand at both ends. The thermal

deformations induced due to the temperature gradient across the steel stud (Thermal

bowing) was determined using the following expression (Cooke, 1987)

1 = L2T / 8D (2.1)

Where,

1 = Horizontal deformation at midspan due to thermal bowing

= Expansion coefficient for steel.

L = Length of member

T = Temperature difference across the member

D = Member depth

Mean stud temperature was used to calculate the expansion coefficient of steel. The

expression gave reasonably good predictions upto temperatures less than 400oC. At

higher temperatures of steel, the temperature difference across the steel would reduce

giving lower calculated deflections. Actual deflections would not return to the

calculated values due to plastic deformation of steel. Gerlich (1995) conservatively

assumed the calculated deformations (1) to remain constant when temperature

gradients (T) decreased. It was achieved by not allowing the value of calculated 1 to

be less than the one calculated for the previous step. The total horizontal deflection for

the system was calculated by adding the thermal deflection (1) and the deflection

(2) due to P- effects.

To obtain the deflection due to P- effects, the deflection 1 (Stress free thermal

deformation) was treated as the initial eccentricity. The initial bending moment P1

then gave an additional horizontal deflection 2


The P- component was predicted analytically by solving the moment equilibrium

equation

ET Ix d2 2/dZ2 = Pa (1 + 2) (2.2)

Figure 2.5: Total Horizontal Deflection for Load-bearing Systems

(Gerlich et al., 1996)

The solution of the above equation yielded deflection 2 at mid-height as

2 = 1 {[1/cos ( L/2)] -1} (Gerlich et al. 1996) (2.3)

Where,

XTa IEP / (mm-1)

With ET = Modulus of elasticity of steel in MPa at average stud temperature.

Ix = Second moment of area of the stud cross-section in mm4

Pa = Applied axial stud load in N

1 = Initial eccentricity at stud ends in mm (Taken equal to thermal bowing

deflection at midspan)

2 = P- deflection in mm

L = member length (wall height) in mm


The total horizontal deflection () was expressed as = 1 + 2. The studs were then

analysed as compression members subjected to axial load P and bending moment P.

A critical temperature was determined such that the maximum permissible stud load

was equal to the applied axial load. This temperature was then compared with the

compression flange temperature on the ambient side of the wall assembly to find the

failure time.

To model the heat transfer and steel framing temperatures, Gerlich (1995) used a

commercially available heat transfer model TASEF (Sterner and Wickstorm, 1990).

TASEF was found to give good correlations with measured values on the exposed

face of the wall. Discrepancy was noted on the ambient side of the wall, with TASEF

values lower (unconservative) than the actual measured values. This was attributed to

the inability of TASEF to model mass transfer and consider ablation or degradation

(with opening of joints) of the exposed lining allowing sudden rise in measured

temperature on the ambient face.

It was observed from specimen FR2031 that the temperature predictions of TASEF

for fires significantly hotter than ISO 834 were unconservative giving low values. As

TASEF under-predicted the lining and framing temperatures on the ambient side, it

gave a greater temperature difference across the stud cross section leading to a higher

calculated horizontal deflection due to thermal bowing effects (Since 1 is

proportional to T). Gerlich (1995) observed that the use of TASEF gave greater steel

stresses (due to higher predictions of deflection) and thus conservative failure time

predictions within 80 to 90% of test results.

Gerlich (1995) used the TASEF and temperature sensitive structural analysis

spreadsheet model to generate data and correlate the failure times of LSF wall

specimens with the load ratios and thickness of plasterboard linings when exposed to

ISO 834 fire conditions. The load ratio versus failure time curves drawn for various

board thicknesses were found to give 62% to 73% accuracy when compared with

measured values.

Milke (1999) observed that fire resistance of structural assemblies depends on fire

exposure conditions, material properties at elevated temperatures, thermal response of

the structure and structural response of the heated assembly. Heat transmission limits


were established to prevent the ignition of combustibles in contact with the unexposed

side of the assembly (Schwartz and Lie, 1985).

Alfawakhiri (2001) used a computer program TRACE (Temperature Rise Across

Construction Elements) to model heat transfer through LSF walls exposed to fire. This

program developed by him was based on an explicit finite difference integration

algorithm (Sultan 1991) to solve one dimensional transient heat transfer equations.

The presence of the steel frame was neglected in the heat transfer simulations.

A large number of numerical trials was conducted on uninsulated wall specimens (W4

and W6) and the properties of gypsum board were calculated by matching the

measured and simulated temperature histories at all the simulated boundaries of the

wall specimens. Further numerical trials were conducted on insulated walls (W1, W3

and W5) using the calibrated gypsum board properties so as to obtain the properties of

insulation materials (Glass fibre batts, loose fill cellulose and Rock fibre batts). The

apparent thermal properties so calculated, also accounted to some extent the physical

phenomena other than heat transfer, such as mass transfer, phase change, etc. This

Alfawakhiri (2001) observed, was due to the fact that the temperature rise in LSF

walls exposed to fire was affected by processes not described by heat transfer, such as

migration of moisture vapours within the board, penetration of cool ambient air or hot

furnace gases into the cavity etc.

The other parameter affecting the simulated temperature histories was the fall off of

gypsum boards. The TRACE model used by Alfawakhiri models the spalling of

gypsum boards by removing it from the simulation at a user specified time. It was

observed that the simulated temperature histories were not very sensitive to the choice

of emissivity coefficients and even less sensitive to the choice of convection

coefficients.

In formulating the structural model Alfawakhiri (2001) made the same basic

assumptions as suggested by Klippstein (1978), except for the assumption of eccentric

loading as shown in Figures 2.6 and 2.7, which in part models rotational end

restraints.


According to Alfawakhiri the eccentricity ‘e’, developed due to the following reasons

1) Shift of the centre of steel section towards the cold flange due to the

deterioration of the stiffness of the hot flange with the increase in the

temperature gradient across the stud.

2) Shift of the load towards the hot flange due to rotation of the end studs

associated with thermal bowing.

3) Internal Stresses caused due to non-linear thermal strain gradients.

Figure 2.6: Thermal Bowing Figure 2.7: Stud End Conditions And Secondary Deflection (a) Uniform Heating, (b) Non-uniform Heating (Alfawakhiri, 2001) (Alfawakhiri, 2001)

Alfawakhiri (2001) gave the expression for eccentricity ‘e’

as e = (1-KR) Øβ-2 (2.4)

Where Ø = Thermal bowing curvature = αTδT/D and β2 = P/EI*

KR is a reduction coefficient

αT = Thermal expansion coefficient for steel (Lie 1992)

= (12 + 0.004 TA)10-6 (2.5)

D = Stud section depth

δT = TH - TC = Temperature difference across stud section in 0C


TA = 0.5 [ TH + TC] = average stud temperature in 0C

P = Vertical load applied at stud ends

E = 203000 MPa = Modulus of Elasticity of steel at room temperature

I* = Elastic modulus- weighted moment of Inertia of the unreduced stud section

about the neutral axis parallel to flanges.

Alfawakhiri (2001) gave the shape of the stress free initial imperfection y1(Z) caused

by thermal bowing as y1(Z) = 0.5 ØZ(H -Z)

The secondary lateral deflection y2(Z) caused by the vertical load ‘P’ with eccentricity

‘e’ was obtained by solving the differential equation

EI* d2y2(Z)/dZ2 = P[y1(Z) + y2(Z) –e] (2.6)

Total lateral deflection ‘y(Z)’ was expressed as

Y(Z) = y1(Z) + y2(Z) (2.7)

= (Øβ-2 - e) [tan(0.5βH)Sin(βZ) + Cos(βZ) – 1]

Substitution of Eq. 2.3 in Eq. 2.4 gave the expression for the lateral deflection ‘Δ’ at

the mid-height of the stud,

Δ = KR Øβ-2 {[1/cos(0.5 βH)] -1} (2.8)

Due to the temperature variation from the hot flange to the cold flange, the modulus

of elasticity varies across the steel stud section. Alfawakhiri used the reduction

coefficient ‘nT’ as expressed by Gerlich (1995) based upon Klippstein’s (1978, 1980)

work.

nT = ET/E

= 1.0 - 3T(10-4) + 3.7T2(10-7) – 6.1T3(10-9) + 5.4T4(10-12) (2.9)

The variation of ‘E’ across the flange was accounted by the ‘modulus-weighted’

moment of inertia ‘I*’

I* was quantified numerically by dividing the stud section into sufficiently large

number ‘q’ of two-dimensional elements, so that


q

I* = ∑ ni [Ii + Ai (xi – c)2] (2.10)

i = 1

Where

ni = Reduction factor for temperature Ti, calculated using expression for nT

from Eq. 2.9

Ii = Moment of inertia of element ‘i’ about its own neutral axis parallel to flanges,

Ai = Area of element ‘i’.

xi = Distance of element ‘i’ from the extreme fibre of the cold flange.

Ti = Temperature of element ‘i’ calculated from

Ti = Tc + (δT xi / D) (2.11)

And c= ∑ ni Ai xi / ∑ ni Ai

(2.12)

q q

i = 1 i = 1

Alfawakhiri (2001) incorporated the above expressions in the computer program

STUD that had been developed to model the structural behaviour of load bearing LSF

walls in fire. The program assumed the calculated deflections to remain constant

whenever δT decreased in time. In STUD simulations, the values of Δ and Y(Z) at any

time were not allowed to be less than their values in the previous steps to account for

the creep and stress relaxation in steel studs at temperatures higher than 400oC. The

simulated mid-height lateral deflections were found to be in good agreement with the

measured values for all the specimens when KR was taken equal to 0.6

In case of insulated walls it was seen that owing to a large temperature difference

across the studs cross-section, a large eccentricity ‘e’ would develop at the stud ends

causing negative secondary deflections y2(Z).

The numerical simulations suggested that stress free thermal bowing gets restrained

(reduced), due to restraints at end studs and by nonlinearity of thermal strain gradients

across the stud cross sections. The compressive stresses in the hot flanges were seen

to reach critical levels when δT values exceeded 350 oC leading to compressive failure


of hot flange. STUD checked for this failure mode at the perforated section (Z= 0.2H)

using the expression

ƒH = nH [ (P/A*e ) + P(e-y0.2H)/S*eH ] ≥ FyH (2.13)

where

ƒH = Compressive stress at the extreme fibre of the hot flange

nH = Reduction factor for the temperature TH, calculated using the Eq. 2.9

FyH = Yield strength of steel at temperature TH

A*e = Elastic modulus weighted effective stud section area in compression,

S*eH = Elastic modulus weighted effective stud section modulus in bending that

causes compression of hot flange.

The stud cross-section calculations were done in accordance with S136-94. The

effective cross-sectional dimensions were assumed to be insensitive to temperature

and were based on steel properties at room temperature and compressive stress ƒ = Fy.

The effective cross-sections were then used in the calculation of the temperature

dependent ‘modulus weighted’ properties A*e and S*eH as per Equations given below

A*e = ∑ ni Ai

(2.14)

q

i = 1

S*eH = ∑ ni [Ii + Ai (xi – c)2] / (D-c) (2.15)

q

i = 1

For uninsulated walls (W4 and W6), the section at mid-height was checked by STUD

for the compressive failure of cold flange at perforated section (Z= 0.4H) using the

expression

ƒc = nc [ (P/A*e ) + (P y0.4H)/S*ec ] ≥ Fyc (2.16)

where,

ƒc = Compressive stress at the extreme fibre of the cold flange


nc = Reduction factor for the temperature Tc, calculated using Eq. 2.9

Fyc = Yield strength of steel at temperature Tc

S*ec = Elastic modulus weighted effective stud section modulus in bending that causes

compression of cold flange calculated from

S*ec = ∑ ni [Ii + Ai (xi – c)2]/ c (2.17)

q

i = 1

The section properties A*e, S*eH and S*ec were based on three different effective cross-

sections, as the configuration of compression elements was different in each case.

Also the value of ‘e’ was taken as zero and was not used in the expression (2.16). This

was because as δT decreases in the final stages of tests (on uninsulated walls) the

heating rate in cold flange becomes higher than in hot flange. This combined with

creep and stress relaxation in steel at temperatures greater than 400oC causes gradual

reduction of eccentricity leading to an assumed value of zero near failure time in an

non-insulated wall specimen.

The STUD program carried out structural failure checks considering both failure

modes at every time step. It was seen that for uninsulated walls (W4 and W6) the

predictions of failure times from STUD showed reasonable agreement with test

structural failure times. For insulated walls (W1, W3 and W5) predicted failure times

agreed well with the initiation of the structural failure in central studs. This was

because of the quasi-elastic approach used in the STUD model formulation and the

effect of load redistribution to colder studs at wall ends in the final phases of the tests.

Predictions based on measured temperatures showed a better agreement with test

results than predictions based on simulated temperatures. The STUD simulations

showed that the major part of deflections observed in the fire resistance tests was due

to thermal bowing and not due to deterioration of steel stiffness.


2.3: Mechanical and Thermo Physical Properties of Steel Stud Wall Assembly

Components at Elevated Temperatures.

The fire resistance behaviour of load bearing, gypsum board protected, stud wall

assemblies can be modelled only with a thorough understanding of the mechanical

and thermo-physical response of individual components of the assembly at elevated

temperatures.

2.3.1 Gypsum Plasterboards

2.3.1.1: Introduction

Gypsum wallboards have been in use since the early 1900’s. They are widely used as

wall or ceiling linings in domestic housing or commercial buildings. The core of these

wallboards or plasterboards is made up of Gypsum i.e. calcium sulphate dihydrate

(CaSO4.2H2O), a naturally occurring non-combustible mineral. The core is

sandwiched between two layers of paper (see Figure 2.8), which are chemically and

mechanically bonded to the core to form flat sheets available in a range of sizes. The

papers provide sufficient tensile strength to the board to assist in handling and

transportation. Gypsum plasterboards have become very popular due to their non-

combustible core and fire resisting properties. Most gypsum boards are made with a

thickness between 10 and 20 mm.

Paper covers Gypsum Core

Figure 2.8: Gypsum Plasterboard

2.3.1.2: Types of Gypsum Plasterboards

The plasterboards are generally available in three main varieties i.e. Regular, Type X

and Special Purpose Boards.

1) Regular Plasterboards generally do not have any fire resistance rating and are made

up of low density Gypsum core without use of reinforcing fibres. They are mostly


used in constructing non-load-bearing walls. When exposed to fire the regular boards

tend to crack up and fall off soon after the burning up of paper facings, which takes

place at around 300oC.

2) Type X Board is a generic term that describes a Gypsum board with a specially

formulated core to provide a greater fire resistance than a regular board of the same

thickness. All type X boards contain some additives such as Vermiculite and Glass

fibre reinforcing to enhance the fire resisting properties. Vermiculite expands when

exposed to heat and thus partly helps in compensating the shrinkage of the Gypsum

core during calcination (i.e. dehydration). Glass fibres improve the mechanical

properties of the board, reduce shrinkage and ablation and thus enhance the stability

and integrity of the board, when exposed to fire.

3) Special Purpose Boards (some called as Type C Boards) are proprietary products

made by manufacturers to obtain superior fire or structural performance over Regular

or Type X boards. For the Gypsum boards to stay in place, in the stud wall assembly

they should possess sufficient tensile ductility to accommodate the thermal strain

incompatibility with steel studs i.e. as the boards tend to shrink and the studs tend to

elongate with rise in temperature. Special additives are used in proprietary

formulations to reduce shrinkage and enhance strength and ductility characteristics of

the Gypsum core.

2.3.1.3 Chemical Properties of Gypsum Plasterboard Linings

Gypsum contains approximately 21% by weight chemically bound water of

crystallization and about 79% calcium sulphate, which is inert below a temperature of

1200oC (Goncalves et al., 1996). In addition to water of crystallization it is found that

approximately 3% free water is also present inside Gypsum plaster, depending upon

the ambient temperature and relative humidity (Buchanan, 2001).

The fire retarding property of the gypsum board primarily stems from this water

content (Free water and water of crystallization). When the gypsum board is exposed

to fire, the free water and water of crystallization is gradually released and evaporated.

The dehydration i.e. release of water occurs in two phases. In the first phase also


known as calcination, Gypsum dehydrate loses some amount of water to yield

Gypsum hemihydrate (CaSO4.1/2H2O) commonly known as Plaster of Paris.

CaSO4.2H2O CaSO4.1/2H2O + 3/2H2O (2.18)

The above reaction (calcination) is endothermic in nature occurring between 100oC to

120oC and consumes large amounts of energy in order to evaporate the free water and

release as steam the chemically bound water of crystallization.

This absorption of energy delays the heat transmission through the board and causes a

temperature plateau on the unexposed face of the lining. The length of this plateau is a

function of the lining thickness, density and composition and is commonly referred to

as the ‘Time Delay’ (Gerlich et al., 1996).

Calcination leads to shrinkage and loss of strength of the sheet material. The progress

of calcination through the sheet thickness is retarded by the exterior layer of calcined

Gypsum on the fire exposed side which acts as a protective layer and adheres well

with the inner uncalcined layers.

The second phase of dehydration, i.e. complete dehydration occurs when the Gypsum

hemihydrate is transformed to Gypsum anhydrite.

CaSO4.1/2H2O CaSO4 + ½ H2O (2.19)

This reaction occurs at about 210oC according to Andersson and Jansson (1987) and

at about 600oC according to Sultan (1996) and at about 225oC according to Bakhtiary

et al. (2000). The temperature at which the second phase occurs much depends upon

the rate of heating. (Thomas, 2002).

2.3.1.4 Mechanical Properties

Most gypsum boards have a density between 550 and 850 kg/m3 .

Goncalves et al., (1996) reported the mechanical properties of Gypsum plasterboards

subjected to fire. The results were based on the tests conducted on plasterboards from

three Australian companies; Boral, C.S.R and Pioneer. The tensile strength

characteristics and modulus of elasticity (MOE) of plasterboard at 300oC and 500oC

are summarised in Table 2.6.


Table 2.6: Mechanical Properties of Australian Manufactured Plasterboards (Goncalves et al., 1996)

Board Type Failure Stress at 3000C (MPa)

Failure Stress at 5000C (MPa)

MOE at 3000C (MPa)

MOE at 5000C (MPa)

Boral 0.121 0.098

C.S.R 0.163 0.098

Pioneer 0.107 0.117

8.5 – 18.8 6.0 – 9.9

2.3.1.5 Thermo-Physical Properties of Gypsum

Gerlich et al. (1996) have quoted Thomas et al. (1994) who summarised the data

measured by Mehaffey (1991) for the thermal conductivity and enthalpy of glass fibre

reinforced gypsum plasterboard as a function of temperature. Thomas’s values for

thermal conductivity and enthalpy are shown in Figures 2.9 and 2.10, respectively.

The enthalpy values represent the summation of the product of specific heat and

temperature, expressed per unit volume. The authors have used the enthalpy values in

modelling to avoid numerical instabilities resulting from sharp peaks that may occur

in the specific heat of materials containing water, due to evaporation of moisture.

Sultan (1996) conducted tests at NRCC to obtain the thermo-physical properties of

Type X Gypsum Board. Measurements were carried out at a heating rate of 2oC/min

as it provided the maximum specific heat at approximately 100oC. The author has

given the results in the form of equations for Specific heat, Thermal conductivity and

Density. The equations are represented in the form of graphs in Figures 2.11, 2.12 and

2.13, respectively.


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400

Temperature [°C]

Th

erm

al c

on

du

ctiv

ity

[W/m

K]]

Figure 2.9: Thermal Conductivity of Gypsum Plasterboard (Franssen, 1999)

0

200

400

600

800

1000

1200

1400

1600

1800

0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400

Temperature [°C]

En

thal

py

[MJ/

K]]

Figure 2.10: Specific Volumetric Enthalpy of Gypsum Plasterboard

(Franssen, 1999)


0

2 000

4 000

6 000

8 000

10 000

12 000

14 000

16 000

18 000

20 000

0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400

Temperature [°C]

Sp

ec

ific

He

at

[J/k

gK

]

Figure 2.11 Specific Heat of Type X Gypsum Board (Franssen, 1999)

The two peaks in the specific heat curve indicate the dehydration of Gypsum, which

appears at temperatures around 100 oC and 650oC. The first peak is the main cause of

delay in the temperature rise of protected steel studs in the stud wall assembly. The

area under the peak gives the energy consumed per kg of Gypsum board to drive out

the water of crystallization and free water from its core.

At this stage (i.e. around 100oC) there is a drop in density and thermal conductivity of

the board. There is a steady rise in thermal conductivity beyond 400oC. Thermal

conductivity also depends upon density variations of Gypsum board (Clancy, 1999).

Thermal conductivity values above 400oC get affected by the presence of shrinkage

cracks in the board which depend upon the type and composition of board and the

nature of fire (Buchanan, 2001).

Conductivity increases on account of radiative heat transfer caused due to the opening

of cracks in the Gypsum boards at high temperatures and also due to ablation, a

process in which thin layers of calcined gypsum due to their cohesionless nature tend

to fall off the board. Cracking is more severe in fire with greater initial temperature

gradient. According to Manzello et al. (2005), cracks will be propagated in the order

of opening at plasterboard joint, cracks at screw points along the stud and transverse


cracks. Manzello et al.’s (2006) study recommends incorporating plasterboard

contraction and crack formation in thermal models.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400

Temperature [°C]

Th

erm

al c

on

du

ctiv

ity

[W/m

K]]

Figure 2.12 Thermal Conductivity of Type X Gypsum Board (Franssen, 1999)

0

100

200

300

400

500

600

700

800

0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400

Temperature [°C]

Den

sity

[kg

/m³]

]

Figure 2.13: Density Variation of Type X Gypsum Plasterboard on Heating (Franssen, 1999)


Harmathy (1988) gave a value of 0.88 kJ/kg.k as the specific heat of gypsum board

at ambient temperature with a peak of 7.32 kJ/kg.k at about 1000C giving

approximately 201 kJ/kg of gypsum as the area under the peak.

Mehaffey et al. (1994) gave a base value of 0.95 kJ/kg.k for specific heat. The

authors used 2oC/min and 20oC/min as the two scanning rates in their differential

scanning calorimeter. For 2oC/min they obtained a peak of 29kJ/kg.k at 95oC and for

20oC/min they got a peak of 14kJ/kg.k at 140oC. The area under both the peaks is

about 490 kJ/kg. Mehaffey et al. measured specific heat only upto 200oC and thus did

not record the second peak (Thomas, 2002).

Andersson and Jansson (1987). The authors did not mention a base value of specific

heat but reported that at 1000C, 75% of bound water is evaporated requiring 515 kJ/kg

of gypsum for the process to occur, and the balance 25% of bound water being driven

off at 2100C requiring 185 kJ/kg of gypsum (Thomas, 2002).

Thomas (2002) studied the two consecutive dehydration reactions that gypsum

undergoes when heated and suggested modifications for the thermo-physical

properties, so as to obtain smooth curves for enthalpy and thermal conductivity

suitable for input into a finite element heat transfer model. The values were calibrated

and validated using furnace and fire test data. Experiments were conducted on

plasterboards similar to the North American type C boards. The author assumed the

base value for specific heat of gypsum as 0.95 kJ/kg.k. This value was adopted from

Mehaffey et al. (1994). The energy required to complete the two stages of dehydration

was adopted from Andersson and Jansson (1987) and was taken to be 515 kJ/kg and

185 kJ/kg of gypsum, respectively. The first is assumed to occur between 100oC and

120oC and the second between 200oC and 220oC. Figure 2.13 gives its density

variation at elevated temperatures and Figure 2.14 gives the mass loss in gypsum

plasterboard undergoing heating (Thomas, 2002)


Figure 2.14: Mass Loss in Gypsum Plasterboard Undergoing Heating (Thomas,

2002)

Baux et. Al. (2008) conducted several fire tests on gypsum plasterboards using

varying proportions of silica fume as filler introduced into the plaster. They observed

that in the case of conventional gypsum plasterboards, the dehydration on exposure to

elevated temperatures leads to large thermal shrinkage due to the loss of bound water

molecules. This resulted in rapid development of cracks allowing passage of heat. To

reduce this tendency of crack formation the researchers added silica fumes as filler

material to the hemihydrates before hydration with an amount ranging from 10 to 60

wt%. From the fire tests they observed that the number of cracks on the fire exposed

face of the gypsum plasterboard decreased with the increase in the filler content with

no cracks developing at a 40 wt% filler amount. The reduction in cracking also was

observed to reduce the overall shrinkage of the plasterboard during the fire tests.

A serious drawback, however, was a decrease of the latent heat effect due to the

substitution of the plaster by silica fume. It was observed that as compared to pure

plaster, the higher the filler content, lower was the latent heat, thus reducing the heat

absorbing capacity of the plasterboard. A 30 wt% of silica fume was found to be

acceptable at temperatures below 10000C. Above this temperature reactions between

the binder and filler material led to melting and geometric instability. To overcome

this problem they are currently trialling aluminosilicate filler in place of silica fumes.

However, they have observed that higher density and thermal conductivity could be

the setback for this material.


Ghazi and Hugi (2009) observed that the thermo-physical properties of Gypsum

Plasterboards are affected by the chemical composition of the ingredients (i.e. the

various carbonates in the plasterboard). Depending on different ingredients different

endothermic reactions were observed to occur between room temperature and 9000C

impacting the sensitivity of thermal conductivity, effective heat capacity and density

with respect to temperature.

Frangi et. al. (2010) carried out experimental and numerical analysis on the fire

behaviour of protective cladding made of gypsum plasterboards at ETH Zurich. 17

small-scale fire tests were performed with non-loaded specimens lined with gypsum

plasterboards and subjected to ISO fire exposure. The fire tests were carried out in the

EMPA’s horizontal small furnace with the internal dimensions of 1.0 x 0.8 m.

Gypsum plasterboards of type A and F according to EN 520 and gypsum fibreboards

according to EN 15283-2 were studied. Gypsum plasterboards of type A are similar to

regular type of plasterboards with a porous gypsum core and no reinforcement or filler

material. Type F plasterboards are similar to type X used in North America with

gypsum core reinforced with glass fibres or fillers to improve the core cohesion at

higher temperatures. Gypsum fibreboards have a gypsum core reinforced with paper

fibres and are usually denser than plasterboards of type A and X.

The researchers observed that the overall thermal behaviour of different types of

gypsum board with different density, fibres and fillers was quite similar although a

general improvement was noticed in the mechanical properties (shrinkage, cracking,

and ablation) of the boards after complete dehydration. The tests indicated that the

layer backing the gypsum board may have a strong influence on the thermal behaviour

of the gypsum board. Insulating batts were observed to cause the fire exposed gypsum

boards to heat more rapidly and fail sooner.

2.3.1.6 STRUCTURAL BEHAVIOUR

The plasterboards play an important role in providing lateral stability to the steel

studs. They provide adequate restraint against torsional buckling and flexural buckling

of the stud about the minor axis. When the assembly is exposed to fire, this ability of

the plasterboard to provide lateral restraint reduces, due to the calcination of the

gypsum board on the fire exposed side, whereas the plaster board on the ambient side


of the assembly continues to provide lateral support as it is less affected by

temperature.

Sultan (1996) reported that fall-off of plasterboard occurs when the unexposed face of

the board reaches about 600°C (Buchanan and Gerlich, 1997). The temperature at

which the gypsum boards lose their restraining capacity depends on the type of board

used. However, according to Ranby (1999) a common temperature of 550°C was

proposed. In the numerical study of Kaitila (2002), the boundary conditions providing

lateral restraints at both flanges were assumed to be valid until 600C. Thermal

properties of gypsum plasterboard are required to determine the extent to which the

plasterboards offer lateral support to the cold-formed steel studs at elevated

temperatures resisting buckling about the minor axis.

2.4 Literature Review Findings Relevant to this Research

Researchers have attempted to improve the fire ratings of wall systems by using

different types of insulations in the wall cavities. Their observations, however, were

found to be contradictory. Sultan and Lougheed (1994) observed that by use of rock

or cellulose fibre as cavity insulation the fire resistance improved by approximately

30 minutes when compared with uninsulated wall assemblies. Sultan (1995) remarked

that only rock fibre when used as cavity insulation in non-load bearing wall

assemblies gave an improvement in the fire performance whereas assemblies using

cellulose fibre actually showed reduced fire resistance. Feng et al. (2003) observed

that the fire performance of non-load bearing wall panels improved with the use of

cavity insulation.

As other researchers have not been able to conclude the effect on the fire ratings of

wall specimens using cavity insulation, it is considered necessary to conduct further

detailed experimental studies to fully understand the benefits or drawbacks of the

traditional method of wall construction using cavity insulation and to recommend new

wall models to improve the fire performance.

In the research conducted by Kodur et al. (1999) and Alfawakhiri (2001), the studs

had perforations in the web and they were found to be failing at these particular


points. All the frames in their study were cross-braced and it was assumed that the

flexural-torsional buckling and weak axis buckling failure modes were prevented by

this lateral restraint. Gerlich (1995) tested frames with a central row of nogs. Feng et

al. (2005) conducted studies using studs with service holes at the centre of the webs.

The studs were braced laterally by four flat bars fixed horizontally, two on either side

of the framing. No tests were conducted on unperforated studs or frames without

lateral bracing. Therefore further study is necessary to investigate the behaviour of

stud panels without perforations and/or bracing.

Gerlich (1995) study was limited to steel grades of 300 and 450. Kodur et al.’s (1999)

and Alfawakhiri’s (2001) studies involved the use of cold-formed steel studs of yield

strength 230 MPa. Feng et al. (2005) extended their study to 1.2 mm and 2 mm

thickness steels with the grade of S350. Hence it can be concluded that past research

on the behaviour of LSF stud wall panels at elevated temperatures was limited to

lower steel grades. Therefore further research is needed on LSF wall systems made of

higher steel grades.

Uniform temperature values were assumed for flanges and lips of lipped channel studs

on both the hot and cold sides in the studies of Kodur et al. (1999), Alfawakhiri

(2001) and Feng et al. (2005). These observations demonstrate the need to study the

true temperature profiles across and along the studs under fire conditions and to

recommend a simplified temperature profile for use in numerical and theoretical

studies of LSF wall systems.

Most of the numerical models of the past tests were not fully validated due to lack of

experimental results and the complexity of the problem. Also the previously

developed elevated temperature calculation methods are extremely complex. Hence

further study is required to recommend a simplified temperature profile for use in

numerical and theoretical studies and to develop simple models

In the study of Gerlich (1995), too many variables were incorporated in the tests

conducted on only three test specimens. More than one variable between the tests

made the comparison difficult to draw any specific inferences. Also, only one layer of

lining was used in this study. There is very limited data about the effect of multiple



plasterboards, joints in plasterboard as well as the influence of density and thickness

of cavity insulation on the fire performance of the wall assemblies. Therefore a series

of fire tests to establish the fire performance characteristics of the plasterboards, non-

load bearing wall assemblies and load bearing wall assemblies using different types of

insulations of varying density and thickness needs to be undertaken.

In most of the previous experiments it has been noted that the actual Standard time-

temperature heating profile in the furnace could not be controlled satisfactorily. Also,

the instrumentation has not been detailed enough to provide a complete understanding

of the temperature gradient across the thickness and along the height of the wall

specimen. Information regarding furnace pressures is not documented adequately. A

negative furnace pressure in most cases has led to a drop in temperature of the edge

studs as compared to the central ones leading to unequal thermal expansions of the

studs. Also, the commonly observed loading of the studs using a steel beam does not

ensure equal load on all the studs. The problem is further compounded by unequal

expansion of the studs leading to inaccurate collapse load predictions.

To overcome all these problems it was considered necessary to build a special furnace

capable of delivering accurately the required heating regime along with a facility to

monitor and control the furnace pressure, choose appropriate instrumentation to

measure the temperature development across the wall specimens, custom build a

loading frame to enable the application of load to individual studs and maintain a

constant load ensuring free thermal expansion during the test procedure. Also, since

most of the research has been carried out outside Australia, representing specific

materials and method of construction used in those countries and as there has been no

research on LSF load bearing stud walls in Australia, it is important that these

investigations are carried out to assess the behaviour of Australian LSF stud wall

systems and provide recommendations to improve their fire ratings.

Chapter 3: Experimental Work to Determine the Mechanical

Properties of G500 Cold-formed Steel at Elevated Temperatures

3.1: Introduction

The deterioration of the mechanical properties of steel at elevated temperatures is the

primary cause of concern in the design of steel structures exposed to fire. The problem

is even more severe when thin sections made of high grade cold-formed steel are

used.

However, the mechanical properties of cold-formed steel at elevated temperatures

have not yet been fully understood and appropriate design values are not available to

the designers. Research carried out in the past has mostly focused on the reduction in

mechanical properties of hot-rolled steels at elevated temperatures. Consequently the

reduction factors for mechanical properties adopted in various steel design standards

are based on the results of hot-rolled steels.

The reduction factors for the mechanical properties of hot-rolled steels are considered

to be different from those of cold-formed steels. Sidey and Teague (1988) state that

the strength reduction factors of cold-formed steels at elevated temperatures may

differ by as much as 20% than those of hot-rolled steels at corresponding

temperatures. Most steel design standards do not provide the reduction factors for

mechanical properties of cold-formed steels at elevated temperatures except for

Eurocode 3: Part 1.2 (ECS, 2001) and BS 5950: Part 8 (BSI, 1990). However,

Eurocode 3: Part 1.2 (ECS, 2001) adopts the same reduction factors for both cold-

formed and hot-rolled steels at elevated temperatures, while BS 5950: Part 8 gives the

strength reduction factors for cold-formed steels at 0.5%, 1.5% and 2.0% strain levels

over a limited temperature range of 2000C to 6000C. The 0.2% proof stress which is

the most commonly used yield strength value in the designs is not included in BS

5950: Part 8 (BSI, 1990).


Gerlich (1995), Makelainen and Miller (1983), Outinen et al. (2000), Lee et al.(2003),

Chen and Young (2004), and Ranawaka and Mahendran (2009) have investigated the

mechanical properties of cold-formed steels at elevated temperatures. However,

limited data is available for the mechanical properties at elevated temperatures of

cold-formed steels of Grade 500 and thickness 1.15 mm, which is fast gaining

popularity in the construction industry.

Use of high grade cold-formed steel is becoming popular in the construction industry

in Australia. G500 is being increasingly used in the construction of steel stud wall

systems. It is necessary to determine accurately the mechanical properties of cold-

formed steel at elevated temperatures to be able to determine the load carrying

capacity of stud wall systems under fire conditions. The mechanical properties of

cold-formed steels such as the yield strength, elastic modulus, ultimate strength and

ultimate strain at different temperatures can be obtained from the corresponding

stress-strain curves of steels at elevated temperatures. Therefore a series of tensile

tests was conducted to determine the mechanical properties of G500 steel at different

temperatures using steady state conditions.

Tensile testing was preferred over compression testing due to its simplicity, as past

research by Ranawaka and Mahendran (2009) has shown that the mechanical

properties obtained from tension and compression tests show minimal differences. In

steady state tests the specimen is heated up to a specified temperature and then the

tensile test is carried out by controlling either the loading rate or the strain rate.

Alternatively, the mechanical properties of steel at elevated temperatures could also

be determined by using the transient state method, wherein the load on the tensile

specimen is kept constant and then the temperature is raised until the specimen fails.

The transient state tests are considered to be more realistic in predicting the behaviour

under fire conditions than the steady state tests, because, in a real fire scenario, the

structural members are exposed to varying temperature under constant loads in which

the creep effect is also included (Outinen and Makelinen, 2002, Chen and Young,

2007, Lee et al., 2003). The creep effect is time dependent and influenced by both the

applied load and the temperature. As the tests are conducted generally within an hour

the effect of creep may be ignored. In transient test methods the temperature-strain

curves need to be converted into stress-strain curves to obtain the mechanical


properties of steel. In this conversion some approximation in the mechanical

properties of steel cannot be avoided.

In this study the steady state test method under controlled strain rate was preferred

due to the simplicity it offered in developing the stress-strain curves along with

accurate data acquisition.

3.2: Experimental Investigation

3.2.1 Test Specimens

Test specimens were prepared in accordance with the Australian Standard AS 2291

(SA, 1979). They were cut from structural steel sheets in the longitudinal direction

giving test specimens with dimensions as shown in Figure 3.1.

(a) Dimensions

(b) Strain Gauged Test Specimen

Figure 3.1: Tensile Test Specimen

The specimen included two holes at the ends to enable fixing to the loading shafts

located at the top and bottom ends of the furnace.


Specimens were obtained from steel sheets of grade G500 with a thickness of 1.15

mm. The actual cross-sectional dimensions of the coupons were determined using a

micrometer after the removal of the zinc coating from the specimens. The coating was

removed by immersing the specimens in a dilute hydrochloric acid bath. The base

metal thickness (BMT) was then used in the determination of the mechanical

properties. The specimens were tested at ambient and eight elevated temperatures,

giving a total of 9 temperatures. Three specimens were tested per temperature, giving

a total of 27 test results.

3.2.2 Test Rig

An electrical furnace was used for the simulated fire tests to determine the mechanical

properties of G500 steel (see Figure 3.2). Four glow bars positioned at the four

corners of the furnace generated heat. Tensile specimen was located at mid-height of

the furnace equidistant from the four glow bars to ensure uniform heating. A

programmable logic controller was used to regulate the furnace temperature. The

furnace could deliver a maximum temperature of 11000C. Two internal thermocouples

located inside the furnace were used to measure the furnace temperature.

An additional thermocouple kept in contact with the specimen measured the surface

temperature of the specimen. The specimen was mounted between the two loading

shafts made of 253 MA stainless steel for satisfactory operation at elevated

temperatures.

Figure 3.3 shows the details of the test rig and its components. The upper loading

shaft passing through an insulated hole in the roof of the furnace was connected to a

hydraulic actuator of capacity 45 kN. A load cell of capacity one ton was connected as

shown in Figure 3.3 (b) for load determination. The hydraulic actuator was rigidly

connected to a cross-head at the top. The bottom loading shaft passing through an

insulated hole in the base of the furnace was rigidly connected to the floor of the

structural laboratory (see Figure 3.3 (c)). Special care was taken to align the loading

shafts in order to avoid eccentric loading of the test specimen. Figure 3.3 (e) shows

the installation of the test specimen in the furnace.


Tensile load

Hydraulic actuator

Load cell

Top loading shaft Glow bars Thermocouples Specimen Bottom loading shaft

Figure 3.2: Furnace Details

The hydraulic actuator was connected to a Multi-Purpose Test Ware System. The

furnace and specimen temperatures were recorded using an automatic data acquisition

system (EDCAR) at intervals of one minute. Data channels from the load cell, strain

gauges and control device of laser speckle extensometer were connected to EDCAR

to process the data and obtain the stress-strain curves (see Figure 3.4).

The elongation of the specimen was measured in the middle portion of the specimen

using an advanced Laser Speckle Extensometer (LSE) as resistance type strain gauges

are good at only room temperatures and get easily damaged at elevated temperatures.

Also the use of mechanical extensometers to measure the strains at elevated

temperatures was considered unsatisfactory as its accuracy largely depended upon the

precision of the connecting devices. The LSE comprises of a PC based video

processor, laser diodes (class 3A), video cameras, lens and a frame grabber. The


frame grabber digitalizes the analogue video signal and these digitized images are

displayed on the computer screen. The video processor is capable of continuously

measuring the displacements of two speckle patterns developed at a specified distance

apart, by the laser beams and recorded by the video cameras on the specimen gauge

length.

(a) Furnace (b) Loading shaft at the top connected to the actuator

(c) Loading shaft connected to floor (d) LSE mounted on a frame outside the glass window

Figure 3.3: Details of Test Rig and its Components


(e) Specimen installed in furnace (f) LSE being used to measure the extension of test specimen

Camera lens

(g) Main Components of LSE

Figure 3.3: Details of Test Rig and its Components

Laser outlets

Pivot mounted cameras

Data transferring

cable


Figure 3.4 EDCAR (Experimental Data Collection and Recorder)

The displacement of the speckle patterns are converted to a strain signal and sent to an

external control system. The instrument was set up in front of the observation window

built of special heat resistant glass, located on one side of the furnace (see Figure 3.3

(d)). The gauge length of the extensometer is the distance between its two cameras,

which was set at 50 mm. Infrared filters were used in front of the camera lens at

temperatures greater than 5000C for better visibility of the test coupons. Figure 3.3 (f)

shows the instrument in use and Figure 3.3 (g) shows the main components of LSE.

The extensometer was calibrated before testing. The laser beam was located behind

the furnace such that the cameras could be directed onto the specimens gauge length

through the special window made from fire resistant glass. In this method of strain

measurement, two laser beams along with two cameras are oriented targeting the

specimens as shown in Figure 3.5. The upper camera is referred to as the slave and the

lower camera as the master.


Figure 3.5 Strain Measurement using LSE

The measuring principle of the extensometer is based on tracking laser speckle

patterns by the use of two upper and lower cameras named as slave and master,

respectively. The extensometer works on the following principle. When a coherent

laser beam is directed on to an optically rough surface, the light gets diffused in

different directions. If the diffused light rays travel through the original beam, the

light is spatially eliminated, resulting in a granular looking speckle pattern as shown

in Figure 3.6. Each camera records unique speckle patterns relevant to the zones it is

directed. Speckle patterns corresponding to the two zones on the specimen which are

separated from each other by a predetermined distance in the elongation direction are

initially stored as reference speckle patterns. When the tensile load is applied to the

specimen, targeted zones of the two cameras change causing a shift in the speckle

patterns observed by the cameras. The video processor is able to locate the new

position of a stored reference pattern and calculate the distance it has moved between

images. Figure 3.6 (a) shows typical speckle patterns of master and slave cameras

before the test while Figure 3.6 (b) shows typical speckle patterns during the test.

Before the tests, the laser-speckle extensometer was calibrated with a special

calibration method, which enables accurate strain measurements. Since the distance

between initial reference patterns given by the two cameras is adjusted to 50 mm and

is stored in the program before starting the testing, the processor is able to calculate

the strain at any time using Equation 3.1.

50 mm

Slave camera

Laser beams

Specimen

Master camera


o

masterslave

l

dd ………………………………………………….............….... (3.1)

where, - Sum of displacements on Slave Camera slaved

masterd - Sum of displacements on Master Camera

ol - Distance between initial reference patterns

Selected speckle patterns

Output of slave camera Output of

slave camera

(a) Speckle Output before the Test

Distance travelled from master camera

Distance travelled from slave camera

(b) Speckle Output during the Test

Figure 3.6: Typical Speckle Output for Strain Measurements


The LSE was validated by using the strain measurements taken at room temperatures

using strain gauges. Figure 3.7 shows a good correlation between the readings

obtained by the LSE and the conventional strain gauges.

Figure 3.7: Comparison of Stress-Strain Curves using Strain Gauges and Laser Speckle Extensometer

3.2.3 Test Procedure

Tensile tests were carried out based on the steady state test method. The specimens

were heated at a rate of 20-250C per minute from ambient temperature up to the pre-

selected temperature and then loaded up to failure while maintaining the same

temperature. A very small tensile load was maintained in the specimen during the

heating phase in order to eliminate the development of compressive forces in the

specimen due to thermal expansion of the specimen and the connecting shafts. After

reaching the specified temperature, a time of approximately 10 minutes was allowed

to elapse for the temperature to stabilize before the application of load. A constant

displacement rate of approximately 0.2 mm/min was adopted for the loading shaft

during the tests. This was equivalent to a strain rate of 0.0033/minute, which was

within the range of 0.001 to 0.005/minute according to the testing standard for

metallic materials, SFS-EN10002-5 (ECS, 1992).


The tests were carried out on the specimens in the elevated temperature range of

1000C to 8000C at intervals of 1000C.

3.2.4 Test Results

The yield strengths at strain levels of 0.2%, 0.5%, 1.5% and 2.0% were obtained for

the purpose of comparison as these strain levels are accepted widely. The 0.2% yield

strength (f0.2) is the intersected value of the stress-strain curve and the proportional

line, offset by 0.2% strain whereas, the yield strengths of f0.5, f1.5 and f2.0 at the strain

levels of 0.5%, 1.5% and 2.0% respectively are obtained from the intersected values

of the stress–strain curve and the non-proportional vertical lines drawn from the

specified strain levels as shown in Figure 3.8(a).

(b) (a)

Figure 3.8: Determination of (a) Yield strength and (b) Elastic modulus. (Lee et al., 2003)

The yield strength values determined based on 0.2% proof stress were used in

deriving the reduction factors for the yield strengths at elevated temperatures. The

modulus of elasticity was determined from the tangent modulus of the initial elastic

linear part of the stress-strain curve, as shown in Figure 3.8(b).

Yield Strength: Table 3.1 presents the yield strengths at various percentages of strain

(f0.2, f0.5, f1.0, f2.0), ultimate strength (fu), percentage strain at failure (εu) and modulus

of elasticity (E) at ambient and elevated temperatures.


From Table 3.1 and Figure 3.9, it is observed that the mechanical properties are not

much affected up to a temperature of 3000C. Beyond 3000C the drop in mechanical

properties is observed to be very rapid with about 90% of the strength lost at around

6000C. Table 3.1 shows that the ductility values increase with increasing temperature.

The ductility is observed to increase very rapidly for temperatures greater than 4000C.

Table 3.1: Mechanical Properties of 1.15 mm G500 CFS at Ambient and Elevated Temperatures.

Temperature f0.2

(MPa) f0.5

(MPa) f1.5

(MPa) f2.0

(MPa) fu

(MPa) εu

(%) E

(GPa)

27 569.00 583.25 583.50 584.5 589.00 10.70 213.52

100 565.00 578.25 577.83 579.81 590.60 7.10 209.92

200 560.00 570.50 580.20 580.72 600.00 6.40 193.55

300 539.00 556.59 582.17 585.03 606.50 14.35 166.61

400 400.00 408.27 455.13 455.91 457.50 13.10 154.15

500 219.00 221.63 274.25 280.56 283.00 10.03 85.25

600 69.00 70.00 70.02 70.14 74.00 30.00 63.16

700 39.50 40.82 40.85 46.76 45.00 30.00 43.18

800 23.00 23.33 29.18 29.225 28.00 47.00 11.88

The Stress-Strain graphs for 0.2% Proof stress obtained at different temperatures are

shown in Figure 3.9.


Figure 3.9: Stress-Strain Graphs at Different Temperatures

The reduction factors (fy,T/fy,20) determined as the ratios of different yield strengths at

elevated temperatures to that at ambient temperature corresponding to the strain levels

of 0.2%, 0.5%, 1.5% and 2.0% are presented in Table 3.2 along with the reduction

factors (ET/E) determined for the modulus of elasticity at different elevated

temperatures to that at ambient temperature.

Table 3.2: Reduction Factors for Yield Strength and Modulus of Elasticity of 1.15 mm G500 Steel

Temperature (0C)

27 100 200 300 400 500 600 700 800

f0.2,T/f0.2 1.00 1.00 0.98 0.95 0.7 0.39 0.12 0.07 0.04

f0.5,T/f0.5 1.00 0.99 0.98 0.95 0.70 0.38 0.12 0.07 0.04

f1.5,T/f1.5 1.00 0.99 0.99 0.99 0.78 0.47 0.12 0.07 0.05

f2.0,T/f2.0 1.00 0.99 0.99 1.00 0.78 0.48 0.12 0.08 0.05

1.15

mm

G500

ET/E 1.00 0.98 0.91 0.78 0.72 0.4 0.3 0.2 0.06

Figure 3.10 shows the Strength Reduction Factors (more appropriately Strength

Retention Factors) corresponding to various percentages of yield strength. The yield

strength values are seen to drop suddenly from 3000C to 6000C. The drop in strength


values for cold-formed steel is observed to be more rapid than in hot-rolled steel

probably due to the loss in strength initially gained by the cold-forming process.

Figure 3.10: Graph Showing Strength Reduction Factors associated with Various

Percentages of Yield Strength as Obtained from Tests

Note:

fy (0.2%): 0.2% Proof Stress

fy (0.5%): Yield strength corresponding to a total strain of 0.5%



3.3: Comparison of Reduction Factors with Results as Obtained by Other

Researchers and as Recommended by Steel Design Codes

Figures 3.11 and 3.12 show the comparison of test results, with the reduction factors

as obtained by other investigators for 0.2% proof stress and modulus of elasticity,

respectively.


Figure 3.11: Yield Strength Reduction Factors

Figure 3.12: Modulus of Elasticity Reduction Factors

The yield strength as observed by other researchers is seen to drop progressively from

ambient temperature with increase in temperature. The yield strength as obtained by

most researchers is seen to be unconservative beyond 4500C. The reduction in


modulus of elasticity of the test specimens and as observed by other researchers is

seen to observe a progressive reduction in its value as the temperature increases.

Figures 3.13 ((a) to (c)) show the reduction in the yield strength of cold-formed steel

test specimens as compared to that given by BS 5950.8 (BSI, 1990). It is observed

that BS 5950.8 (1990) gives values on the conservative side up to temperature of

around 3250C, beyond which the yield strength of cold-formed steel is seen to fall

rapidly making the values given by BS 5950.8 (1990) more and more unconservative.

Similar observations can be made in Figure 3.14 where the yield strength of cold-

formed steel test specimens corresponding to 0.2% proof stress is compared with the

yield strength of hot-rolled steel as given by AS 4100 (SA, 1998).

(a) Comparison of 0.5% Strength Reduction Factors as Suggested by BS 5950-8 (BSI, 1990) with Test Results



(b) Comparison of 1.5% Strength Reduction Factors as Suggested by BS 5950-8 (BSI, 1990) with Test Results

(c) Comparison of 2.0% Strength Reduction Factors as Suggested by BS 5950-8 (BSI, 1990) with Test Results



Figure 3.14: Comparison of 0.2% Strength Reduction Factors with AS 4100 (SA, 1998) Recommendations

Figure 3.15: Comparison of Modulus of Elasticity Reduction Factors with AS 4100 (SA, 1998) Recommendations

The yield stress ratios determined from the test results compared well with the values

given in AS 4100 (SA, 1998) up to a temperature of 2150C. Beyond this temperature

and up to 3800C the values given by the code were seen to be conservative. However



beyond 3800C the code values were seen to be unconservative right up to 8000C with

a notable difference at 6000C, at which point the code considers a yield strength ratio

of 0.44 (i.e. a 56% reduction in yield strength) for design purposes, whereas the test

results at the same temperature give a yield stress ratio of 0.12 (i.e. a 82% reduction in

yield strength). The strength and stiffness values adopted by AS 4100 (SA, 1998) are

seen to be grossly unconservative in the steel temperature range of 5000C – 7500C as

can be seen in the Figures 3.14 and 3.15, respectively. The modulus of elasticity ratios

of AS 4100 (SA, 1998) are seen to be unconservative from 1400C onwards when

compared with the test results. At 6000C the code adopts a reduction factor of 0.5 (i.e.

a 50% reduction in stiffness) whereas the test results give a value of 0.3 (i.e. a 70%

reduction in stiffness)

As the mechanical properties of high grade cold-formed steel are significantly

different from the values proposed by AS 4100 (SA, 1998), following empirical

equations based up on the tensile coupon tests have been proposed to better represent

the behaviour of high grade cold-formed steels at elevated temperatures. The

developed equations relate the reduction factors of strength and stiffness with respect

to temperature.

Although various strain levels (0.2%, 0.5%, 1.5% and 2.0%) were considered in

determining the yield strength, only the reduction factors based on the 0.2% proof

stress method were used in deriving the empirical equations. This is because the yield

strengths based on other strain levels have not been accepted widely.

Equations 1((a) to (c)) give the yield strength reduction factors with respect to

temperature.

(fyT / fy20 ) = -1.891 X 10-4 T + 1.012 27 < T ≤ 300 Eq. 1(a)

(fyT / fy20 ) = -2.8 X 10-3 T + 1.8 300 < T ≤ 600 Eq. 1(b)

(fyT / fy20 ) = -4 X 10-4 T + 0.356 600< T ≤ 800 Eq.1(c)

Equation 2 gives the reduction factors of modulus of elasticity with respect to

temperature.

ET/E = -5.733 X 10-7 T2 – 8.419 X 10-4 T + 1.0637 100< T ≤ 800 Eq.2

Figures 3.16 and 3.17 show a good agreement between the values predicted from

these equations and the test results.

Figure 3.16: Comparison of Yield Strength Reduction Factors from Test Results and Predictive Equation

Figure 3.17: Comparison of Elastic Modulus Reduction Factors from Test Results and Predictive Equation


Figures 3.18 and 3.19 show the comparison of the proposed equations based on the

experimental work for G500 grade of steel with the generalised equations as given by

Ranawaka and Mahendran (2009).

Figure 3.18: Comparison of Ranawaka and Mahendran’s (2009) Equation and Predictive Equation in the Determination of Yield Strength Reduction Factors

Figure 3.19: Comparison of Ranawaka and Mahendran’s (2009) Equation and Predictive Equation in the Determination of Elastic Modulus Reduction Factors

The yield strength reduction factors as obtained by Ranawaka and Mahendran’s

(2009) equation are slightly unconservative in the temperature range of 3500C to



6000C as compared to the proposed predictive equation. Also the strength reduction as

observed in the experimental work is seen to be linear with increase in temperature

instead of being non-linear as given by Ranawaka and Mahendran’s (2009) equation.

The elastic modulus reduction factors given by the proposed predictive equation are

observed to conform closely to Ranawaka and Mahendran’s (2009) equation.

3.5: Conclusion

To address the lack of reliable mechanical property data for cold-formed steel at

elevated temperature, detailed experimental work was carried out as part of this study

in the Structural laboratory of Queensland University of Technology. Tensile coupon

tests of G500 cold formed steel were undertaken at temperatures ranging from 1000C

to 8000C under steady state conditions to obtain the reduction factors (more

appropriately retention factors) for strength and modulus of elasticity.

The current Australian and European Standards do not present accurate reduction

factors for the yield strength and elasticity modulus of cold-formed steels at elevated

temperatures. These factors are observed to be grossly unconservative at temperatures

beyond 3500C. To provide more accurate mechanical properties and facilitate safer

design of cold-formed steel structures at elevated temperatures, predictive equations

have been developed based on the results from this research for calculating yield

strength and elastic modulus. These equations were seen to compare well with the test

results and conservatively predict the yield strength and elastic modulus at elevated

temperatures as compared to the existing standards.

P.N.Kolarkar: Structural and Thermal Performance of Cold‐formed Steel Stud Wall Systems under Fire Conditions 75

Chapter 4: Thermal Performance of Gypsum Plasterboards and

Composite Panels

4.1: Introduction

Fire resistance of non-load bearing and load bearing wall systems depends to a large

extent on the level of protection offered to the steel frame against fire attack. The

most popular method of providing this protection to the steel frame is by attaching

gypsum plasterboard sheets on either side of the frame. To improve the fire resistance

of the wall, multiple sheets are attached on either side. When the wall is exposed to

fire from one side the plasterboards form the first line of defence by protecting the

steel from the intense heat generated by the fire. A temperature gradient inevitably

develops across the depth of the wall. A substantial drop in temperature occurs across

the thickness of each layer of plasterboard used. With prolonged exposure to fire the

plasterboards calcine and develop cracks allowing the heat to penetrate, eventually

leading to the failure of the wall.

To better understand the thermal performance of the gypsum plasterboards many

experiments were conducted in the Fire Research Laboratory of Queensland

University of Technology. Fire tests were performed on Type X gypsum plasterboards

supplied by the company Boral Plasterboards under the product name FireSTOP.

Thermal performance of single, double and triple layers of plasterboards was studied.

Different types of insulations were also used to help improve the fire performance.

Composite panels were developed with a layer of insulation between two sheets of

plasterboard.

This chapter presents the details of a series of fire tests performed on individual and

multiple layers of plasterboard. It also examines and compares the thermal

performance of composite panels developed from different insulating materials of

varying densities and thicknesses and makes suitable recommendations.


4.2: Test Setup and Procedure

Fire tests were carried out by exposing one face of the specimens to heat in a propane-

fired vertical gas furnace. An adapter was specially designed to fit into the large

furnace so as to isolate a single burner to facilitate the testing of small scale specimens

(see Figures 4-1 and 4-2). The internal dimensions of the adapter were 1290 mm x

1010 mm. The specimen to be fire tested was mounted on a platform extending from

the base of the adapter such that the specimen would enclose the open furnace

chamber. On starting the furnace, the specimen was exposed to heat from one side

only. The furnace temperature was measured using four Type K mineral insulated and

metal sheathed thermocouples, each being placed at the centre of the four quarters

formed by the horizontal and vertical centre lines. Care was taken to ensure that the

distance of the hot junction of all the furnace thermocouples from the fire surface of

the test specimen was about 100 mm. The average temperature rise of these

thermocouples served as the input to the computer controlling the furnace heat

according to the cellulosic fire curve (Standard time-temperature curve) given in AS

1530.4 (SA, 2005), which is similar to ISO 834-1 (1999) and ASTM E119 (1995).

Additionally four more thermocouples were uniformly distributed in the chamber so

as to give a reliable indication of the average temperature of the furnace chamber in

the vicinity of the test specimen. These thermocouples were connected to the data

logger and used for the plotting of the furnace time-temperature graphs.

The adapter was able to utilise the existing exhaust vent built into the left side wall of

the large furnace as shown in Figure 4-1. To establish proper hot air circulation within

the small chamber another exhaust vent was built by drilling a hole into the right side

wall of the adapter. Both the exhausts were fitted with control gates to monitor the

exhaust opening sizes and achieve greater control on the convection currents within

the chamber. This provided a more uniform temperature over the height of the

chamber. The specimens were installed in the furnace as shown in Figure 4-3. A

pressure transducer was also used to measure the pressure inside the furnace chamber

during the fire test (see Figure 4-4). Time-temperature profiles at various locations

across the thickness of the Test Specimens were plotted from the data available to

help assess their fire performance. The tests were stopped once the plasterboard paper

on the ambient side of the specimen started to burn.

Large Furnace

Existing left vent in the large furnace

Right vent built in the adapter

Adapter for fire testing of small scale specimens

Figure 4-1: View Showing Adapter Attached

to Large Furnace for Carrying Out Fire Testingof Small Scale Specimens

Supports for attaching top clamps to hold the specimen in place

Valve to control vent opening

150 mm thick ceramic fibre lining

Slots to attach platform for supporting small scale specimens

Figure 4-2: Adapter Details


Figure 4-3: View Showing Plasterboard Specimen Installed for Fire Testing

Figure 4-4: Pressure Transducer used for Determining Furnace Chamber Pressure during Testing

4.3: Test Specimens

Fire tests were conducted on gypsum plasterboard specimens each measuring 1350

mm x 1080 mm in dimensions. The specimens were built using either single or

multiple plasterboards. Composite panel specimens using different types of

insulations placed between the plasterboards were also built. Insulation densities were

varied to study their effect on the fire performance. K type wire thermocouples were

inserted within the body of the plasterboard by drilling holes so as to measure the

temperature variation at different depths across the thickness of the plasterboard when

exposed to fire from one side. The holes were drilled normal to the plane of the

plasterboard to the required depth at mid-height of the specimen. The hot junction of

the wire thermocouple was then inserted into the hole, and the hole was then sealed

off using moist powdered gypsum plasterboard. Minimum of two holes were drilled

and thermocouples installed to determine the temperature profile at any particular

depth. Fifteen test specimens were built as shown in Table 4-1. The position of

thermocouples is indicated by the coloured dots. Construction details of each

specimen have been discussed after the table.


Table 4-1: Details of Plasterboard and Composite Panel Test Specimens

No. Configuration Specimen Description

1 Pb = 13 mm

2

Pb = 16 mm

3

Pb1 = 13 mm (Fire Side)

Pb2 = 16 mm (Ambient Side)

4

Pb1 = 16 mm (Fire Side)


5 Pb1 = 16 mm (Fire Side)

Pb2 = 16 mm (central)


6, 7,

8, &

9

Pb1 = 16 mm (Fire Side) Insulation: Glass Fibre of varying thickness, density and type. Pb2 = 16 mm (Ambient Side)

10 &

11

Pb1 = 16 mm (Fire Side) Insulation: Rock Fibre of varying thickness, density and type. Pb2 = 16 mm (Ambient Side)

12, 13

& 14

Pb1 = 16 mm (Fire Side) Insulation: Cellulose Fibre of varying thickness, density and type. Pb2 = 16 mm (Ambient Side)

15 Pb1 = 16 mm (Fire Side) Insulation: Isowool. Pb2 = 16 mm (Ambient Side)


To measure the temperature of the ambient surface of the Test Specimens, five

thermocouples were attached to the unexposed surface of the plasterboard, one

thermocouple at the centre of the specimen and one at the centre of each quarter

section of the specimen as shown in Figure 4-5.

Thermocouples

Figure 4-5: Thermocouples on the Ambient Side of the Specimen

Fifteen Test Specimens were built and tested to achieve the following objectives:

1) To study the thermal performance of a single layer of plasterboard including

the temperature distribution across the thickness of the plasterboard.

2) To determine the influence of chemically bound and free water present in the

body of the plasterboard on the ambient side temperature of the plasterboard

specimens.

3) To study the effect of joints between plasterboards on the temperature

distribution across the thickness of the plasterboards.

4) To study the effect of joints between plasterboards on the ambient side

temperature of the plasterboard specimens.

5) To study the effect of multiple boards on the thermal performance of

specimens.

6) To study the effect of various types of insulation with varying density and type

on the thermal performance of the composite boards.


4.3.1: Test Specimen 1

4.3.1 A): Construction Details

This specimen was made of a single layer of 13 mm thick gypsum plasterboard.

Thermocouples were located on the specimen as shown in Figure 4-6. Five additional

thermocouples were attached on the ambient face to measure the temperature of the

unexposed surface of the wall. Thus a total of nine Thermocouples (4+5) were used to

measure the temperature across the specimen.

Fire Side (Exposed Surface)

13 mm Plasterboard

Ambient Side (Unexposed Surface)

Two thermocouples on the fire side at mid height

Two thermocouples at 7 mm from the fire side at mid height

Five thermocouples on the ambient side as shown in Figure 4-5

Figure 4-6: Instrumentation for Test Specimen 1

4.3.1 B): Observations, Results and Discussions

One side of Test Specimen 1 was subjected to the standard time-temperature heating

regime in the furnace (see Figure 4-7 (a)). By the end of 3 minutes smoke was seen to

start coming from the edges of the specimen. This was on account of the burning of

the plasterboard paper on the exposed side. The smoke subsided after the paper got

completely burnt out. By the end of 6 minutes steam was seen to come out from the

specimen and condense on the top front face of the furnace adapter. By the end of 12

to 13 minutes the steam subsided and the specimen soon was seen to burn steadily

without letting out smoke or steam. By the end of 18 minutes the ambient side paper

of the plasterboard started to discolour. The specimen was also seen to bow laterally

outward (see Figure 4-7 (b)). This was probably caused due to the shrinkage of the

surface exposed to furnace following the expulsion of water. By the end of 33 minutes

the outside paper had started to burn and the test was stopped. The specimen was seen

to have lost most of its strength and had developed deep vertical cracks on the

exposed surface, although the cracks were not visible on the ambient surface.



Figure 4-8 shows the time-temperature profile across the plasterboard thickness for

Test Specimen 1. The time-temperature profiles obtained within the thickness of the

specimen at a distance of 7 mm from the exposed surface (fire side) and on the

ambient surface show the development of the temperature in three phases. The first

phase displayed a steady rise in temperature from the ambient temperature to around

1000C. This was followed by the second phase where the temperature was maintained

close to 1000C thus giving a plateau. In this phase the heat energy was primarily used

up to convert the free and chemically bound water in the plasterboard to steam. The 7

mm depth and 13 mm depth profiles have their second phase extending up to

approximately 6 and 12 minutes respectively. The third phase started when moisture

in the plasterboard was no longer available for conversion into steam. The

temperature in the third phase was seen to increase gradually reaching 4500C at 7 mm

depth and 2750C on the ambient surface towards the end of the test.

The test was stopped on account of the paper on the ambient side burning. The

thermocouple at 7 mm depth from the exposed surface did not record any sudden rise

in temperature in the third phase up to the end of the test. This implies that the layer of

plasterboard up to 7 mm depth, though calcinated, was still intact and prevented any

sudden ingress of heat. A temperature difference of approximately 3500C was

observed from the exposed surface to a depth of 7 mm. With a further drop of

approximately 2000C from 7 mm to 13 mm depth giving a temperature of 2750C on

the ambient surface.

Figure 4-9 shows a graph of temperature versus depth plotted at intervals of ten

minutes. The profile at the end of ten minutes shows the curve linear from the

exposed surface up to the 7 mm depth and then gradually flattening out and merging

with 1000C line at a depth of 8.5 mm. This implies that, at the end of 10 minutes of

fire exposure, the specimen still had moisture from 8.5 mm to 13 mm depth. This can

be verified from Figure 4-8 where it can be noticed that, at the end of 10 minutes, the

7 mm depth profile had entered the third phase but the 13 mm depth profile was still

in the second phase. The 20 minute profile in Figure 4-9 shows the temperature at all

the depths above 1000C implying that moisture had been completely driven out of the

entire thickness of the specimen.

With the expulsion of water from across the thickness of the specimen, the

plasterboards became more and more calcinated with the development of several

shrinkage cracks over the surface and within the body of the plasterboard. At this

stage, the graph of temperature vs depth (see Figure 4-9) approaches linearity. The

two profiles, 20 minutes and 30 minutes, in Figure 4-9 obtained after the expulsion of

water from the plasterboard body are seen to run almost parallel to each other (i.e. the

slopes are almost the same) suggesting that the thermal properties of the calcinated

plasterboard does not undergo much change with respect to temperature as long as the

integrity is maintained.

(a) Test Specimen 1 at the start of the test (b) Thermal bowing visible towards the end of the test

Figure 4-7: Fire Testing of Test Specimen 1


0

100

200

300

400

500

600

700

800

900

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 FS 7 mm Amb

Figure 4-8: Time-Temperature Profile of Test Specimen 1

Note:

AS 1530.4: Standard time-temperature curve from AS 1530 Part 4

F.S.: Temperature profile of the exposed surface of the specimen (Fire Side surface)

7 mm: Temperature profile at a depth of 7 mm from the exposed surface

Amb: Temperature profile of the unexposed surface of the specimen

0

100

200

300

400

500

600

700

800

900

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Depth (mm)

Tem

per

atu

re (

oC

)

10 min 20 min 30 min

Figure 4-9: Temperature-Depth Profiles of Test Specimen 1




This specimen was made of a single layer of 16 mm thick gypsum plasterboard.

Thermocouples were located on the specimen as shown in Figure 4-10. A total of 13

thermocouples (8+5) were used to measure the temperature across the specimen.


16 mm thick plasterboard


Two thermocouples on the fire side at mid-height

Two thermocouples at 4 mm from the fire side at mid-height



Five thermocouples on the ambient side as shown in Figure 4-5.



The specimen was fire tested for about 78 minutes. The observations pertaining to the

evolution of smoke and steam were similar to that of Test Specimen 1. Figure 4-11 (a)

shows Test Specimen 2 mounted in the small adapter for the fire test. By the end of 29

minutes, the paper on the ambient surface started to discolour uniformly. By 40

minutes, the ambient surface had become quite dark. Towards the end of the test, the

paper was partially burnt and the specimen had begun to bow laterally in the outward

direction (see Figures 4-11 (b) and (c)).

The plateaus for the 4, 8, 12 and 16 mm depth profiles in Figure 4-12 were seen to

extend approximately up to about 3, 7, 12 and 18 minutes respectively. That is, on

average, 1 min of fire exposure was required to expel water from 1mm thickness of

the plasterboard. Thus, at the end of 10 minutes, a plasterboard thickness of

approximately 10 mm from the exposed surface would have its water expelled. This

means, the plateau of the unexposed surface, in the case of 13 mm plasterboard would


extend up to 13 minutes and for 16 mm plasterboard it would extend approximately

up to 16 minutes.

In Figure 4-12 it can be clearly seen that the 12 mm depth profile has entered the third

phase at the end of 15 minutes whereas the 16 mm depth profile was still in the

second phase implying the presence of moisture in the last few mm thickness of the

plasterboard. The 75 minutes profile in Figure 4-13 is seen to approach linearity. It is

seen to be almost parallel to the 30 minute profile signifying very little change in the

thermal properties of the plasterboard over that duration of time.

(a) Test Specimen 2 at the Start of Test (b) Test Specimen 2 at the End of Test

(c) Thermal Bowing Visible at the End of Test Figure 4-11: Fire Testing of Test Specimen 2


0

100

200

300

400

500

600

700

800

900

1000

1100

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 FS 4 mm 8 mm 12 mm Amb


0

100

200

300

400

500

600

700

800

900

1000

1100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17Depth (mm)

Tem

per

atu

re (

oC

)

15 min 30 min 45 min 60 min 75 min





This specimen consisted of two layers of gypsum plasterboards, one of 13 mm

thickness and the other of 16 mm thickness attached to each other by 30 mm screws

spaced at 300 mm centres along the periphery. The 13 mm thick plasterboard was

labelled as Pb1 and formed the exposed side of the specimen, whereas the 16 mm

thick plasterboard was labelled as Pb2 and formed the ambient side of the specimen.

Thermocouples were located on the specimen as shown in Figure 4-14. Thirteen

thermocouples were used to measure the temperature across the specimen.


13 mm thick Plasterboard (Pb1)

16 mm thick Plasterboard (Pb2)


Two thermocouples on the fire side (No. 2)

Two thermocouples at 7 mm from the fire side and within Pb1 at mid-height

Two thermocouples at 13 mm from the fire side i.e. in the interface of Pb1-Pb2 at mid-height





Test Specimen 3 was exposed to fire for a period of 171 minutes. The fire side paper

of the exposed plasterboard caught fire by the end of 3 minutes when the temperature

of the exposed surface was around 4000C. The smoke was soon followed by steam

which continued for 4 to 5 minutes. This was followed by a period of steady burning

during which time there was hardly any emission of smoke or steam. By the end of 20

minutes, smoke reappeared. This was probably due to the plasterboard paper on the

ambient side of Pb1 (Plasterboard 1) burning. By the end of 62 minutes the



plasterboard on the unexposed surface of the specimen started to discolour, when its

temperature was about 2000C.

Figure 4-15 (a) shows Test Specimen 3 with all its instrumentation at the start of the

test. Towards the end of the test the paper on the ambient surface of the specimen was

noticed to have blackened uniformly (see Figure 4-15 (b)). Horizontal and vertical

folds in the paper indicated that deep cracks in the plasterboard body had reached the

ambient surface with only the paper holding the pieces together. The test was stopped

at this stage. The specimen had undergone lateral deformation in the outward

direction towards the end of the test. The plasterboard pieces of the test specimen

when closely observed showed that the glass fibres used in the making of Type X

gypsum boards were intact only in the layers close to the unexposed surface (see

Figures 4-15 (c) and (d)) of the specimen.

Plasterboard 1 had undergone severe calcination with a large network of shrinkage

cracks. The glass fibres in this plasterboard had completely melted. In spite of the

severe calcination of Plasterboard 1 (13 mm board) compared to Plasterboard 2 (16

mm board) towards the end of the test, the drop in temperature across the thickness of

the specimen, (slope of the 150 minutes profile in Figure 4-17) was seen to be almost

uniform, again suggesting that the thermal properties of the plasterboard do not

undergo much change even when it is severely calcinated.

In Figure 4-16 the plateaus for the 7, 13, 21 and 29 mm depth profiles were seen to

extend up to about 6, 14, 31 and 54 minutes, respectively. As observed from the

previous specimens, Plasterboard 1 had expelled its moisture across the thickness of

13 mm in approximately 13 – 14 minutes. However, Plasterboard 2 showed extended

periods of plateau. This was probably because of the heat that was allowed to escape

from the interface Pb1-Pb2 thus lowering the severity of the fire on Plasterboard 2.

The 13 mm curve in Figure 4-16 shows the time-temperature profile of the interface

(Pb1-Pb2). A temperature drop of approximately 4000C was observed from the

exposed surface to the ambient side of Plasterboard 1 (i.e. Pb1-Pb2 interface) and a

further drop of approximately 5500C up to the unexposed surface of the specimen

towards the end of the test. Both plasterboards remained intact until the end of the

test. The thermocouples recording the fire side temperature displayed a sudden

increment in temperature at around 32 minutes. This has been assumed to be on

account of the malfunctioning of the wire thermocouples at temperatures approaching

9000C.

(a) Test Specimen 3 at the Start of Test (b) Test Specimen 3 at the End of Test

(c) Glass Fibres Intact Close to the (d) Glass Fibres Seen at the Top

Unexposed Surface of the Specimen Right Corner


The joint in the plasterboard is seen to have an influence on the length of the plateau

(second phase) in the time-temperature profile of the ambient surface of the

plasterboard. The plateau is seen to last up to 44 minutes as against the expected 29

minutes which is the combined thickness of the two plasterboards (see Figure 4-16).

The extra 15 minutes is probably on account of the possible recondensation of steam

in the interface.


0100200300400500600700800900

1000110012001300

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 FS 7 mm 13 mm 21 mm Amb


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30Depth (mm)

Tem

per

atu

re (

oC

)



The 7 mm and 13 mm depth temperature profiles of Specimen 3 are seen to display

higher temperatures than the equivalent depth temperature profiles of Specimen 1 at

corresponding times. This is due to the influence of Plasterboard 2 in Specimen 3


which blocks the escape of heat and redirects most of it back onto Plasterboard 1

causing it to heat up faster.



This specimen consisted of two 16 mm thick plasterboards attached to each other by

40 mm screws spaced at 300 mm centres along the periphery. Thermocouples were

positioned to measure the temperature profiles of the exposed surface i.e. the fire side

surface (FS), the interface between the two plasterboards (Pb1-Pb2), the unexposed

surface, i.e. the ambient side (amb), and also within the thickness of the plasterboards

as shown in Figure 4-18.


Pb1 (16 mm thick)

Ambient Side (Unexposed Surface) Pb2 (16 mm thick)



Two thermocouples at 16 mm from the fire side and in the interface of Pb1-Pb2 at mid-height





Test Specimen 4 was subjected to the fire test for approximately 222 minutes. The

behaviour of the specimen was very much similar to that of Test Specimen 3. After

intermittent evolution of smoke and steam, the ambient side of the specimen started to

discolour at the end of 78 minutes. The test was continued for some time even after

the burning of the ambient side paper. The specimen displayed a small amount of


lateral deformation in the outward direction. The test was finally stopped when most

of the ambient side paper started to peel and burn.

Figures 4-19 ((a) and (c)) show Test Specimen at the start and towards the end of the

test. The plateaus for 8, 16, 24 and 32 mm depth profiles in Figure 4-20 were seen to

extend up to 6, 23, 40 and 60 minutes, respectively. Plasterboard 2 showed much

extended periods of plateau. The interface is seen to increase the plateau on the

ambient side by approximately 28 minutes more than the expected duration of 32

minutes. Towards the end of the test a temperature difference of approximately 3200C

was noticed across Plasterboard 1 and approximately 6300C across Plasterboard 2.

Both plasterboards were intact up to the end of the test.

Figure 4-20 shows the temperature-time profiles for various depths across the

specimen cross-section while Figure 4-21 shows the temperature-depth profiles at

specific intervals of time.

As expected the 8 mm and 16 mm depth temperature profiles are seen to display

higher temperatures than the equivalent depth profiles of Test Specimen 2 at

corresponding times due to the heat redirected by the ambient side plasterboard.

(a) Test Specimen 4 at the Start of Test (b) Evolution of Smoke and Steam from Test Specimen 4



Clamps permitting vertical movement of the specimen

Uniform discolouration of paper

K type thermocouple wires

(c) Test Specimen 4 Showing Thermal Bowing at the End of Test


0

100

200300

400

500

600700

800

900

10001100

1200

1300

0 20 40 60 80 100 120 140 160 180 200 220Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 FS 8 mm 16 mm 24 mm Amb



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35

Depth (mm)

Tem

per

atu

re (

oC

)

30 min






This specimen consisted of three plasterboards, each of 16 mm thickness, firmly

attached together using 50 mm long screws spaced at 300 mm centres along the edges.

Thermocouples were located on the surface as well as between the plasterboards to

measure the interface temperature. Eleven thermocouples were used to measure the

temperature profiles across the specimen as shown in the Figure 4-22.


Pb2 (16 mm thick)

Pb1 (16 mm thick)

Ambient Side (Unexposed Surface) Pb3 (16 mm thick)








ing regime for a period of slightly

over 3 hours. The test was stopped when the ambient side plasterboard paper started

temperature reaching 9000C

by the end of 155 minutes of test. At about 165 minutes Plasterboard 1 must have

Test Specimen 5 was subjected to the furnace heat

to burn. Figure 4-23 shows the time-temperature profiles across the specimen

thickness at various depths. The 16, 32 and 48 mm depth temperature profiles show

their second phases extending up to 23, 62 and 120 minutes, respectively. The

interfaces are seen to increase the plateau on the ambient side by approximately 64

minutes more than the expected duration of 48 minutes.

Plasterboard 1 was seen to heat up quite rapidly with its

partially or fully collapsed as the curve is seen to rise rapidly and merge with the fire

side (FS) curve. At the end of the test, the temperature across the Pb2-Pb3 interface

had reached 7500C and the unexposed surface had crossed 2000C.

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 FS 16 mm 32 mm Amb


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50Depth (mm)

Tem

per

atu

re (

oC

)

30 min 60 min 90 min 120 min 150 min 180 min


Figure 4-24 shows the temperature-depth profiles at different time intervals. With the

passage of time the curves are seen to become more and more linear up to 120

minutes. Beyond 150 minutes Plasterboard 1 starts to deteriorate very rapidly forcing

the portion of graph between 0 to 16 mm (representing Pb1) to become horizontal. At

180 minutes the initial portion of graph from 0 to 16 mm was horizontal, signifying

that Pb1 had collapsed and was no longer effective. The temperature drop from 16

mm to 32 mm and from 32 mm to 48 mm signifies the presence of Pb2 and Pb3 until

the end of the test.

The advantage of three layers of plasterboard over two layers is observed only during

the initial two hours of the test due to the extended plateau of the temperature profile

on the ambient surface of the specimen. After two hours the advantage starts reducing

rapidly and at around three hours both specimens are equivalent and display similar

thermal performance.



4.3.6: Test Specimens 6, 7, 8 and 9


Test Specimen 6 consisted of a composite panel formed by sandwiching a layer of

insulation between the plasterboards. The insulation was laid in the cavity formed on

Pb1 by bordering it with plasterboard strips along the edges. This was achieved by

first fixing 16 mm plasterboard strips of 50 mm width along the periphery of the

board to form the cavity. The desired depth for the cavity was obtained by choosing

the appropriate thickness and number of plasterboard strips to be used along the

border. In the making of this test specimen a cavity of 32 mm was decided as it could

be easily provided using two 16 mm strips mounted on each other. The strips were

then fixed to the exposed plasterboard (Pb1) using 50 mm screws spaced at 300 mm

centres along the centreline of the frame.

Glass fibre insulation in the form of a mat was then cut to appropriate dimensions and

laid inside the cavity (see Figures 4-25 (a) and (b)). The mat used was 50 mm in

thickness and of 13.88 kg/m3 density. The mat was then compressed to a thickness of

32 mm (depth of the cavity) by firmly pressing it down with the help of a second layer

of 16 mm thick plasterboard (Pb2) which formed the ambient side of the test

specimen. This plasterboard was then screwed to the frame by 70 mm long screws

running along the periphery at 300 mm centres and staggered with the screws initially

applied to form the frame. The compressing of the 50 mm thick glass fibre mat to 32

mm thickness increased its density from 13.88 kg/m3 to 21.68 kg/m3 (ρ2 = ρ1 x t1/t2 =

13.88 x 50/32 = 21.68 kg/m3)

Thermocouples were installed during the construction process at various locations

across the thickness of the test specimen as shown in Figure 4-25. The wires were

carefully installed in the interfaces formed between plasterboard and insulation so as

to obtain the temperature profile on either side of the insulation. This would help in

understanding the behaviour of the insulation and its effectiveness at high

temperatures during the fire test. Eleven thermocouples were used to study the fire

performance of the test specimen.

(a) 50 mm Thick Glass Fibre Mats (b) Laying of Single Layer Glass Fibre Mat

Figure 4-25: Construction of Test Specimen 6



Insulation

16 mm Plasterboard (Pb2)

16 mm Plasterboard (Pb1)


Two thermocouples at the interface of Pb1-Ins at mid-height

Two thermocouples at the interface of Ins-Pb2 at mid-height


Figure 4-26: Instrumentation for Test Specimens from 6 to 15

Test Specimen 7 was built in a manner similar to Test Specimen 6. However two

glass fibre mats each of 50 mm in thickness were laid in a cavity of depth 32 mm. The

glass fibre mats were held pressed down to a thickness of 32 mm by the use of

washers held down by screws passing through the base plasterboard (see Figure 4-27).

The compressed glass fibre mat was then further compressed by fixing the second

layer of plasterboard in a manner similar to the previous specimen. The compressing

of the glass fibre mats from a combined thickness of 100 mm to 32 mm increased its

density from 13.88 kg/m3 to 43.4 kg/m3. This test specimen was built to study the

effect of insulation density on the fire performance of the panel. Thermocouples were

installed on the interfaces and on the plasterboard surfaces to measure the temperature


profiles across the panel during the fire test as in Test Specimen 6. Eleven

thermocouples were used to study the fire performance of the panel.


(a) Laying of Two Layers of 50 mm thick Glass Fibre Mats

(b) Glass Fibre Mats Compressed to 32 mm Thickness

Test Specimen 8 was built in a manner similar to Test Specimens 6 and 7. In this

specimen semi-rigid glass fibre mat of density 37 kg/m3 was used as insulation

material. The mat was 25 mm in thickness and easier to cut and lay. The cavity for

laying this insulation was formed by using two strips of 13 mm thick plasterboard as

border along the periphery giving a total depth of 26 mm. The construction of the

border frame was identical to that in Specimens 6 and 7. The glass fibre mat was then

cut to fit into the cavity as shown in Figure 4-28. This was followed by the fixing of

the ambient side plasterboard (also 16 mm in thickness as the exposed plasterboard).

Thermocouples were fixed in the same manner as Test Specimen 6 to record the

temperature variation across the body of the panel. Eleven thermocouples were used

to study the fire performance of the panel.

Test Specimen 9 was built similar to Test Specimens 6 to 8 with the only variation of

using a different thickness of glass fibre mat. This specimen was built using glass

fibre board of thickness 13 mm and density 168 kg/m3. The cavity was formed using a

single layer of 13 mm thick plasterboard strip to form the frame border with the

construction being similar to the previous specimens (see Figure 4-29). After the mat

was cut and placed inside the cavity, it was covered by the ambient side plasterboard

of 16 mm thickness and fixed to the frame to form the panel. Eleven thermocouples

were used to determine the temperature gradient across the panel. The instrumentation

was identical to that of Test Specimen 6.


(a) Laying of Single Layer of 25 mm (b) Border Frame of two 13 mm Thick Thick Semi-Rigid Glass Fibre Mat Plasterboard Strips to facilitate the

attachment of the Ambient Side Plasterboard


Washers to hold the insulation in place

13 mm thick Glass fibre mat

Plasterboard strip of thickness 13 mm and width 50 mm



Test Specimens 6, 7, 8 and 9 were exposed to the standard time-temperature heating

regime for slightly over three hours. The initial behaviour of all the specimens was

similar to the previously tested specimens. All the specimens displayed a small

amount of thermal bowing in the outward direction towards the end of the test. The

ambient surface of these specimens showed uniform discolouration after about 110

minutes of testing (see Figures 4-30 (a), 4-36 (a) and 4-39 (b)). The tests were stopped



when the paper on the ambient surface started to burn. When the specimens were

inspected after the test, it was noted that the glass fibre insulation in all the specimens

had been almost completely consumed by heat with only small amounts still visible

along the edges of the specimens.

The time-temperature graphs show that the interface Pb1-Ins in all the specimens

showed a very rapid rise in temperature in the third phase crossing 6000C at about 35

minutes from the start of the test. The temperature profile of the interface (Pb1-Ins)

tended to become horizontal when its temperature approached 7000C. The glass fibre

insulation at this temperature began to disintegrate and lose its insulating properties as

could be seen from the temperature-depth profiles (see Figures 4-32, 4-35, 4-38 and 4-

41). The central portions of the graphs representing the insulation tended to become

horizontal from 90 minutes onwards for Test Specimens 6, 7 and 8 and 109 minutes

for Test Specimen 9 indicating that the insulation was no longer capable of bringing

about a temperature drop across its thickness.

As the heat energy was probably used up for disintegrating the glass fibre insulation,

the temperature on the ambient side of Pb1 (i.e. Pb1-Ins) did not rise. Also less heat

was getting redirected due to the continuous loss of insulation. These factors kept the

temperature on the ambient side of Pb1 steady (under 7000C) almost up to the end of

the test. The temperatures of the two interfaces Pb1-Ins and Ins-Pb2 are seen to merge

together soon after the disintegration of the glass fibre insulation due to direct

transmission of heat by radiation.

Regardless of insulation thickness and density, it was seen that the glass fibre

insulation became ineffective at about 7000C making the composite panels follow

similar temperature-time profiles up to the end of the test. In all these specimens Pb1

and Pb2 were found to remain intact until the end of the test.

(b) Glass Fibre Insulation Consumed by Heat

(a) Uniform Discolouration of Paper on Ambient Side


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 FS Pb1-Ins Ins-Pb2 Amb


Note:

Pb1-Ins: Temperature profile of the interface between Pb1 (exposed plasterboard) and the insulation

Ins-Pb2: Temperature profile of the interface between the insulation and Pb2 (unexposed plasterboard)


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70Depth (mm)

Te

mp

era

ture

(oC

)



Figure 4-33: Test Specimen 7 Installed

in the Furnace for Testing


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180Time (min)

Tem

per

atu

re (

oC

)



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70Depth (mm)

Te

mp

era

ture

(oC

)




(a) Uniform Discolouration of Paper (b) Glass Fibre Insulation is Totally

on Ambient Side Consumed Leaving a Dark Stain Behind


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55 60Depth (mm)

Te

mp

era

ture

(oC

)

30 min 60 min 90 min 120 min 150 min 180


(a) Test Specimen 9 Installed (b) Uniform Discolouration of Paper in the Furnace on the ambient Side for Testing of the Specimen



0

100200

300

400

500600

700

800

9001000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Te

mp

era

ture

(oC

)



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50Depth (mm)

Tem

per

atu

re (

oC

)


Figure 4- 41: Temperature-Depth Profiles of Test Specimen 9

The temperature development on the ambient side of the insulation in Test Specimens

6, 7, 8 and 9 has been shown in Table 4-2 over 10 minute intervals. Table 4-2 clearly

shows that the thickness, number of layers or the density of glass fibre insulation does

not significantly affect the temperature development of the Ins-Pb2 interface. In the



initial stages the Insulation in Test Specimen 9 was seen to perform slightly better

than the insulations in the other test specimens. However, this advantage was lost by

the time the fire side temperature of the insulation reached 7000C.

Table 4-2: Time–Temperature profile of the ambient side of the insulation (Ins-Pb2 interface) in Test Specimens 6, 7, 8 and 9 using glass fibre as insulation

material

Time

(min)

Test Specimen 6

T = 32 mm

P = 21.7 kg/m3

Test Specimen 7

T = 32 mm

P = 43.4 kg/m3

Test Specimen 8

T = 25 mm

P = 37 kg/m3

Test Specimen 9

T = 13 mm

P = 168 kg/m3

30 220 190 180 120

40 300 250 240 200

50 350 300 300 250

60 390 320 350 300

70 450 380 410 370

80 520 470 500 430

90 580 540 560 500

100 600 600 660 600

110 620 600 660 700

120 650 600 670 700

130 660 640 660 680

140 680 680 670 670

150 690 690 690 670

160 700 700 700 700

170 710 700 705 720

180 700 720 740

The temperatures on the ambient side of the insulation in all the test specimens (6 to

9) were seen to be in close comparison after the exposed surface of the insulation

crossed 7000C. Hence, for all practical purposes the thermal performance of the glass

fibre insulated composite panels can be assumed to remain unchanged regardless of

the thickness or density of the insulations used. However, the use of semi-rigid glass

fibre mats is recommended as the construction of Test Specimens 8 and 9 was much

easier than Test Specimens 6 and 7 due to the ease of handling of the insulation mats.

4.3.7: Test Specimens 10 and 11

4.3.7A): Construction Details

Test Specimen 10 was built using rock wool insulation of density 100 kg/m3 and

thickness 25 mm. The construction and instrumentation of this specimen was identical

to that of Test Specimen 8. Test Specimen 11 was built using 13 mm thick rock wool

insulation of density 114 kg/m3 (Figure 4.42). The construction and instrumentation

of this specimen was similar to Test Specimen 9.

Rockwool strips of width 225 mm and thickness 13 mm



Both test specimens (10 and 11) were subjected to the fire test for nearly three hours.

Slight thermal bowing in the outward direction was noted in both specimens towards

the end of the test. Figures 4-44 and 4-46 show the time-temperature profiles across

the specimens at various depths.


In Test Specimen 10 (Figure 4-44), the Pb1-Ins profile was seen to rise rapidly in the

third phase crossing 6000C by the end of 32 minutes, beyond which it flattened out,

with the temperature gradually increasing to 9000C by the end of 147 minutes.

Around this time Plasterboard 1 must have collapsed as the curve jumps rapidly to

merge with the FS curve. The Ins-Pb2 curve is seen to rise gradually up to 6000C by

the end of 147 minutes at which time the temperature is seen to rise sharply on

account of the collapse of Plasterboard 1. The profile of the interface Ins-Pb2

continued to maintain a temperature difference of over 2500C with the FS curve even

beyond 147 minutes implying that the insulation was still intact and functional. The

150 mm depth temperature profile in Figure 4-45 shows a horizontal segment from

exposed surface to 16 mm depth signifying the collapse of Plasterboard 1. Beyond 16

mm and up to 41 mm depth the temperature drop is brought about by the 25 mm layer

of insulation and beyond 41 mm up to 56 mm the drop in temperature is on account of

Plasterboard 2.

(a) Test Specimen 11 at the Start of Test (b) Thermal Bowing at the End of Test

(c) Rock Wool Insulation Intact even after the Fire Test



In the case of Test Specimen 11 (Figure 4-46), the Pb1-Ins graph crossed 6000C by

about 45 minutes and then increased gradually to reach 9000C at about 174 minutes.

The graph was seen to rise sharply at this stage merging with the curve signifying the

collapse of Pb1. The graph of Ins-Pb2 maintained a profile well below that of Pb1-Ins

giving a minimum temperature difference of 2000C across the thickness of the

insulation up to the end of the test signifying the presence of the insulation until the

end of the test. This can be verified from the temperature-depth graph in Figure 4-47

where the 180 minute profile shows a horizontal segment from exposed surface to 16

mm depth signifying the collapse of Plasterboard 1. Beyond 16 mm and up to 29 mm

depth the temperature drop is brought about by the 13 mm layer of insulation and

beyond 29 mm up to 45 mm the drop in temperature is on account of Plasterboard 2.

Contrary to glass fibre insulation, the rock wool insulation showed greater resistance

to disintegration. The physical presence of the insulation was blocking and redirecting

the heat flow back to Plasterboard 1. This resulted in the rising of the temperature of

Pb1-Ins to values beyond 7000C and steadily kept rising up to 9000C when Pb1 started

to breach. Even after getting directly exposed to fire after the collapse of Plasterboard

1, the insulation remained intact and continued to offer protection to Plasterboard 2.

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180Time (min)

Te

mp

era

ture

(oC

)




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55 60Depth (mm)

Te

mp

era

ture

(oC

)



0100

200300400

500600700800

90010001100

12001300

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)




0100

200300

400500600

700800900

10001100

12001300

0 5 10 15 20 25 30 35 40 45 50Depth (mm)

Te

mp

era

ture

(oC

)



Table 4-3 shows the temperature profile of the ambient side of the insulation in Test

Specimens 10 and 11. The thermal performance of both insulations is seen to be

nearly the same in spite of the thicknesses being different. The temperatures were

seen to differ only after 145 minutes from the start of the test after the collapse of the

external plasterboard (Pb1) in Test Specimen 10. Temperatures shown in red indicate

the absence of Plasterboard 1 at that time.



Table 4-3: Time–Temperature Profile of the Ambient Side of the Insulation (Ins-Pb2 interface) in Test Specimens 10 and 11 using Rock Fibre as Insulation

Material

Time (min)

Test Specimen 10 T = 25 mm

P = 100 kg/m3


P = 114 kg/m3

30 160 180

40 240 230

50 290 300

60 310 340

70 330 400

80 390 480

90 450 520

100 500 520

110 520 540

120 540 550

130 550 580

140 580 590

150 700 600

160 850 605

170 880 650

180 860

4.3.8: Test Specimens 12, 13 and 14


Test Specimens 12, 13 and 14 were built using cellulose fibre as insulation. The

insulation was wet sprayed onto the plasterboards using a special wall nozzle (see

Figure 4-48). Insulation thickness and density were varied to study their effect on the

fire performance of the composite panels. Test Specimens 12, 13 and 14 were built

with an insulation thickness of 32 mm (density = 102 kg/m3), 25 mm (density = 108

kg/m3) and 20 mm (density = 131 kg/m3), respectively. Eleven thermocouples (6+5)

were used to study the temperature gradient across each composite panel.

Wall Nozzle

Wet cellulose spray

Figure 4-48: Construction of Test Specimens 12, 13 and 14


The tests for the cellulose fibre composite panels lasted slightly over two hours when

the tests were stopped following the burning of the paper on the ambient side of the

composite specimens. All the specimens displayed outward thermal bowing at the end

of the test (see Figure 4-49(d)). The paper on the ambient side started to discolour

after about 100 minutes. The discolouration in all three specimens was observed to be

non-uniform (see Figure 4-49 (b)) indicating the burning of cellulose fibre within the

specimen in certain areas creating pockets of high temperature. This allowed the heat



to penetrate the insulation compromising the integrity of the composite panel. The test

was stopped soon after the burning of the ambient side paper (see Figure 4-49(c)).

Closer inspection after the fire test revealed that in all the test specimens Plasterboard

1 had collapsed fully or partially and the cellulose fibre had completely burnt out at

the end of the test leaving behind traces of ash sticking to the fire side of Plasterboard

2 (see Figure 4-49(e)).

Figures 4-50 to 4-53 show the temperature profiles across Test Specimens 12 and 13,

respectively. The profiles of both the composite panels were observed to be almost

identical with the plateaus of the Pb1-Ins interface extending up to 18 minutes and

crossing 6000C at about 35 minutes. By the end of 120 minutes the ambient side

temperature of Plasterboard 1 (Pb1-Ins) in both the specimens had reached

approximately 9000C. The third phase of the Ins-Pb2 profile in both test specimens

started at about 36 minutes and by the end of 120 minutes of fire test a temperature

difference of approximately 2000C was recorded by the thermocouples across the

thickness of the insulation (i.e. the difference in the interface temperatures of Pb1-Ins

and Ins-Pb2) indicating the presence of insulation. The sudden increase in temperature

of the Pb1-Ins interface at 125 minutes for Test Specimen 12 and at 119 minutes for

Test Specimen 13 indicate the breaching of the exposed plasterboard. This was soon

followed by the burning of the paper on the ambient side of the composite panel and

the test was terminated.

The temperature profile of the ambient side as seen in both test specimens show the

plateau extending up to 90 minutes. Beyond 90 minutes the temperature (the average

of the five thermocouples mounted on the ambient side on the composite panel)

however was seen to climb quickly crossing 2000C by 120 minutes and 112 minutes

for Test Specimens 12 and 13, respectively. This rapid rise in temperature was

probably on account of the insulation burn out in certain areas creating pockets of

high temperature and allowing the heat to penetrate the composite panel (see Figures

4-50 and 4-52).

Figures 4-51 and 4-53 show the temperature-depth graphs for Test Specimens 12 and

13, respectively. The graphs in both test specimens are almost linear at 90 minutes.

Beyond this time deterioration in the insulation can be noted in Test Specimen 12 as


the profile tends to become horizontal in the central portion, although a drop of 2000C

across the insulation thickness can still be noted at 120 minutes. However in Test

Specimen 13, beyond 90 minutes, the deterioration of Plasterboard 1 seems to have

started along with the disintegration of the insulation as the initial portion of the

profile from 0 mm to 16 mm (representing Pb1) at 120 minutes has become almost

horizontal suggesting the cracking of the exposed plasterboard, whereas a temperature

drop of about 3000C can still be seen from 16 mm to 41 mm in the profile

(representing the 25 mm thick insulation) signifying the physical presence of the

insulation.

Test Specimen 14 was seen to deteriorate more rapidly when compared to Test

Specimens 12 and 13. Figure 4-54 shows the temperature profile across the composite

panel. The plateau for the Pb1-Ins interface was seen to extend up to 21 minutes

beyond which the temperature was seen to increase rapidly crossing 6000C by 40

minutes. Beyond 6000C the temperature rise was gradual reaching 8000C by about

100 minutes at which time the plasterboard appeared to have partially collapsed as the

temperature of the interface (Pb1-Ins) increased sharply merging with the fire side

curve at 120 minutes from the start.

The 60 minute profile in Figure 4-55 is almost linear beyond which the central portion

of the graph is seen to gradually flatten out indicating the disintegration of the

insulation in the composite panel. At 120 minutes the profile is seen to be horizontal

from 0 mm to 36 mm indicating the collapse of the exposed plasterboard and the

complete burnout of the insulation.

The temperature profile on the ambient side of the composite panel was seen to have

its plateau extending up to 80 minutes beyond which it started to increase quickly

crossing 2000C by about 112 minutes. The test was terminated following the burning

of the ambient side paper.

(a) Test Specimen 11 at the Start of Test (b) Non-Uniform Discolouration of Paper

on the Ambient Side

(c) Burning of Ambient Side Paper



(d) Thermal Bowing of Specimen (e) View Showing Collapse of Pb1 and

Disintegration of Cellulose Fibre Insulation

(f) Cellulose Fibre Sample Before and After the Fire Test



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

per

atu

re (

oC

)



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70Depth (mm)

Tem

per

atu

re (

oC

)

30 min 60 min 90 min 120 min



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

per

atu

re (

oC

)



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55 60

Depth (mm)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130

Time (min)

Tem

per

atu

re (

oC

)



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55

Depth (mm)

Tem

per

atu

re (

oC

)



In the case of Test Specimens 12 and 13 the 120 minute profile showed a fall in

temperature across the entire thickness of the composite panel signifying that both the

plasterboards and the insulation were still intact, whereas in the case of Test Specimen

14 the 120 minute profile was horizontal from 0 mm to 36 mm indicating the collapse



of the exposed plasterboard (Pb1) and the complete disintegration of the cellulose

fibre.

The temperature development on the ambient side of the insulation in Test Specimens

12, 13 and 14 is shown in Table 4-4

Table 4-4: Time–Temperature profile of the ambient side of the insulation (Ins-Pb2 interface) in Test Specimens 12, 13 and 14 using cellulose fibre as insulation material.

Time (min)


P = 102 kg/m3


P = 108 kg/m3


P = 131 kg/m3

30 90 100 90

40 150 120 140

50 230 205 250

60 300 300 340

70 350 340 450

80 400 340 530

90 460 400 600

100 510 450 650

110 600 520 880

120 670 620 1000

The thermal performance of these specimens has been seen to differ with the change

in thickness and density of insulation, unlike the specimens built using glass fibre and

rock fibre. It is assumed that the influence of density is more dominant than the

influence of thickness in the case of cellulose fibres. As there is less control in

maintaining the density of the insulation layer over the entire interface it is likely that

certain areas burn faster than others leading to the formation of hot pockets leading to

an early insulation failure of the specimen.



Test Specimen 15 was built using 25 mm thick Isowool insulation. The purpose of

using this high quality insulation, which is used as lining material in the interior of the

furnace, was to only compare the performance of other insulations in relation to

Isowool insulation. The construction and instrumentation methods were similar to

Test Specimen 8.



Test Specimen 15 was subjected to the fire test for slightly over three hours. The

specimen displayed slight thermal bowing towards the end of the test. The ambient

side of the composite panel was seen to discolour uniformly from about 140 minutes

(see Figure 4-57).

Figure 4-58 shows the temperature profiles at various depths across the width of the

composite panel. The interface temperature (Pb1-Ins) had its plateau extending up to

21 minutes beyond which it was seen to increase sharply crossing 6000C at 35

minutes and reaching 7000C at 45 minutes. Beyond this point the temperature was

almost constant up to 80 minutes and then increased very gradually reaching 9000C at

160 minutes. The exposed plasterboard (Pb1) must have breached at this time as the

temperature of the interface (Pb1-Ins) increased suddenly merging with the fire side


curve. The temperature drop across the thickness of the insulation at this time was

about 4000C indicating that the insulation had maintained its integrity until the end of

the test. The ambient side temperature was seen to have its plateau extending up to

120 minutes beyond which it increased gradually crossing 2000C at about 170

minutes. The test was stopped following the burning of the ambient side paper.

(a) Test Specimen 15 at the Start of Test (b) Uniform Discolouration of Paper on the Ambient Side


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190Time (min)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 5 10 15 20 25 30 35 40 45 50 55

Depth (mm)

Te

mp

era

ture

(oC

)



Figure 4-59 shows the temperature-depth profile of the specimen at intervals of 30

minutes. The 150 minute profile is seen to be linear whereas the 180 minute profile is

seen to have its initial portion of the curve (from 0 mm to 16 mm) horizontal

indicating the collapse of the exposed plasterboard (Pb1). The temperature of the

interface (Ins-Pb2) and the ambient side temperature of the test specimen are seen to

rise rapidly soon after the collapse of the external plasterboard.

Figure 4-60 shows the interface temperature (Ins-Pb2) of Test Specimen 15 in

comparison with that of other specimens using glass fibre, rock fibre and cellulose

fibre as the insulation material. Isowool insulation being capable of withstanding very

high temperatures is seen to perform better than all other types of insulation in the

initial stages of the test. However, the rising temperature of the interface (Pb1-Ins) on

account of the redirected heat forced the exposed plasterboard to heat up rapidly

leading to its collapse. The advantage gained by the use of superior insulation was lost

soon after the collapse of the external plasterboard (Pb1).


0

100

200

300

400

500

600

700

800

900

1000

1100

0 20 40 60 80 100 120 140 160 180

Time (min)

Te

mp

era

ture

(oC

)

Sp. 6 Sp. 7 Sp. 8 Sp. 9 Sp. 10

Sp. 11 Sp. 12 Sp. 13 Sp. 14 Sp. 15

Figure 4-60: Time-Temperature profiles for interface Ins-Pb2 of Test Specimens

6 to 15

4.4: Conclusions

This chapter has described the details of 15 fire tests on the thermal performance

of Plasterboards and Insulations. Following is a list of the main findings.

1) The time of exposure to the cellulosic fire curve determines the approximate depth

up to which the free and chemically bound water present in the gypsum plasterboard

gets expelled. On an average, 1 minute of fire exposure is required to expel water

from 1 mm thickness of plasterboard. Hence in the case of 13 mm thick plasterboard

exposed to standard time-temperature curve from one side, the temperature on the

ambient surface would be maintained at about 1000C up to 13 minutes and in the case

of 16 mm plasterboard it would be maintained for up to 16 minutes.

2) In spite of the numerous shrinkage cracks which are seen to develop over the

surface and within the body of the plasterboard due to the expulsion of free and

chemically bound water, the thermal gradient across the thickness of the plasterboard

when exposed to the standard time-temperature heating regime is seen to be

unaffected by the period of fire exposure and rising plasterboard temperature up to

about 900oC beyond which the plasterboard is seen to lose integrity probably because



the cracks get interconnected and reach up to the ambient surface allowing the

passage of heat.

3) Use of multiple boards introduces interfaces between adjacent plasterboards. These

interfaces between plasterboards are seen to help in improving the fire performance of

the wall system i.e. two plasterboards would give a better thermal performance than a

single plasterboard of double thickness. The interface leads to an improved thermal

performance by extending the duration of the plateau (second phase) of the

temperature profile on the ambient side of the specimen. The interface, however, does

not influence the linear variation of temperature across the specimen thickness after

the water is completely driven out from the plasterboards as transmission of heat

across the joint is very rapid on account of radiation.

4) Test Specimen 5 built using three layers of plasterboard is seen to perform better

than the Test Specimen 4 built using two layers only up to two hours from the start of

the test. Beyond two hours the external plasterboard of the triple layered specimen fell

off probably because of the bending of the heat softened screws under the dead weight

of the plasterboards making it equivalent to a double layered specimen. Hence it

appears that better fixing methods are needed if multiple layers (more than two) of

plasterboards are to be used.

5) Glass fibre insulation is seen to disintegrate at about 700oC. This resulted in similar

time-temperature profiles for all the composite panels using glass fibre insulation of

varying thickness and density.

6) In the case of Test Specimens 6, 7, 8 and 9 both plasterboards Pb1 and Pb2 were

found to be intact until the end of the test. This was probably because the temperature

of the Pb1-Ins interface did not climb beyond 7500C on account of the disintegration

of the glass fibre insulation from 7000C onwards. In the case of Test Specimens 10 to

15 the external plasterboard Pb1 fell off on account of the continuous build up of the

Pb1-Ins interface temperature reaching 9000C on account of the redirected heat from

the longer lasting insulation. Following the collapse of the external plasterboard Pb1

the temperature on the ambient side of the insulation and subsequently on the ambient

side of the test specimen was observed to build up rapidly.

7) Compared to Test Specimens using double plasterboards, the composite panels

using glass fibre insulation are seen to perform better as can be seen in Figure 4-61.

The average Ins-Pb2 interface temperature is used for the plotting of the temperature

profile for the specimens using glass fibre insulation (Test Specimens 6 to 9)

Figure 4-61: Average Time-Temperature profile for interface Ins-Pb2 of Test

Specimens 6 to 9 compared with Time-Temperature profile of Pb1-Pb2 interface

temperature of Test Specimen 4

8) Composite panels made of rock fibre insulation of varying density and thickness

also did not display any appreciable difference in their thermal performances although

the insulation lasted nearly up to the end of the test. Rockwool insulation showed

much greater resistance to disintegration when compared with glass fibre and

cellulose fibre insulation.

9) Fire resistance of cellulose insulation appears to depend upon its density. The

higher density cellulose fibre in Test Specimen 14 was totally burnt, whereas the

lower density cellulose fibre in Test Specimens 12 and 13 lasted longer time. The

distribution of cellulose fibre within a specimen could also be non-uniform as the

technique of spraying of wet cellulose fibre onto the plasterboard is not a standardized

process and can lead to areas of varying density within a specimen. This was probably

the cause of the burning of cellulose fibre within the individual specimens at different

rates giving rise to pockets of high temperature and thus lowering the integrity.


Chapter 5: Thermal Performance of Non-Load Bearing Wall

Systems

5.1 Introduction

Fire safety of light gauge cold-formed steel frame (LSF) stud wall systems is critical

to the building design as their use has become increasingly popular in all areas of

construction throughout Australia. Partition wall panels composed of a cold-formed

steel frame lined with one or two plasterboards as side sheathing are being widely

used as they are very easy to assemble, thus improving the speed of construction. In

Australia, plasterboard lining manufacturers provide fire resistance ratings of non-

load bearing LSF stud wall systems. They have prescribed steel stud walls with single

or multiple plasterboard linings achieving fire resistance ratings, ranging from 60 to

120 minutes. These systems are based on full-scale fire resistance tests using the

standard fire curve recommended by ISO 834 and AS 1530.4 (SA, 2005). With

increasing demand for higher fire ratings of these walls, more than two layers of

plasterboard linings are being prescribed, which not only make the construction

process very laborious but also the resulting walls become very heavy.

Efforts have also been made to improve the fire ratings of the wall systems by using

different types of insulations in the wall cavities, but contradicting results were

obtained. Sultan and Lougheed (1994) performed several small scale fire resistant

tests on gypsum board clad steel wall assemblies (914 mm x 914 mm) using glass

fibres, rock fibres and cellulose fibres as cavity insulation. They noted that the rock

and cellulose fibre cavity insulations improved fire resistance rating by approximately

30 minutes when compared with non-insulated wall assemblies, whereas only a small

benefit was noted in the case of specimens using glass fibres. The cavity side of the

exposed gypsum board of insulated wall assemblies heated up more rapidly reaching

temperature levels of 7000C much earlier when compared to that in non-insulated wall

assemblies. Following the calcination of the exposed plasterboard, the exposed side of

the cavity recorded higher temperatures when compared to that in non-insulated wall

assemblies. Sultan (1995) carried out full scale fire resistance tests on non-load

bearing gypsum board wall assemblies and noted that when rock fibre was used as


cavity insulation the fire resistance rating increased by 54% over the non-insulated

wall assembly. Use of glass fibre as cavity insulation did not affect the fire

performance while cellulose fibre insulation reduced the fire resistance. Feng et al.

(2003) conducted fire tests on non-load bearing small scale wall systems and reported

that the thermal performance of wall panels improved with the use of cavity

insulation.

In summary, past research has produced contradicting results about the benefits of

cavity insulation to the fire rating of stud wall systems and hence further research is

needed. There is also a need to develop new wall systems with increased fire rating.

This chapter introduces a new wall system that uses a thin insulation layer between

two plasterboards on each side of stud wall frame instead of cavity insulation. It then

presents the details of a series of fire tests of non-load bearing (NLB) walls, examines

and compares their thermal performance, and makes suitable recommendations.

5.2 Test Specimens

Fire tests were conducted on nine small scale wall assemblies each measuring 1280

mm in width and 1015 mm in height. The wall assemblies typically consisted of three

commonly used cold-formed steel lipped channel section studs (90 x 40 x 15 mm)

spaced at 500 mm. The studs were fabricated from galvanized steel sheets (G500)

having a nominal base metal thickness of 1.15 mm and a minimum yield strength of

500 MPa. Test frames were built (see Figure 5-1) by attaching the studs to the top and

bottom tracks made of 1.15 mm G500 steel unlipped channel sections (92 x 50 mm)

using 12 mm long self-drilling wafer head screws. Test specimens were built by lining

the test frames with one or two layers of gypsum plasterboards manufactured by Boral

Plasterboard under the product name of FireSTOP. All the plasterboards used were

1280 mm in width and 1015 mm in height with a thickness of 16 mm and a mass of

13 kg/m2. The nine wall specimens built were divided into four categories as shown in

Table 5-1.


Table 5-1: Details of non-load bearing wall specimens

Category Specimen

No. Configuration Objective

1 To study the effect of single layer of plasterboard (1x1) on both sides of frame on the fire rating of wall specimens

I

2

To study the effect of vertical joint in the exposed plasterboard over the central stud.

II 3

To study the effect of dual layers of plasterboard (2x2) on both sides of frame on the fire rating of wall specimens

4 To study the effect of glass fibre used as cavity insulation in a wall specimen with two layers of plasterboard (2x2).

5

To study the effect of rock fibre used as cavity insulation in a wall specimen with two layers of plasterboard (2x2).

III

6

To study the effect of cellulose fibre used as cavity insulation in a wall specimen with two layers of plasterboard (2x2).

7

To study the effect of glass fibre used as external insulation between the two layers of plasterboard on each side of a wall specimen with two layers of plasterboard (2x2)

8

To study the effect of rock fibre used as external insulation between the two layers of plasterboard on each side of a wall specimen with two layers of plasterboard (2x2)

IV

9

To study the effect of cellulose fibre used as external insulation between the two layers of plasterboard on each side of a wall specimen with two layers of plasterboard (2x2)


1.15 mm G500 steel unlipped channel track: 92 x 50 mm

1.15 mm G500 steel lipped channel stud: 90 x 40 x 15 mm

Figure 5-1: Typical steel wall frame used to build NLB test wall specimens

5.3 Construction Details of Test Specimens

Test Specimen 1

Test steel frame shown in Figure 5-1 was lined on both sides by a single layer of

plasterboard (1x1 assembly) covering the entire frame without any joints (see Figure

5-2).

Thermocouple wires



The plasterboards were attached to the three studs of the steel frame by 25 mm long

self-drilling bugle head screws at 300 mm centres and to the top and bottom tracks at

250 mm centres along the top and bottom edges of the board.

K type thermocouple wires were located on the steel frame, three on each stud at mid-

height to measure the temperatures of the hot flange (flange attached to the exposed

plasterboard), web, and the cold flange (flange attached to the ambient plasterboard).

These thermocouples allowed the determination of the average stud temperature and

the temperature gradient across the stud at mid-height. Additional thermocouples were

attached at the mid-height of the plasterboard to measure temperatures inside the wall

cavity and on the fire exposed surface (Figure 5-3 shows the locations of 15

thermocouples used across the wall assembly). To measure the temperature of the

ambient surface of the wall assembly, five more thermocouples were attached to the

unexposed surface of the plasterboard, one thermocouple at the centre of the wall and

one at the centre of each quarter section of the assembly giving a total of 20

thermocouples.

Pb1

Pb2

Figure 5-3: Thermocouple Locations for Test Specimen 1

Test Specimen 2

Construction of Test Specimen 2 was identical to that of Test Specimen 1, but with a

vertical joint in the exposed plasterboard located on the hot flange of the central stud.

A screw spacing of 200 mm on centres was adopted along each of the two

plasterboard edges forming the joint. The screw positions along both the edges were

staggered giving a screw spacing of 100 mm along the stud. The joint was taped and

covered by two applications of joint compound. The instrumentation was identical to

Test Specimen 1 (see Figure 5-4).


Vertical Joint

Pb1

Pb2

Figure 5-4: Test Specimen 2 with a Joint in the Exposed Plasterboard over the Central Stud and Thermocouple Locations

Test Specimen 3

Test Specimen 3 was built by lining the steel frame with two layers of plasterboard on

either side (2x2 assembly). The base layer plasterboards were first attached to the

three studs by 25 mm long self-drilling bugle head screws at 300 mm centres. The

face layer plasterboards were then attached by 45 mm long self-drilling bugle head

screws spaced at 300 mm centres and penetrating the studs midway between the base

layer screws. Nineteen thermocouples were used to measure the temperature variation

across the mid-height of the wall assembly as shown in Figure 5-5. Five additional

thermocouples were used as in the previous specimens to measure the temperature of

the ambient surface of the wall assembly.

Thermocouple

Figure 5-5: Thermocouple Locations for Test Specimen 3

Test Specimen 4

This specimen was built similar to Test Specimen 3, but with the cavity filled with

two layers of 50 mm thick glass fibre mats of original density 13.88 kg/m3

compressed to 90 mm thickness (i.e. the depth of the cavity) giving the insulation a

density of 15.42 kg/m3 (ρ2 = ρ1 x t1/t2 = 13.88 x 100/90 = 15.42 kg/m3). Figure 5-6


shows the laying of the glass fibre mats in the cavity of the wall assembly between the

studs. Care was taken to pack the cavities of the individual studs and tracks with

insulation so as not to leave any air pockets in the wall cavity. Insulation was also

packed into the wall cavity beyond the end studs to establish conditions as close to the

central stud as possible. The instrumentation used was identical to that used for Test

Specimen 3 (see Figure 5-7).

(a) (b)

Non-load bearing wall specimen using glass fibre as cavity Insulation

packed into the wall cavity

beyond the end studs

Thermocouple wires

Steel platform to support the Test Specimen

(c)

Figure 5-6: Construction and Placement of Test Specimen 4 in the Furnace


Cavity Insulation

Figure 5-7: Thermocouple Locations for Test Specimens 4, 5 and 6

Test Specimen 5

Test Specimen 5 was built similar to Test Specimen 4, but with rock fibre of density

100 kg/m3 used as cavity insulation. Two mats each of 25 mm in thickness were

placed in the cavity of the wall. This left a gap of 40 mm between the insulation and

the cavity facing side of Plasterboard number three (Base layer plasterboard on the

ambient side) (see Figure 5-8). The thermocouple locations and numbers were

identical to that of Specimens 3 and 4 (see Figure 5-7).



Test Specimen 6

Test Specimen 6 was built similar to Test Specimens 4 and 5, but with cellulose fibre

used as cavity insulation. The insulation was wet sprayed into the cavity until it was

filled completely (see Figure 5-9). The calculated density of cellulose insulation in the

cavity was 125 kg/m3. This was obtained by using the expression ρ = {[weight of Test

Specimen 6 (with insulation) - weight of Test Specimen 3 (without

insulation)]/volume of cavity} Instrumentation was the same as Specimens 3, 4 and 5.


Test Specimen 7

In Test Specimen 7, a layer of 25 mm thick glass fibre insulation of density 37 kg/m3

was sandwiched between the two plasterboards, thus forming composite panels on

either side of the steel frame. The face plasterboard layer was attached through the

insulation layer to the base layer and the frame with 65 mm long drywall screws with

bugle heads, spaced at 300 mm centres along the studs and 250 mm centres along the

top and bottom edges connecting to the tracks. A total of 28 thermocouples including


five of them on the ambient surface were used to measure the temperature profiles

across the wall assembly (see Figure 5-10).

Figure 5-10: Thermocouple Locations for Test Specimens 7, 8 and 9

Test Specimen 8

Test Specimen 8 was built similar to Specimen 7, but with a 25 mm layer of rock fibre

of 100 kg/m3 used as insulation to form the composite panels. The methods of

construction and instrumentation used were the same as for Test Specimen 7.

Test Specimen 9

Test Specimen 9 was built similar to Specimens 7 and 8, but with a 25 mm layer of

cellulose fibre of density approximately 108 kg/m3 used as insulation to form the

composite panels. The fibre was wet sprayed on to the base plasterboard layer and

then covered by the face layer (see Figure 5-11). Sixty mm wide plasterboards were

used along the edges to enable the inclusion of cellulose insulation within a firmly

connected specimen. The instrumentation used was the same as for Test Specimen 7.


Border frame built using two layers of plasterboard strips each of 60 mm width and thickness 13 mm

Specimen surface ready for wet spraying of cellulose fibre

Thermocouple wires

(a)

Facing Plasterboard on ambient side with thermocouple wires

Wet spray up of cellulose fibre

(b)


The thermocouple wires in all the test specimens were taken directly to the outside of

the wall (ambient side) through small holes drilled in the unexposed plasterboard.

This was done so as to have minimum length of thermocouple wires within the wall

thus minimizing the possible contact of the lead wires forming hot junctions at

locations other than at those where the measurements are desired. The ends of each

wire were colour coded so as to be able differentiate between them once the wall

assembly was completed.


5.4 Test Set-up and Procedure

The custom built adapter described in Chapter 4 was used to test the non-load bearing

specimens. Fire tests were carried out by exposing one face of the specimens to heat

in this propane-fired vertical furnace as shown in Figure 5-12. The specimens were

subjected to a heating profile, which followed the standard-time temperature fire

curve as given in AS 1530 Part 4 (SA, 2005).

Large Furnace

Vents on both sides of furnace

Gate to control exhaust opening

Adapter

Figure 5-12: Test Specimen placed in the specially built adapter in the large furnace

Similar to the procedure followed to test the plasterboard panels (as described in

chapter 4) the furnace temperature was measured using four thermocouples

symmetrically placed about the horizontal and vertical centre lines and the average

temperature of which was used by a software to control the furnace heat according to

the cellulosic fire curve (Standard time-temperature curve) given in AS 1530.4 (SA,

2005). Additional thermocouples were placed within the furnace to measure the

chamber temperature. The average temperature of these thermocouples was used as

furnace temperature for the plotting of graphs. The specimens were installed in the

furnace as shown in Figure 5-13.


Test specimen with vertical edges free to deflect laterally

Gaps sealed by Isowool

Horizontal wooden beam to attach the LVDTs

(a)

Top clamps that allow vertical expansion of specimen

Pressure Transducer

Data Logger

LVDTs

(b) Figure 5-13: Test Specimen subjected to fire on one side

Specially designed clamps positioned at the top to hold the specimen in place allowed

the specimen to expand freely during the test. The vertical edges of the specimen were

kept free to allow lateral deformations. All the gaps and openings around the

specimen were sealed using Isowool. Three Linear Variable Differential Transducers

(LVDTs) mounted on a wooden beam were used to measure the mid-height lateral

deflections of the studs. One LVDT was positioned parallel to the specimen at the


centre to measure axial deformations. A pressure transducer was used to measure the

chamber pressure variations during the test. The failure of the small scale test

specimens was based on the integrity and insulation criteria in AS 1530.4 (SA, 2005).

The furnace and specimen temperatures were recorded using an automatic data-

acquisition system at intervals of one minute.

Based on AS 1530.4 (SA, 2005), the assembly was deemed to have failed if any one

of the following occurred first:

1. A single point temperature reading on the unexposed surface of the specimen

exceeded the ambient temperature by 180 0C;

2. The average of the five thermocouples on the unexposed surface of the specimen

exceeded the ambient temperature by 140 0C

3. Passage of flame or smoke for a minimum duration of 10 s through the unexposed

surface of the specimen.

5.5 Observations, Results and Discussion

5.5.1 Test Specimens 1 and 2

5.5.1.1 Visual Observations

Both specimens were exposed to the cellulosic fire curve in the furnace for slightly

more than three hours (Specimen 1-200 minutes and Specimen 2-190 minutes). When

the specimens were examined at the end of the test, it was noted that both the exposed

and ambient side plasterboards were severely affected in both specimens, but were

intact i.e. did not fall off during the test. The ambient surface of the unexposed

plasterboard in both specimens showed discolouration of paper and folds indicating

development of cracks on the cavity facing surface of the ambient plasterboard. All

the studs in both the frames were in good condition as seen in Figures 5-14 and 5-15.

The joint in the exposed plasterboard of Test Specimen 2 had opened up about 5 to 10

mm over the height of the stud. A small part of the joint at the bottom of the central

stud is visible in Figure 5-15.


Cavity facing side of ambient

Figure 5-14: Test Specimen 1 after the fire test

Joint

Figure 5-15: Test Specimen 2 after the fire test

(Exposed plasterboard fell-off after the test during handling)


5.5.1.2) Time-Temperature Profiles

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Te

mp

era

ture

(oC

)

AS 1530.4 Furnace FS HF WebCF Pb1-Cav Pb2-Cav Amb

Figure 5-16: Time-Temperature Profile for Test Specimen 1 (No joints in

plasterboard)

0

100200

300400

500600

700

800900

10001100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Te

mp

era

ture

(oC

)

AS 1530.4 Furnace FS HF Web

CF Pb1-Cav Pb2-Cav Amb

Figure 5-17: Time-Temperature Profile for Test Specimen 2 (With a joint in the

exposed plasterboard over the central stud)

Note:

AS 1530.4: Cellulosic fire curve (standard time-temperature curve) given by Australian Standard 1530 Part 4


Furnace: Average time-temperature curve followed by the furnace

FS: Average time-temperature profile of exposed surface of Plasterboard 1

HF: Average time-temperature profile of the hot flanges

Web: Average time-temperature profile of the webs

CF: Average time-temperature profile of the cold flanges

Pb1-Cav: Average time-temperature profile of the cavity facing surface of Plasterboard 1

Pb2-Cav: Average time-temperature profile of the cavity facing surface of Plasterboard 2

Amb: Average time-temperature profile of the unexposed surface of the wall

Figures 5-16 and 5-17 show the time-temperature profiles across Test Specimen 1 and

Test Specimen 2, respectively, when exposed to fire from one side. The steel stud

temperatures are seen to remain in a very narrow band in Specimen 1 (giving almost a

uniform temperature distribution across the cross-section) whereas a slight dispersion

is seen in Specimen 2. The central studs were critical in both the specimens as they

showed higher temperature than the end studs over the entire test. The temperature

rise of the studs was seen to occur in three phases. The first phase (which consisted of

the initial 7-8 minutes) showed a rapid gain in the stud temperatures up to about

1000C. Around this temperature the second phase started with the heating rate

decreasing and almost becoming zero due to the presence of moisture (free and

chemically bound water) in the gypsum plasterboard. The duration of this phase

depends on the time required to vaporize and drive away the moisture across the

thickness of the shielding plasterboard. The third phase (this was around 20 minutes

for both specimens) started soon after the evaporation of the moisture leading to a

rapid increase in the stud temperatures. Both specimens followed the same pattern

identically without showing any influence of the joint on the heating rates of the studs

as seen in Figures 5-18 to 5-20.


0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Te

mp

era

ture

(oC

)

Sp1-S1HF Sp1-S1CF Sp2-S1HF Sp2-S1CF


0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Te

mp

era

ture

(oC

)



Sp1-S1/2/3HF: Time-temperature profile followed by the hot flange

of Stud 1/2/3 in Specimen 1

Sp 1-S1/2/3CF: Time-temperature profile followed by the cold flange


Sp2-S1/2/3HF: Time-temperature profile followed by the hot flange


Sp 2-S1/2/3CF: Time-temperature profile followed by the cold flange


The opening of the vertical plasterboard joint, caused by the shrinkage of the gypsum

plasterboard (calcination), appears to affect the central stud after the initial period

(time required for the weakening of the joint) of 70 minutes of fire exposure, as up to

that time the hot flange temperatures of both specimens were identical (see Figure 5-

19). A sharp increase in the temperature of the central stud in Test Specimen 2 is seen

with the deterioration and opening of the joint beyond this initial period (see Figure 5-

19 and Table 5.2). Table 5.2 compares the central stud temperatures of both the

specimens up to the end of the test.


Table 5-2: Central Stud Temperatures of Test Specimens 1 and 2

Test Specimen

Time

Temp.

30

(min)

60

(min)

90

(min)

120

(min)

150

(min)

180

(min)

HF 282 524 573 603 634 677

W 231 465 531 576 606 643

1

(Without

Joint)

CF 182 416 495 560 591 628

HF 263 504 683 847 930 957

W 212 439 553 648 720 782

2

(With

Joint)

CF 164 390 517 648 724 787

Note: HF – Hot Flange, W – Web, CF – Cold Flange

In spite of the increase in the central stud temperatures in Test Specimen 2, the

ambient side (plasterboard) temperatures of both specimens (i.e. with and without the

joint) remained almost identical up to 130 minutes from the start of the test (see

Figure 5-21)

0

25

50

75

100

125

150

175

200

225

250

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Tem

per

atu

re (

oC

)

Sp1-Amb Sp2-AmbCritical avr temp. 170 Insulation Failure Time (Sp-1) 89 minInsulation Failure Time (Sp-2) 92 min

Figure 5-21: Average Unexposed Surface Temperature of Test Specimens 1 and 2


Note:

Sp1/2-Amb: Average time-temperature profile on the ambient surface of specimen 1/2

This implies that although the joint had begun to open after the initial period (70

minutes) exposing the hot flange of the central stud in Test Specimen 2, the

plasterboards at the joint had not been detached from the screws. This allowed the

plasterboards to remain attached to the studs and prevented any sudden ingress of heat

from the furnace chamber into the wall cavity (see Figure 5-22).

0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Tem

per

atu

re (

oC

)

Sp1 Pb2-Cav Sp2 Pb2-Cav

Figure 5-22: Time-Temperature Profiles of Cavity facing surfaces of Specimens 1

and 2

Note: Sp1/2 Pb2-Cav: Average time-temperature profile on the cavity facing surface of Plasterboard 2

5.5.1.3) Specimen Behavior

Lateral deflections of both specimens showed the same pattern with the end studs

deflecting less than the central stud. After the initial 20 minutes of protection offered

to the studs by the single layer of plasterboard, the hot flange temperatures of the

studs started rising sharply creating a temperature gradient across the cross-section of

the studs. This caused the studs to elongate more on the hot side than on the cold side

and forced the studs to bow towards the fire. The lateral deflection profiles of the

central stud in both specimens shown in Figure 5-23 are almost identical implying

little or no effect of the joint on the stud deformation. The lateral deflections, although

very small, were seen to be negative (i.e. towards the furnace) for both specimens


with a maximum of 2.75 mm occurring at approximately 45 minutes from the start of

the test.

-3.00

-2.50

-2.00

-1.50

-1.00

-0.50

0.00

0.50

0 10 20 30 40 50 60 70 80 90 100

Time (min)

De

fle

cti

on

s (

mm

)

Sp1-S2 Sp2-S2

Figure 5-23: Lateral Deflections of the Central Studs in Test Specimens 1 and 2

Note:Sp1/2-S2: Lateral deflection profile of stud 2 (central stud) in specimen 1/2

5.5.1.4) Wall Failure

Insulation failure (i.e. thermal failure) of Test Specimens 1 and 2 occurred at 89 and

92 minutes, respectively. At this time the average temperature of the unexposed

plasterboard surface of test specimens exceeded the ambient temperature of 300C by

140 0C.

5.5.1.5) Significance

A single layer of 16 mm FireSTOP (Boral) plasterboard on the fire side was seen to

offer an initial protection of around 20 minutes to the studs after which the stud

temperatures increased rapidly. Test Specimen 2 also showed consistent results. Test

Specimen 1, although built without joints in the plasterboard, did not show any

improvement in the fire rating when compared with the performance of Test

Specimen 2 built with a vertical central joint in the exposed plasterboard. Test

Specimen 2 had failed by insulation, before the effect of the joint, could be noticed on

its ambient surface. However, the vertical joint is likely to reduce the fire rating of


load bearing walls as the rapidly rising temperatures in the studs is likely to cause a

premature structural failure of the studs.

5.5.2 Test Specimens 3, 4, 5 and 6

5.5.2.1 Visual Observations

Test Specimen 3 (No cavity insulation), Test Specimen 4 (Glass fibre as cavity

insulation), Test Specimen 5 (Rock fibre as cavity insulation) and Test Specimen 6

(Cellulose fibre as cavity insulation) were subjected to heat in the furnace for slightly

more than 3 hours. Inspection soon after the test showed that the Plasterboards 1 and 2

(Fire side plasterboards) in Specimen 3 were still intact whereas they had partially

fallen off in the case of Specimens 4, 5 and 6. They fully collapsed due to their

extreme brittleness when they were removed from the furnace and placed on the

laboratory floor for inspection. Plasterboard 3 (base plasterboard on the ambient side)

in Specimen 3 was not as severely damaged as it was in the case of Specimens 4, 5

and 6. The cavity insulations of Specimens 4 and 6 were seen to be completely burnt

out with only traces of cellulose fibre ashes left behind in Specimen 6, whereas the

rock fibre insulation in Specimen 5 was still visible although it had completely lost its

integrity. Studs of Specimen 3 were seen to be in good condition whereas the studs in

the cavity insulated specimens were seen to be severely damaged, in particular the

ones in Specimen 6 using cellulose fibre as cavity insulation (see Figures 5-24 to 5-

27). The unexposed surface of all the specimens showed no signs of damage or effect

of temperature until the end of the test.


Studs in good condition

Exposed plasterboards were still intact at the end, but fell off during handling

Cavity facing surface of base layer plasterboard on ambient side.

Figure 5-24: Test Specimen 3 after the fire test (no cavity insulation)

Figure 5-25: Test Specimen 4 after the fire test (glass fibre cavity insulation)


(a) Front view of specimen after fire test (b) Side view of specimen after fire test

(c): Central stud showing a missing (d): Damaged central stud flange in the middle portion

Figure 5-26: Test Specimen 5 after the fire test (rock fibre as cavity insulation)


Central stud has totally disappeared in the middle portion.

(a) Side view of test specimen after fire test

(b) Badly deformed end stud showing (c) Central stud totally consumed separation of web from the flanges.

Figure 5-27: Test Specimen 6 after the fire test (cellulose as cavity insulation)

5.5.2.2 Time-Temperature Profiles

a) Plasterboard Surfaces: (Figures 5-28 to 5-30)

i) Average temperature of the interface surface between the exposed

Plasterboards 1 and 2 (Pb1-Pb2)

The temperature on this surface was seen to increase in all the specimens (in three

phases) from approximately four minutes after the test was started. In the initial


couple of minutes (first phase) all the specimens showed a sharp increase in the

temperature from ambient to around 800C beyond which the temperature did not

increase much (second phase) due to the hydration of the exposed plasterboard with

the temperature gradually increasing to 1200C by the end of 20 minutes. By this time

all the free and chemically bound water was probably driven out with the gypsum

board starting to dry and shrink developing and propagating fine shrinkage cracks.

Beyond 20 minutes (third phase) all the specimens showed a sharp increase in the

temperature of the interface (Pb1-Pb2), which continued until the end of the test,

except in the case of Specimen 3 where the temperature rise became gentler after

crossing 7400C at 90 minutes. By the end of 130 minutes the interface temperature in

Specimen 3 had reached 8000C whereas this temperature was reached 40 minutes

earlier by the cavity insulated specimens, ie. at about 90 minutes (95 minutes for

Specimen 6). This sustained rise in the temperature of the interface in cavity insulated

specimens was considered to be due to the heat being blocked and redirected back to

the exposed plasterboards by the insulation in the cavity.

The exposed plasterboard (Pb1) must have fallen off from Specimens 4, 5 and 6 after

about 132 minutes, 124 minutes and 137 minutes, respectively, as can be seen in the

sudden jump in temperature of the interface. In the case of Specimen 3 the increase in

temperature gradient at about 180 minutes probably indicates the loss of integrity of

Plasterboard 1. The temperature of the interface between Plasterboards 1 and 2 (Pb1-

Pb2) was seen to be between 9000C and 10000C in all the specimens when

Plasterboard 1 started to disintegrate.

ii) Average temperature on the cavity facing surface of the exposed Plasterboard

2 (Pb2-Cav)

An initial increase in temperature (Phase 1) was followed by a plateau (Phase 2)

extending up to 60 minutes in Specimen 3 and 55 minutes in the cavity insulated

specimens. The temperature was under 1200C at the end of the plateau. Pb2-Cav side

showed a much longer plateau than the one displayed by the ambient side of

Plasterboard 1 (Pb1-Pb2). This was because the fire side of Plasterboard 2 was

exposed to a fire curve defined by the interface temperature between Plasterboards 1

and 2, which was much less severe than the standard time-temperature curve

experienced by the fire side of Plasterboard 1.


In Specimen 3 the plateau was followed by a gradual increase in temperature (Phase

3), reaching 4000C by the end of 115 minutes. Specimens 4 and 5 showed a much

more rapid rise in temperature from 55 to 70 minutes with the temperature crossing

4000C beyond which it became gentler (but much steeper when compared with

Specimen 3). In Specimen 4 the temperature reached 7000C in 124 minutes and

crossed 10000C by 170 minutes whereas in Specimen 5, it reached 7000C in about 120

minutes and crossed 10000C by 145 minutes (25 minutes earlier than Specimen 4).

Specimen 6 also showed a rapid increase in temperature from 55 to 75 minutes

reaching 4100C beyond which it continued to rise gradually reaching 7000C by the

end of 140 minutes. This was followed by a very rapid rise in temperature crossing

10000C by 155 minutes.

The temperature in Specimen 3 did not rise as rapidly as in the case of cavity

insulated specimens probably because in Specimen 3 the base layer plasterboard on

the fire side was allowed to lose heat via radiation in the empty cavity. The fast

passage of heat across the cavity and into Plasterboard 3 which served as a heat sink

to the fire side plasterboard checked its temperature escalation. In contrast, in

Specimens 4, 5 and 6, the insulation in the cavity due to its very low conductivity was

blocking the flow of heat and redirecting it back to the cavity facing surface of the

exposed plasterboard thus forcing a sharp and sustained rise in its surface temperature.

Fastest increase in temperature was noted in Specimen 5 with rock fibre insulation in

the cavity whereas the responses of Specimens 4 and 6 were nearly the same.

Specimens 5 and 6 show a rise in temperature gradient at about 124 minutes and 137

minutes, respectively, which coincides with the fall off times of Plasterboard 1 in

these specimens. The fall off times of Plasterboard 2 in the case of Specimens 4, 5 and

6 appears to be around 148 minutes, 145 minutes and 146 minutes as seen in the sharp

rise in the temperature recorded by the thermocouple on the cavity side of

Plasterboard 2 (Pb2-Cav). Plasterboard 2 in Specimen 3 was intact throughout the

test.


0100200300400500600700800900

100011001200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Te

mp

era

ture

(oC

)

AS 1530.4 Furnace FS Pb1-Pb2Pb2-Cav Pb3-Cav Pb3-Pb4 Amb

Figure 5-28: Time-Temperature Profiles of Plasterboard surfaces in Test Specimen 3

(No Cavity Insulation)

Figure 5-29: Time-Temperature Profiles of Plasterboard surfaces in Test Specimen 4

(Cavity Insulation – Glass Fibre)


0100200300400500600700800900

100011001200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Te

mp

era

ture

(oC

)

AS 1530.4 Furnace FS Pb1-Pb2Pb2-Ins Ins-Pb3 Pb3-Pb4 Amb

Figure 5-30: Time-Temperature Profiles of Plasterboard Surfaces inTest Specimen 5

(Cavity Insulation-Rock Fibre)

0

100

200

300

400

500600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Te

mp

era

ture

(oC

)

AS 1530.4 Furnace FS Pb1-Pb2Pb2-Ins Ins-Pb3 Pb3-Pb4 Amb


(Cavity Insulation – Cellulose Fibre)

Note:

Pb1-Pb2: Average time-temperature profile of the interface between Plasterboards 1 and 2

Pb3-Pb4: Average time-temperature profile of the interface between Plasterboards 3 and 4

Pb2-Ins: Average time-temperature profile of the interface between Plasterboard 2 and insulation

Ins-Pb3: Average time-temperature profile of the interface between insulation and

Plasterboard 3


iii) Average temperature on the cavity facing surface of the ambient side

Plasterboard 3 (Pb3-Cav)

In Specimen 3 the initial rise in temperature of this surface almost coincided with the

rise in temperature of Pb2-Cav surface, probably due to the fast transmission of heat

via radiation across the cavity. The plateau extended up to 65 minutes (as compared to

60 minutes on Pb2-Cav surface) reaching a temperature of 1280C. The time-

temperature profile of Pb3-Cav surface followed the corresponding profile of Pb2-

Cav very closely but with a slight lag that never exceeded 700C.

In Specimen 4 the initial temperature rise on Pb3-cav surface started 6 minutes after

the initial rise on Pb2-Cav surface due to the protection offered by the insulation in

the cavity. The plateau was seen to extend up to 66 minutes (11 minutes longer than

that of Pb2-Cav surface). The plateau was followed by a gradual increase in

temperature up to 124 minutes reaching a temperature of 2800C when the temperature

on the Pb2-Cav surface had reached 7000C. The disintegration of the glass fibre

insulation must have started around this temperature (7000C) from across the cavity

reducing its density and thickness as the Pb3-Cav side started recording a very rapid

rise of temperature reaching 7850C by 153 minutes and crossing 10000C by 170

minutes. At this point of time the glass fibre insulation must be considered totally

useless as the time-temperature curve merged with that of the Pb2-Cav surface. Once

the disintegration of the glass fibre started at around 124 minutes, it took 29 minutes

for the insulation to become totally ineffective.

In Specimen 5 the temperature started rising with an 8 minute delay, reaching 760C by

around 75 minutes. From 75 minutes to 130 minutes the temperature rose gradually,

reaching 2260C by 130 minutes at which time the temperature across the insulation on

the Pb2-Cav surface was around 8500C developing a difference of 6240C across the

cavity. This temperature difference across the insulation was maintained fairly

constant up to 145 minutes, beyond which it started to converge gradually with the

decrease in the insulation integrity, crossing 10000C at 170 minutes before merging

with the time-temperature curve of Pb2-Cav. Disintegration of rock fibre insulation

started at about 145 minutes and took approximately 25 minutes to become totally

ineffective.


In Specimen 6, the temperature started rising with a 9 minute delay and reached 980C

by about 94 minutes. This plateau was the longest when compared with 66 minutes of

Specimen 4 and 75 minutes of Specimen 5, indicating the superior initial insulating

properties of the cellulose fibre over the other two types. Beyond 94 minutes the

temperature was seen to rise gradually reaching 2150C at 110 minutes at which instant

the temperature across the cavity on the Pb2-Cav surface was 6300C giving a

temperature of 4150C across the cavity. A very sharp increase in temperature was

noticed beyond 145 minutes reaching 10000C in less than 10 minutes before merging

with the Pb2-Cav surface curve. The cellulose fibre insulation in the cavity of

Specimen 6 was intact for almost 145 minutes beyond which it disintegrated in just 10

minutes as against the 29 minutes taken by glass fibre insulation and 25 minutes taken

by the rock fibre insulation.

iv) Average temperature on the ambient side of unexposed Plasterboard 3

(Pb3-Pb4)

Specimen 3 showed a plateau up to 131 minutes beyond which it rose very gradually

to 2200C by 170 minutes. Specimen 4 showed a plateau up to 144 minutes beyond

which it rose very rapidly crossing 7000C by 187 minutes. Specimens 5 and 6 both

showed a plateau up to 160 minutes beyond which the graphs showed a sharp rise

crossing 7000C by 196 and 187 minutes, respectively. The gradual increase in the

temperature of Pb3-Pb4 surface in Specimen 3 was because the fire side plasterboards

were still intact without loosing their integrity whereas the rapid increase in

temperature in the cavity insulated specimens was probably due to the extreme

damage suffered by the fire side plasterboards due to their accelerated calcination

caused by the redirected heat from the cavity insulation. This led to the early collapse

of the fire side plasterboards and the subsequent quick disintegration of the cavity

insulation, leaving the ambient side plasterboards to face the full impact of fire from

the furnace.

v) Average temperature on the ambient side of unexposed Plasterboard 4

The average temperature on the ambient surface of the cavity insulated specimens was

only marginally lower than that recorded by Specimen 3 with no insulation up to

approximately 130 - 150 minutes beyond which the cavity insulated specimens


showed a sharper increase in temperature following the disintegration of half the wall

on the fire side and surpassed the ambient side temperature of Specimen 3.

b) Steel Surfaces (Figures 5-32 to 5-35)

The first phase involving the initial rapid temperature gain from ambient to about

1000C in the studs of all the specimens occurred in the first 20 minutes. This was

followed by the second phase wherein the temperature gradient became almost zero

up to about 60 minutes in Specimen 3 and 50 minutes in the case of Specimens 4, 5

and 6. The length of this plateau depended upon the time required to evaporate the

water contained in the two external plasterboards. Once this water had vaporized and

been driven out a rapid increase in the stud temperatures (Phase 3) was noted in all the

specimens until the end of the test.

The steel temperatures in the specimens depended upon the Pb2-Cav surface

temperature history. The studs of Specimen 3 picked up heat from this surface by

conduction (through the physical contact of the studs with the plasterboard surface),

convection (movement of hot air within the wall cavity) and by direct radiation. In

Specimens 4, 5 and 6 whose cavity was filled with insulation, the heat was passed on

to the studs from the plasterboard surface by conduction alone through steel and

insulation material. As the heat transfer via radiation is the fastest, the studs of

Specimen 3 had a more uniform temperature gradient across the studs as compared to

the large temperature variations across the stud cross-sections in the cavity insulated

wall specimens. The large temperature gradients across the cross-sections of studs in

cavity insulated specimens was due to the low conductivity of the insulation in the

cavity, which reduced the heat flow towards the cold flanges of the studs and

accelerated the temperature rise of the hot flanges due to the additional heat redirected

from the surface of insulation. This caused the hot flanges of the studs in cavity

insulated specimens to heat up more rapidly than those of Specimen 3. The

temperatures of the studs remained high over the entire test period leading to their

earlier damage. This is why severe damage and burn-out was observed in the studs as

shown in Figures 5-26 and 5-27.


0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Te

mp

era

ture

(oC

) S1-HF

S2-HF

S3-HF

S1-W

S2-W

S3-W

S2-CF

S3-CF

Figure 5-32: Time-Temperature Profiles across Studs in Test Specimen 3


Note:

S1/2/3-HF: Time-temperature profile followed by the hot flange of Stud 1/2/3

S1/2/3-W: Time-temperature profile followed by the web of Stud 1/2/3

S1/2/3-CF: Time-temperature profile followed by the cold flange of Stud 1/2/3

0

100200

300400

500600

700800

9001000

11001200

1300

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Tem

per

atu

re (

oC

)

S1-HF

S2-HF

S3-HF

S1-W

S2-W

S3-W

S1-CF

S2-CF

S3-CF




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

1300

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Tem

per

atu

re (

oC

)

S1-HF

S2-HF

S3-HF

S1-W

S2-W

S3-W

S1-CF

S2-CF

S3-CF


(Cavity Insulation-Rock Fibre)

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

1300

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Tem

per

atu

re (

oC

)

S1 HF

S2 HF

S3 HF

S1 W

S2 W

S3 W

S1 CF

S2 CF

S3 CF



The central studs in all the specimens showed higher temperatures at any time than

the end studs. This was probably because the end studs could dissipate heat in to the

atmosphere faster as they were closer to the end of the walls and thus had lower

confinement when compared to the central stud.


Table 5.3 shows the times taken by the hot flanges of the central studs of Specimens 3

to 6 to attain temperatures ranging from 4000C to 7000C.

Table 5.3: Hot Flange Temperature versus Time for the Central Stud

Time in Minutes Hot Flange Temperature

(oC) Specimen 3 Specimen 4

(Cav. Ins. GF) Specimen 5

(Cav. Ins. RF) Specimen 6

(Cav. Ins. CF)

400 100 78 74 91

500 144 91 82 106

600 218 107 97 125

700 119 115 139

The hot flange of Specimen 5 (with rock fibre as cavity insulation) heated up the

fastest whereas Specimen 6 (with cellulose fibre as cavity insulation) was the slowest

to heat up amongst the cavity insulated specimens. Specimen 3 with no insulation in

the cavity showed the best results with the studs remaining at temperatures much

lower than the cavity insulated specimens over the entire test period. The results of

Specimen 3 however could not be compared beyond 180 minutes as the furnace

heating regime in the case of Specimen 3 showed a sudden deviation from the

standard time-temperature curve profile due to some errors in the settings of the auto-

controlled valves regulating the flow of gas in the burner. The problem was fixed and

did not recur when Specimens 4, 5, and 6 were tested (see Figures 5-29 to 5-31).

In the cavity insulated specimens the sudden rise in the temperature of the studs was

observed within few minutes of the partial collapse of Plasterboard 1 and the severe

calcination and cracking of Plasterboard 2. As the exposed plasterboards were more

severely affected in the central portion, the middle stud was the first to show a sudden

increment in temperature. Studs 1 and 3 followed with a time delay of approximately

5 to 10 minutes. As the plasterboards of Specimen 3 were intact throughout the test,

the studs always had a gradual increase in temperature.

5.5.2.3 Entire Wall: Time-temperature graphs of Specimens 3 to 6 shown in Figures

5-36 to 5-39 display the temperature histories across the entire wall thickness with

plasterboard and steel taken together. Steel temperatures used in these figures are the

average temperatures of the three studs.


0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)

FS Pb1-Pb2 Pb2-Cav H F Web

CF Pb3-Cav Pb3-Pb4 Amb

Figure 5-36: Time-Temperature Profiles across the Cross-section of Test Specimen 3


0

100200

300

400

500600

700

800

9001000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)

FS Pb1-Pb2 Pb2-Cav HF WebCF Pb3-Cav Pb3-Pb4 Amb




0

100200

300

400

500600

700

800

9001000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)

FS Pb1-Pb2 Pb2-Cav HF W

CF Pb3-Cav Pb3-Pb4 Amb


(Cavity Insulation – Rock Fibre)

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Time (min)

Tem

per

atu

re (

oC

)

FS Pb1-Pb2 Pb2-Cav HF WebCF Pb3-Cav Pb3-Pb4 Amb




5.5.2.4 Behaviour of Specimens:

Figures 5-40 to 5-43 show the lateral deflection and axial deformation (A.D.) of studs

with time. Specimen 3 when exposed to the heating regime was observed to bow

away from the furnace reaching a maximum lateral deflection of 4 mm (in Stud 1) at

approximately 190 minutes from the start of the test. The axial deformations

(elongations) of studs were observed to start after 65 minutes of fire exposure. This is

consistent with the start of Phase 3, which is the steep increase in the stud

temperatures following the end of plateau in the time-temperature profiles of the

studs. The axial elongation steadily increased to reach a maximum of 4.2 mm by 200

minutes. Unlike Specimen 3 for which the temperature gradient of stud was small, the

cavity insulated specimens were seen to bow towards the furnace with the maximum

lateral deflections of the central studs being 3.7 mm, 5.5 mm and 5.6 mm, respectively

(negative deflection means towards the furnace). The deflections of the central studs

were seen to reverse sharply past these points in the cavity insulated specimens. This

was probably induced by the higher temperatures in the hot flanges leading to the loss

of their strength and stiffness faster than the cold flanges and thus undergoing local

and flexural buckling and/or deformations. For larger non-load-bearing walls this is

expected to occur earlier due to the greater slenderness of the studs and the larger self

weight of the walls. Specimen 3 also moved away from the furnace at the end of

testing for the same reasons given above for other specimens.

The axial elongations were measured only near the central stud for all the specimens.

In Specimens 4 and 5 the axial elongations were observed to start beyond two hours

of fire exposure reaching a maximum of 7.8 mm for Specimen 4 and 6.5 mm for

Specimen 5 in 170 and 167 minutes, respectively, whereas in Specimen 6 the

elongations started from approximately 80 minutes reaching 10.5 mm in 170 minutes.


-1.00

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Def

lect

ion

(m

m)

L.D. Stud 1 L.D. Stud 2 L.D. Stud 3 A.D.

Figure 5-40: Deflection-Time Profiles of Test Specimen 3


Note: L.D. Stud 1/2/3: Lateral-deflection time profile of Stud 1/2/3 A.D.: Axial deformation of studs

-4.00

-3.00

-2.00

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Def

lect

ion

(m

m)





-8.00

-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

8.00

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Def

lect

ion

(m

m)



(Cavity Insulation – Rock Fibre)

-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

8.00

10.00

12.00

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Def

lect

ion

(m

m)





5.5.2.5 Wall Failure

Table 5-4 shows the times at which the different portions of the wall were severely

affected contributing to the failure.

Table 5-4: Failure Times of Wall Components in minutes

Specimen Pb1:Fall

off time

Pb2:Time at partial/full collapse

Period of insulation failure

Local buckling of Hot Flange of central stud

4 132 148 124 to 150 125

5 124 145 145 to 170 145

6 137 146 145 to 155 145

All the specimens (3 to 6) were very stable with the ambient side temperature well

below the insulation failure temperature of 1650C (Ambient temperature was 250C)

throughout the test i.e. no insulation failure. If the reversal of the lateral deformations

of the studs could be considered as the failure of the steel frames caused due to the

softening and consequent local buckling of the hot flanges, then the failure times of

Specimens 4, 5 and 6 would be 125, 145 and 145 minutes, respectively. Specimen 3

showed no signs of failure until the end of the test.

5.5.2.6 Significance

1) Two layers of 16 mm FireSTOP gypsum plasterboard used in a 2x2 non-load

bearing wall construction were seen to offer around 60 minutes of initial protection to

the steel frames before losing the moisture content across both the plasterboards. This

period is more than twice that of a single board (20 minutes) as the second

plasterboard was subjected to a much less severe fire curve than the outer plasterboard

layer and hence took a longer time to expel the water held inside of it.

2) Heat transfer in the cavity of walls not filled with insulation took place via

conduction, convection and radiation. As a result of the faster transmission of heat

mostly through radiation, the temperatures across the stud cross-sections were


generally uniform, thereby resulting in minimum lateral deformations (ie. reduced

thermal bowing)

3) Use of cavity insulation was seen to be detrimental to the fire rating of walls. It not

only led to higher temperatures in the steel studs, but also to larger temperature

gradients across their depth (and increased thermal bowing effects).

4) Cavity insulated specimens were seen to bow initially towards the furnace. The

lateral deflections reversed sharply with the hot flanges of the studs undergoing local

buckling leading to the failure of the wall.

5) Among the three types of insulations used, rock fibre developed the maximum

temperature gradient across the studs whereas cellulose fibre developed the minimum.

The hot flange temperatures in the specimen using rock fibre insulation were more

than those of other specimens at any given time.

6) The heat trapped in the cavity by the insulation led to extensive stud damage in the

case of cavity insulated specimens. This was in contrast to the relatively good

condition of the studs in the non-insulated Specimen 3.

5.5.3 Test Specimens 7, 8 and 9

5.5.3.1 Visual Observations: (Figures 5-44 to 5-46)

The exposed Plasterboards 1 and 2 in all the three specimens fell off completely while

the insulation layer between the exposed plasterboards was totally consumed by the

furnace heat. Only small pieces of rock fibre insulation could be seen lying in the

debris whereas the other two insulations had completely vanished. The base layer

plasterboard on the ambient side (Pb3) had also collapsed in the central portion of all

the specimens. The ambient side plasterboard (Pb4), although cracked on the fire side,

was still intact and standing in one piece in all the specimens. The unexposed side of

the wall (ambient side of Plasterboard 4) showed no signs of any damage or

discolouration until the end of the test. Traces of glass fibre and cellulose fibre

insulation could be seen along the periphery between Pb3 and Pb4. In the case of

Specimen 8 the rock fibre insulation was found to be intact on one half of the

specimen. On the other half a portion of it had fallen off from the central area.


The cold-formed steel frames were not twisted or bent. The upper and lower tracks

were in good condition. The central stud in all the three specimens was the most

affected. The central stud in Specimen 9 (using cellulose fibre insulation) showed the

maximum damage. The long screws used to hold the external exposed plasterboards

onto the steel frame were seen to be bent downwards, probably due to the weight of

the external plasterboards acting on the heat softened screws during the test. The

yielding of the screws also may have accelerated the collapse of the external

plasterboards.

Figure 5-44: Test Specimen 7 after the fire test (Glass fibre as external insulation)


Figure 5-45: Test Specimen 8 after the fire test (Rock fibre as external insulation)

Figure 5-46: Test Specimen 9 after the fire test (Cellulose fibre as external insulation)

5.5.3.2 Time-Temperature Profiles

a) Plasterboard Surfaces: (Figures 5-47 to 5-49)


Plasterboard 1 and Insulation (Pb1-Ins)

Similar to the cavity insulated specimens, the thermocouples on the ambient side of

Plasterboard 1 in all the specimens having external insulation used in the form of

composite panels responded in about three minutes showing a rapid rise in

temperature during the first phase. This phase displayed a sharp increase in

temperature reaching about 800C in less than 2 minutes, after which the temperature

increased gradually to 1200C by the end of 18 minutes. In the case of Specimen 9,

using cellulose fibre as composite insulation, the second phase lasted only 12 minutes

(6 minutes less than that observed in Specimens 7 and 8). The third phase in the case

of these specimens was much different than the cavity insulated specimens. Unlike the

cavity insulated specimens which displayed an almost constant temperature gradient


up to failure, the temperature gain was very rapid up to 34 minutes in Specimens 7

and 8 crossing 5500C (whereas the cavity insulated specimens recorded around 3000C

by the end of 35 minutes) and 23 minutes in Specimen 9 crossing 4500C, beyond

which the graph became very gentle with the temperature of the interface in

Specimens 7, 8 and 9 crossing 8000C by the end of 145 minutes, 115 minutes and 125

minutes, respectively. The initial steep rise in temperature in the 3rd phase was

considered to be due to the heat being blocked and redirected by the following layer

of insulation.

The fall off times (either partial or complete) of Plasterboard 1 in Specimens 7, 8 and

9 were considered to be 167 minutes, 145 minutes and 125 minutes, respectively, as

the temperature after these times showed a rapid rise (vertical), merging with the fire

side curve.

0

100200

300

400

500600

700

800

9001000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220

Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 Furnace FS Pb1-Ins Ins-Pb2

Pb2-Cav Pb3-Cav Pb3-Ins Ins-Pb4 Amb




0

100200

300400

500600

700800

9001000

11001200

1300

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220

Time (min)

Tem

per

atu

re (

oC

)




(External Insulation-Rock Wool)

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 Furnace Pb1-Ins Ins-Pb2 Pb2-CavPb3-Cav Pb3-Ins Ins-Pb4 Amb


(External Insulation-Cellulose Fibre)


ii) Average temperature of the interface surface between the insulation and

exposed base layer Plasterboard 2 (Ins-Pb2)

This interface (in Specimens 7, 8 and 9) responded to the initial rise in temperature

(Phase 1) in under 4 minutes of exposure to furnace heat, reaching around 800C

rapidly and then remained almost constant (second phase) up to 22 minutes in

Specimens 7 and 8. In Specimen 9 however the second phase lasted only up to 15

minutes. The rate of temperature rise in the third phase was almost uniform in

Specimen 7 up to 85 minutes, beyond which it showed a sudden increase. This was

considered to be due to the rapid disintegration of the glass fibre insulation as the

temperature on the interface of Pb1- Ins had reached 7000C at this stage. From 85 to

95 minutes most of the glass fibre insulation would have burnt out. Beyond 95

minutes the rate of temperature rise became gentler and was constant up to 151

minutes during which time the insulation must have been totally consumed as the

graph of Ins-Pb2 surface merged with the graph of Pb1-Ins interface.

In Specimen 8, the rate of temperature rise was uniform up to 165 minutes beyond

which a sudden rise in temperature was noticed, probably due to the collapse of the

external Plasterboard 1 at about 145 minutes. A temperature difference of

approximately 2000C between the two sides of the insulation (i.e. between Pb1-Ins

and Ins-Pb2) indicates that the rock fibre insulation was still intact and functional

almost up to 180 minutes, beyond which it began to lose its integrity. Specimen 9

showed a change in the rate of temperature rise, becoming gentler after 23 minutes at

around 2300C. It continued with a uniform rate up to 125 minutes beyond which it

rose sharply, suggesting the collapse of both the external Plasterboard 1 and the

insulation disintegration.

iii) Average temperature on the cavity facing surface of the exposed Plasterboard

2 (Pb2-Cav)

The initial increase in temperature was followed by a plateau extending up to 84

minutes, 90 minutes and 70 minutes in Specimens 7, 8 and 9, respectively (compared

with 55 minutes observed in the cavity insulated specimens). This extended period for

the second phase must be due to the additional protection offered by the insulation


placed between the external plasterboards. Specimen 7 observed a more or less

uniform temperature gradient up to 173 minutes beyond which the gradient increased

as a result of the falling off of the external plasterboard and the disintegration of the

insulation. The sudden jump in the temperature at 198 minutes implies the falling off

of the base layer plasterboard (Pb2) on the fire side.

Specimen 8 also observed an almost uniform temperature gradient up to 200 minutes

beyond which it rose sharply suggesting the collapse of Plasterboard 2. Specimen 9

showed a change in gradient at about 131 minutes becoming steeper on account of the

collapse of the external plasterboard and cellulose fibre insulation at 125 minutes.

Beyond 163 minutes there was a further increase in the temperature gradient implying

a breach in the base layer Plasterboard 2, followed by its collapse.

iv) Average temperature on the cavity facing surface of the ambient Plasterboard

3 (Pb3-Cav)

In the absence of insulation, the transmission of heat across the cavity by radiation

was very quick, forcing the Pb3-Cav surface to heat up almost instantaneously and

trace very closely on the underside of the time-temperature profile of Pb2-Cav surface

with the maximum temperature difference between the two cavity surfaces being

290C, 500C and 580C in Specimens 7, 8 and 9, respectively.

v) Average temperature of the interface surface between base layer Plasterboard

3 and insulation (Pb3-Ins)

The thermocouples positioned in this surface started sensing a rise in temperature

from 12 minutes, 11 minutes and 18 minutes in Specimens 7, 8 and 9, respectively.

The temperatures climbed gradually reaching around 1000C and remained almost

constant up to approximately 130 minutes, 150 minutes and 135 minutes in

Specimens 7, 8 and 9 respectively, beyond which the third phase started with the

temperatures increasing rapidly. In Specimens 7, 8 and 9, a sharp increase in

temperature gradient was noted after about 200 minutes, 204 minutes and 168

minutes, respectively, indicating the possible breaching of Plasterboard 3 in all the

specimens. In Specimen 9, the change in gradient at about 168 minutes is not large

suggesting that cracks developed in Plasterboard 3 at this stage may not have been


very severe but progressed and widened rapidly leading to a collapse at about 184

minutes.

vi) Average temperature of the interface surface between the insulation and

Plasterboard 4 (Ins-Pb4)

The time-temperature graphs of this interface in the three specimens followed very

closely, but on the underside of the Pb3-Ins graph. The initial rise in temperature was

followed by a plateau which lasted up to 140 minutes, 175 minutes and 145 minutes

in Specimens 7, 8 and 9, respectively.

In Specimen 7, the plateau was followed by a gradual increase in temperature up to

200 minutes giving a temperature difference of 2000C on either face of the glass fibre

insulation. Beyond 200 minutes the insulation was seen to disintegrate very rapidly on

account of the steep increase in the temperature of the ambient side of Plasterboard 3

(fire side of insulation ). This led to a sharp rise in the temperature at about 204

minutes on the Ins-Pb4 interface.

In Specimen 8, the temperature rise following the plateau was very gentle up to 205

minutes, with a temperature difference 4000C being developed across the insulation.

After 210 minutes the rock fibre insulation must have lost its integrity as the

temperature of the Ins-Pb3 interface increased suddenly from 1870C to 7000C and

merged with the temperature profile of the ambient side of Plasterboard 3.

In Specimen 9, the constant temperature plateau lasted up to 145 minutes beyond

which the temperature started rising quickly. The cellulose fibre insulation must have

been intact up to 180 minutes as it maintained an almost stable temperature difference

of around 2000C between the ambient side of Plasterboard 3 and fire side of

Plasterboard 4. After about 185 minutes the temperature on the ambient side of

Plasterboard 3 crossed 7500C, leading to the burning up of the insulation. This led to a

quick rise in the temperature profile of the Ins-Pb4 interface merging it with the Pb3-

Ins profile implying the total disappearance of the insulation. The temperature-time

profile of the Ins-Pb4 interface in Specimen 8 with rock fibre insulation was the

lowest when compared with the interface profiles in Specimens 7 and 9. This clearly

indicates the superior insulating properties of rock fibre insulation over the glass fibre

and cellulose fibre insulations.


vii) Average temperature on the ambient side of unexposed Plasterboard 4

The average temperature on the unexposed wall surface was below 1000C for all the

three specimens until the end of the test. This clearly indicates that the failure of these

wall specimens would be due to the structural failure of the steel frame rather than the

thermal insulation failure caused by the heat penetration through the wall.

b) Steel Surfaces (Figures 5-50 to 5-52)

The time-temperature graphs of the studs in Specimens 7, 8 and 9 showed the profiles

of the hot flanges, webs and the cold flanges to lie in a narrow band unlike the wider

profiles that were seen in the cavity insulated specimens. This was due to the quick

transmission of the heat across the cavity by radiation leading to a more uniform

temperature variation across the individual studs in each of the walls. The initial

temperature rise in all the specimens was over within the first 20 minutes. This was

followed by a period of almost constant temperature (plateau) up to 70 minutes for

Specimens 7 and 9. However, the plateau in the case of Specimen 8 using rock fibre

insulation extended up to 80 minutes.

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 20 40 60 80 100 120 140 160 180 200 220

Time (min)

Tem

per

atu

re (

oC

)

S1-HF

S2-HF

S3-HF

S1-W

S2-W

S3-W

S1-CF

S2-CF

S3-CF




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 20 40 60 80 100 120 140 160 180 200 220 240

Time (min)

Tem

per

atu

re (

oC

)S1-HF

S2-HF

S3-HF

S1-W

S2-W

S3-W

S1-CF

S2-CF

S3-CF


(External Insulation-Rock Fibre)

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 20 40 60 80 100 120 140 160 180 200

Time (min)

Te

mp

era

ture

(oC

)

S1-HF

S2-HF

S3-HF

S1-W

S2-W

S3-W

S1-CF

S2-CF

S3-CF



In the third phase of the time-temperature graphs following the plateau, the stud

temperatures increased sharply at about 200 minutes for Specimens 7 and 8 at which

time Plasterboard 2 in these specimens fell off. In Specimen 9 the temperature

gradients became sharp at around 160 minutes which was consistent with the fall off


of Plasterboard 2 at 163 minutes. Akin to the cavity insulated specimens, the central

studs of the composite panel specimens recorded higher temperatures than the end

studs at all times. Table 5-5 shows the times taken by the hot flanges of the central

studs of Specimens 7, 8 and 9 to attain temperatures ranging from 4000C to 7000C.

The hot flange of Specimen 9 (using cellusic fibre (CF) as external insulation) heated

up the fastest whereas Specimen 8 (using rock fibre (RF) as external insulation) was

the slowest to heat up. Specimen 8 gave the best results as the rock fibre insulation

effectively maintained the steel temperatures lower than other specimens for a longer

period and displayed better insulating properties than cellulose and glass fibre.

Table 5-5: Hot Flange Temperature versus Time for the Central Stud

Time in Minutes Hot Flange Temperature

(0C) Specimen 7

(Insulation:GF) Specimen 8

(Insulation:RF) Specimen 9

(Insulation:CF)

400 117 142 122

500 148 160 132

600 159 178 140

700 175 189 151

5.5.3.3 Entire Wall: (Figures 5-53 to 5-55)

Time-temperature graphs of Specimens 7, 8 and 9 display the temperature histories

across the entire wall thickness with plasterboard and steel taken together. Steel

temperatures used are the average temperatures of the three studs. Time-temperature

graphs of Specimens 7, 8 and 9 display the temperature histories across the entire wall

widths with plasterboard and steel taken together. The hot flange temperatures of the

studs were seen to be higher than the cavity facing surface of Plasterboard 2 (Pb2-

Cav). This was probably caused by the close proximity of the hot end of the

thermocouples measuring the flange temperatures with the screws fixing the external

plasterboards to the studs. Due to the thermal bridging a small region of the flange

around the screw would be at a higher temperature, thus causing the thermocouples in

that region to record slightly higher temperatures than actual. On the other hand the

hot end of the thermocouple used for measuring the cold flange temperatures was

placed between the flange and the base layer plasterboard on the ambient side, thus


shielding it from direct radiation coming from Pb2-Cav surface. This probably caused

it to measure slightly lower values than the actual taking the profile slightly lower

than the profile of Pb3-Cav surface.

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220

Time (min)

Tem

per

atu

re (

oC

)

FS Pb1-Ins Ins-Pb2 Pb2-Cav HF WCF Pb3-Cav Pb3-Ins Ins-Pb4 Amb

Figure 5-53: Time-Temperature Profiles over the Entire Cross-section of Test Specimen 7





0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)

FS Pb1-Ins Ins-Pb2 Pb2-Cav HF WCF Pb3-Cav Pb3-Ins Ins-Pb4 Amb



5.5.3.4 Behaviour of Specimens: (Figures 5-56 to 5-58)

Specimen 7 started exhibiting slight deformations after 2 hours of fire exposure. The

specimen was seen to bow away from the furnace with the lateral deformation

reaching its maximum value of 5.8 mm in 213 minutes. The axial deformation was

not significant up to 140 minutes, beyond which axial shortening was noticed

probably caused by the lateral deformations.

Specimen 8 showed no significant lateral or axial deformations throughout the test.

The maximum axial shortening was noted to be 2 mm at which time the maximum

lateral deformation of 1 mm was reached. The thermocouples were removed from the

wall specimen after 3 hours.

Specimen 9 displayed a delayed lateral bowing towards the furnace with the central

stud reaching a maximum of 6.2 mm at around 188 minutes, at which time the axial

deformation was around 9 mm.


-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

8.00

10.00

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220

Time (min)

Def

lect

ion

(m

m)

L.D. Stud 1 L.D. STUD 2 L.D. Stud 3 A.D.



-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

8.00

10.00

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Def

lect

ion

(m

m)





-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

8.00

10.00

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Def

lect

ion

(m

m)




5.5.3.5 Wall Failure

Table 5-6 shows the times at which the different portions of the wall were severely

affected contributing to the failure.

Table 5-6: Failure times of Wall Components in Minutes

Specimen Pb1: Fall off time

Period of insulation failure between Pb1 and Pb2

Pb2: Fall off time

Pb3: Fall off time

Period of insulation failure between Pb3 and Pb4

7 167 85-95 198 200 204-210

8 145 180-190 200 204 205-210

9 125 125 163 184 185-188

The ambient side of the plasterboard in Specimens 7, 8 and 9 showed no signs of

exceeding the insulation failure temperature during the entire course of the test. The

tests were discontinued after about 3 hours of exposure to the furnace heat. For all

practical purposes the failure of Plasterboard 2 would suggest the commencement of

wall failure as the steel stud temperatures would rise quickly leading to the failure of

the cold-formed steel frame.



5.5.3.6 Significance:

1) Use of external insulation offered greater thermal protection to the studs resulting

in a near uniform temperature distribution across their cross-sections thereby

producing minimum early lateral deformation (thermal bowing).

2) The difference in temperature of the individual studs in the externally insulated

specimens was not significant as the radiation of heat in an open cavity is very fast

leading to a quick balance of temperatures in the individual studs. This would help in

reducing the building up of internal stresses in the frame caused by the unequal

expansions of the individual studs.

3) The wall can be considered to have failed when the studs reverse in lateral

deformation or when the external plasterboards collapse, whichever occurs first.

CHAPTER 6: STRUCTURAL AND THERMAL PERFORMANCE

OF LOAD BEARING WALL SYSTEMS

6.1: Introduction

A detailed experimental study was conducted in the Fire Research Laboratory of

Queensland University of Technology to evaluate the fire resistance of full scale, load

bearing steel stud wall assemblies. It included nine test specimens. One specimen was

tested at room temperature to determine its ultimate load bearing capacity while the

remaining eight specimens were exposed to the standard fire condition on one side

under a constant load to assess their fire performance.

This chapter presents the details of the experimental study into the thermal and

structural performance of the load bearing wall assemblies lined with single or dual

layers of plasterboard with or without cavity insulation. The insulations used were

glass, rock and cellulose fibres. Three new stud wall systems were built with the

insulation sandwiched between the plasterboards on both sides of the steel wall frame

instead of being placed in the cavity. Details of the results, including the temperature

and deflection profiles, measured during the tests are presented along with the stud

failure modes. Figures 6-1 (a) and (b) show the likely local and global stud failure

modes that could be encountered during the test.

Basic Section Flange Buckling

Web Buckling Distortional Buckling

Figure 6-1 (a): Basic Local Failure Modes


Flexural Buckling about Major Axis Flexural Buckling about Minor Axis

Torsional Buckling Lateral Torsional Buckling

Lateral Distortional Buckling

Figure 6-1 (b): Basic Global Failure Modes

6.2: Test Specimens

All the steel frames used in the large scale load bearing wall models were built to a

height of 2400 mm and a width of 2400 mm to represent a typical wall in a building.

All the studs and tracks used were fabricated from galvanized steel sheets having a

nominal base metal thickness of 1.15 mm and a minimum specified yield strength of

500 MPa. The frames were made of four vertical studs having 90 x 40 x 15 x 1.15 mm

lipped channel sections as shown in Figure 6-2. The studs were spaced at 600 mm

centres. Test frames were made by attaching the studs to the top and bottom tracks

made of 92 x 50 x 1.15 mm unlipped or plain channel sections using 12 mm long self

drilling wafer head screws.

The steel frames were lined on both sides by single or multiple layers of gypsum

plasterboards manufactured by Boral Plasterboard under the product name of Firestop.

The plasterboards supplied were 1200 mm in width by 2400 mm in length with a

thickness of 16 mm and mass of 13 kg/m2. The sheets were manufactured to the

requirements of Australian Standard AS/NZS 2588 – “Gypsum Plasterboard” (SA,

1998).


(a)

Stud Wall Frame

(b) (c) Stud Section (External Dimensions) Track Section

Figure 6-2: Test Wall Frame


(a) Isometric View

Butt Joints between Plasterboards

Butt Joints between Plasterboards

(b) Plan View Figure 6-3: Stud to Plasterboard Connections


The plasterboards were installed vertically on both sides of the steel frame to build the

single layer wall models. The vertical butt joints were located over the centre line of

stud flanges for proper fixing. The boards were attached by 25 mm long drill point

screws which require much less effort than needle point screws thus reducing the

chance of stud distortion. These screws were spaced at 200 mm centres along the

plasterboard edges and 300 mm centres along the intermediate studs in the field of the

plasterboard as shown in Figure 6-3(a).

The butt joints between the plasterboards on one side were offset in relation to the

corresponding joints on the other side, with the offset being equal to a single stud

spacing of 600 mm (see Figure 6-3(b)). Plasterboards ‘c’ and ‘d’ were connected to

studs only along one edge whereas the other plasterboards had multiple stud supports.

A minimum edge distance of 15 mm was maintained for all the screws from the

plasterboard edges. For wall models requiring double layers, the second layer or the

outer layer consisted of plasterboard sheets installed horizontally with the joint at

mid-height of the wall. The outer layer plasterboards were attached by 45 mm long

self-drilling bugle head screws spaced at 300 mm centres in the field of the

plasterboard and penetrating the studs. The exposed screw heads were given two coats

of joint compound. The joints were sealed with 50 mm wide perforated chamfered

edge joint reinforced paper tape and covered with two coats of joint compound as

shown in Figure 6-4.

Figure 6-4: Protection of Joints


Table 6-1 gives an overview of the nine load bearing test wall specimens used in this

study.

Table 6-1: Details of Test Specimen Configuration

Test

No.

Configuration Test Insulation Objective

1 Ambient None To determine the ultimate load bearing capacity of specimen at ambient temperature

2 Fire None To study the fire performance of a 1x1 LBW

3

Fire None To study the fire performance of a 2x2 LBW

4

Fire Glass Fibre (Cavity

Insulation)

To study the fire performance of a 2x2 LBW with glass fibre as cavity insulation

5

Fire Rock Fibre (Cavity

Insulation)

To study the fire performance of a 2x2 LBW with rock fibre as cavity insulation

6

Fire Cellulose Fibre

(Cavity Insulation)

To study the fire performance of a 2x2 LBW with cellulose fibre as cavity insulation

7

Fire Glass Fibre (External

Insulation)

To study the fire performance of a 2x2 LBW with glass fibre as external insulation

8

Fire Rock Fibre (External

Insulation)

To study the fire performance of a 2x2 LBW with rock fibre as external insulation

9


(External Insulation)

To study the fire performance of a 2x2 LBW with cellulose fibre as external insulation

Note:

LBW: Load Bearing Wall

1 x 1: Single layer of plasterboard on both sides

2 x 2: Two layers of plasterboard on both sides


6.3: Construction Details of Test Specimens


The test steel frame shown in Figure 6-2(a) was lined on both sides by a single layer

of plasterboard (1 x 1 assembly) covering the frame as shown in Figure 6-3.

Thermocouple wires were not installed as the specimen was tested for its ultimate

axial compression capacity at ambient temperature.


Construction of Test Specimen 2 was similar to the construction of Test Specimen 1

in all respects, except for the type of connection adopted at the top end of the studs. In

Test Specimen 2, a gap of 15 mm was left between the stud and the upper track as

shown in Figure 6-5.

16 mm Plasterboard

Figure 6-5: Stud to Track Connection at the Top

Screws were not used to connect the top end of the studs to the upper track. Instead

friction fit connections were adopted to allow for the vertical expansion of the studs

when exposed to elevated temperatures. Studs 2 and 4 had vertical plasterboard joints

on the fire side.

K type thermocouple wires were installed to measure the temperature variations

across the wall (over the plasterboard and steel surfaces) and along the stud lengths

(see Figure 6-6). Their locations on the wall are given in Section 6.4.6(B). The

thermocouple wires were attached to the hot flange, web and cold flange of the stud


by passing the hot junctions of the wire through small holes drilled in the steel

elements. These wire ends (hot junctions) were then pressed flat against the steel to

measure the surface temperature. The wires were then drawn to the ambient side

through tiny holes drilled in the unexposed plasterboard as shown in Figure 6-6(a). A

total of 50 thermocouple wires were installed, out of which 36 were used for

measuring the stud temperatures at three different levels (Figure 6-6b) and 14 were

used for measuring the temperatures on the plasterboard surfaces.

(a) Close up view of TC wires (b) View showing 36 TC wires attached attached to the flanges and web to the studs over 3 levels

(c) Fixing of Ambient Side Plasterboard




Test Specimen 3 was constructed with two layers of plasterboard on both sides of the

steel frame. The first layer was installed vertically while the second layer was

installed horizontally. Friction fit joints (similar to Test Specimen 2) were adopted at

the top end of the studs. Fifty six thermocouple wires were installed in the wall to

measure the temperature variations across the width of the wall and length of the

studs, when the wall specimen was exposed to furnace heat from one side. The

thermocouple wires were pulled across the width of the wall and onto the ambient

side through small holes drilled in the plasterboards. The holes in the second layer

(face layer) of the plasterboard were made to align exactly with the hole locations in

the first (base) layer so as to avoid any damage to the thermocouple wires (Figure 6-

7). Some of the thermocouple wires were bent with their hot junctions sandwiched

between the plasterboard surfaces to measure the interface temperature between the

plasterboards.

Thermocouple wires passed through aligned holes of the ambient side plasterboards

Face layer applied horizontally on ambient side

Base layer applied vertically on ambient side

Figure 6-7: Fixing of Face Plasterboard on the Ambient Side of Test Specimen 3


The construction of Test Specimen 4 was very similar to that of Test Specimen 3. The

only difference was in the use of cavity insulation. After fixing the two plasterboards

on the fire side along with their associated thermocouples the cavity in the wall

between the studs was filled with two layers of 50 mm thick glass fibre mats (with an

original nominal density of 13.88 kg/m3) compressed to 90 mm thickness (depth of

the cavity) giving a density of 15.42 kg/m3. Figures 6-8(a) and (b) show the


installation of glass fibre mats. The cavities of the individual studs and tracks were

also packed with the same insulation to avoid the formation of air pockets within the

wall cavity. Figure 6-8(c) shows the temporary positioning of the base layer

plasterboard on the ambient side to facilitate the passing of thermocouple wires

through holes drilled in it at appropriate places. After fixing the base layer

plasterboard, the face layer plasterboard on the ambient side was fixed in a manner

similar to Test Specimen 3. Fifty six thermocouple wires were used to measure the

thermal response of the wall model when subjected to fire from one side.

(a) (b)

Laying of Glass Fibre Mats in the Wall Cavity

(c) Temporary Position of Base Layer Plasterboard on the Ambient Side for Passing Thermocouple Wires through Aligned Holes

Figure 6-8: Construction of Test Specimen 4 Using Glass Fibre as Cavity Insulation



Construction of Test Specimen 5 was identical to that of Test Specimen 4 in all

respects, except for two layers of rock fibre each of 25 mm in thickness and density

100 kg/m3 being used as cavity insulation. Figure 6-9(a) shows the installation of rock

fibre mats in the wall cavity whereas Figure 6-9(b) shows the passing of thermocouple

wires through small holes drilled into the base layer plasterboard of the ambient side.

Once all the wires were passed through the holes the base layer was lowered (with the

thermocouple wires being gently pulled simultaneously) and fixed onto the steel

frame. The face layer on the ambient side was then positioned at right angles (applied

horizontally) over the base layer with the holes in the face layer aligned perfectly with

the pattern of holes in the base layer. The thermocouple wires were then passed

through the aligned holes of the face layer after which the face layer was lowered

gently and fixed onto the studs through the base layer.

(a) Laying of Rock Fibre Mats in the Wall Cavity



(b) Passing of Thermocouple Wires through the Base Layer Plasterboard on the Ambient Side



Test Specimen 6 was built similar to Test Specimens 4 and 5, but with cellulose fibre

used as cavity insulation. After fixing the two layers of plasterboard on the fire side

along with their associated thermocouple wires the cavity was fine sprayed with plain

water to just moisten the cavity facing surface of the plasterboard. This was quickly

followed by a wet spray of cellulose fibre. Spraying was stopped after the complete

filling of the wall cavity with an approximate insulation density if 100 – 110 kg/m3.

The ambient side plasterboards were then subsequently laid and fixed to complete the

specimen construction.


(a) Preparing the Cavity Surface (b) Spraying of Wet Cellulose Fibres For the Cellulose Spray into the Wall Cavity

(c) Cavity Filled Completely (d) Construction of Test Specimen 6 With Cellulose Fibres Completed with the Fixing of Ambient Side Plasterboard

Figure 6-10: Construction of Test Specimen 6 using Cellulose Fibres as Cavity Insulation


The construction of Test Specimen 7 required the insulation to be laid not within the

cavity as in the previous specimens but outside the cavity (referred from here on as

external insulation) and between the base and face layer plasterboards on both sides of

the wall. To achieve this, the plasterboard layer on the fire side was attached along

with its associated thermocouples to the steel frame in a manner similar to the

previous specimens. This was followed by fixing the base layer plasterboard of the

ambient side along with its thermocouples to the steel frame thus closing the wall

cavity. Before fixing this plasterboard, all the thermocouples attached to the base

layer plasterboard on the fire side were carefully passed on to the ambient side

through small holes drilled at appropriate locations in the base layer plasterboard on


the ambient side. After fixing the base layer plasterboard and closing the wall cavity,

13 mm plasterboard strips of 60 mm width were fixed to the base layer plasterboards

on either side along the periphery of the wall to generate a cavity for external

insulation. To increase the depth of the cavity a second layer of plasterboard strips of

the same width and thickness was mounted on the previous strip giving a total cavity

depth of 26 mm. This was followed by the placing of a single layer of 25 mm thick

glass fibre mat of density 13.88 kg/m3 in the cavity formed on either side of the wall

along with additional thermocouples to measure the temperatures on either side of the

insulation during the fire test. Finally the face layer plasterboards were fixed

horizontally on either side of the wall (sandwiching the insulation between the face

and base layer plasterboards) while taking care to pass all the thermocouple wires

onto the ambient side of the wall through holes in the ambient side plasterboard.

(a) Thermocouple Wires being Passed (b) View Showing Plasterboard Strips through the Base Layer Plasterboard Attached along the Border of the Base layer with the Glass Fibre Mat Installed in Position

Figure 6-11: Construction of Test Specimen 7 using Glass Fibres as External Insulation


The construction of Test Specimen 8 was identical to that of Test Specimen 7 in all

respects, except for a single layer of 25 mm thick rock fibre insulation of density 100

kg/m3 used as external insulation.


(a) Single Layer of Rock Fibre Insulation Laid as External Insulation over the Base Layer

(b) Fixing of the Face Layer Plasterboard over the External Insulation

Figure 6-12: Construction of Test Specimen 8 using Rock Fibres as External Insulation


The fixing of the base layer plasterboards to the steel frame and the strips along the

periphery to develop the cavity for external insulation was identical to that used in

Test Specimens 7 and 8. In this specimen 25 mm thick cubical spacers cut from

plasterboard strips were also positioned in the field of the cavity. This was done to


provide a firm support to the face layer plasterboard which was subsequently attached

after wet spraying the cavity with cellulose fibre insulation of density 100 – 110

kg/m3 (see Figures 13 (a) to (c)).

Spacer blocks

Wet spray of cellulose insulation

Border strips attached to base layer Plasterboard

(a) Spraying of Wet Cellulose as External Insulation on Ambient Side

(b) Spraying of Wet Cellulose (c) Fixing of the Face Layer Plasterboard as External Insulation on the Fire Side On the Fire Side

Figure 6-13: Construction of Test Specimen 9 using Cellulose Fibres as External Insulation

The outer layer plasterboards in Test Specimens 7, 8 and 9 were fixed to the steel

frame by 70 mm long plasterboard screws (with bugle heads) spaced at 300 mm

centres in the field of the plasterboard. As the screws used were not of the self drilling

type, it was necessary to pre-drill the studs before fixing the screws.


6.4: Test Set-up and Procedure

6.4.1: Furnace

A propane fired gas furnace was specifically designed to carry out the tests on the

wall specimens. The furnace has internal dimensions of 2.1 m width, 0.3 m depth and

2.4 m height. The front face of the furnace was left open thus exposing all the burners

(see Figures 6-14 (a) and (b)). The furnace was mounted on a carriage so that it could

roll on wheels (Figure 6-14(c)).

To start the test the carriage was moved forward to make contact with the frame

holding the test wall specimen, thereby completing the combustion chamber. On

starting the furnace the wall was exposed to heat from one side as desired.

The gas burners are nozzle mixing units with a high velocity, spinning, air flow,

creating a negative vortex at the refractory block mouth. When gas enters the vortex it

mixes rapidly producing intense combustion. Also the inverted parabolic shape of the

burner block port works with the vortex and pulls the flame flat on to the furnace wall

(Figure 6-14d). This protects the wall from any localised flame impingement and

ensures a more uniform distribution of the temperature over the wall surface, mostly

by radiation.

Furnace Specifications

1) Structural steel furnace shell lined with ceramic fibre insulating material.

2) Six pyronics model SW 3 infrared flat flame gas burners including ignition

pilot burners.

3) 7.5 KW, 415-volt centrifugal combustion air fan complete with air distribution

manifold.

4) Gas control safety train and six individual control gas trains.

5) Control panel consisting of fan stop, start, programmable temperature control,

oven temperature control and burner controls (see Figure 6-14e).


Gas Burner

Exhaust

Steel Shell

Ceramic Fibre Insulation

(a) Front View of Furnace before Ignition

(b) Front View of Furnace before and after Ignition

Figures 6-14: Details of Furnace Operation and Components


Specimen in Loading Frame Furnace

Control

Data Logger

(c) Side View Showing the Furnace on Wheels

(d) View of Radial Flame (e) Control Panel (Back View of Furnace)

Furnace

Thermocouple

Burner

Viewing Port 100mm x 200mm

450 mm

200mm

200 mm

(f) Rear wall of Furnace Housing the Thermocouples and Viewing Ports

Figures 6-14: Details of Furnace Operation and Components


Furnace Temperature:

The furnace was designed to deliver heat in accordance with AS 1530.4 (SA, 2005) as

given by the following equation

Tt – To = 345 log 10 (8t+1)

Where, t = Elapsed time in minutes

Tt = Furnace temperature (0C) at time t

To is the ambient temperature (0C) at the start of the test.

Furnace Pressure

The specimen holder containing the wall assembly was sealed against the furnace in

order to maintain the furnace pressure by at least 2 Pascal greater than the atmospheric

pressure over the top two thirds of the wall specimen. The positive pressure helped in

preventing the drawing of outside cold air into the combustion chamber.

Observation Ports

Four observation ports were provided on the rear side of the furnace as shown in

Figure 6-14(f) for observing the structural response of test walls.

Instrumentation Ports

Eight ports were provided on the rear wall to facilitate the introduction of

thermocouples to record the furnace temperatures during the test. Three ports were

also provided along the side (top, centre and bottom) to record the internal pressure

(see Figure 6-14(f)).

6.4.2: Compression Loading Frame

Loading Arrangement Used in Ambient and Fire Tests:

The loading frame was specially designed to load the individual studs of a wall

specimen in compression directly from the bottom side. It consisted of two columns

firmly bolted to the ground and a universal beam (UB) connecting the two columns to

form an ‘H’ shaped portal frame. A second universal beam was bolted to the floor.

Four jacks each of 45 kN capacity were mounted on this beam at a spacing of 600

mm. The shafts of the jack were co-axially guided through a hollow sleeve running


across a rectangular hollow section (RHS) attached between the two columns and

fixed parallel to the universal beam as shown in Figure 6-15.

Figure 6-15: Loading Frame

Test Specimen

RHS

Jack

Bottom Universal

Beam

(a) Loading of Each Stud in the Specimen using Jacks

Figure 6-16: Loading Arrangement


Bottom Track of wall

Gap to allow for stud elongation

Cement Plasterboard

Loading Plate

Sliding Plate

RHS

Jack

(b) Close up View of Loading Assembly

Figure 6-16: Loading Arrangement

The RHS supporting the hollow sleeve ensured a vertical movement of the jack.

Loading plates with collars to house the shafts coming out of the RHS were mounted

on top of each jack. The loading frame was built such that, when the test specimen

was mounted into the frame, the bottom track rested on the four loading plates and the

upper track was pressing against a cement board clamped firmly to the underside of

the top UB (the cement board was used to minimize the heat loss from the upper track

of the test specimen to the top UB). The test specimen was mounted in a manner to

ensure that the centroids of the studs aligned with the centroids of the loading plates.

This was made possible as the spacing of the jacks was identical to the spacing of the

studs. A special arrangement was also made using a sliding plate to facilitate

movement of any individual jack along with its shaft and sleeve to the extent of 20

mm on either side so as to achieve a better accuracy in lining up of the centroids of

the loading plate and the studs.

All the jacks were connected to a single hydraulic pump as the aim was to determine

the failure load of the wall specimen. A load cell was attached to the pump to obtain

directly the load being delivered to the studs by the jacks. The use of a single pump

ensured equal loading on all the studs as the same hydraulic pressure operated all the

jacks (see Figure 6-17).


(a) Hydraulic Pump

connecting the Four Jacks

(b) Excess Load Release Valve

Figure 6-17: Hydraulic Pump and its Connections

To test the wall specimens subjected to fire, it was necessary to prevent the heat loss

from the bottom track into the loading plates. This was achieved by inserting cement

plasterboard pieces between the bottom track and the loading plate. The plasterboard

pieces were cut to the same size as the loading plates.

6.4.3: Ambient Temperature Test Procedure

Figure 6-18 shows the test specimen installed into the loading frame. The studs were

centred over the individual jacks and the wall was checked for its verticality using

spirit levels. After proper positioning of the wall, the top track was fastened to the top

beam using G - clamps on either side. This was done to retain the correct positioning

of the wall during testing. The studs were loaded initially to about 5 to 10 kN and then

unloaded. This was done twice to remove any residual strains and initial slackness

which may be present in the system during the assembly of the wall specimen. All


four studs were then reloaded simultaneously in increments of 2 kN. Load and

displacement readings were recorded by the Edcar software at the end of each load

increment. The specimen was assumed to have failed when the oil pressure in the

jacks could not be maintained. This was also confirmed via the Edcar load-

displacement graph which showed rapid load reductions (unloading).

Portal Frame

Wooden beam to mount LVDTs

LVDT

Figure 6-18: Test Set-up for Ambient Temperature Test

6.4.4: Instrumentation for Ambient Temperature Tests

To measure the axial shortening of the studs four Linear Variable Displacement

Transducers (LVDT) were used with each LVDT placed under the loading plate and

as close as possible to the stud as shown in Figures 6-19 (a).

Eight LVDTs were used to measure the out-of-plane movements of the wall

specimen. The transducers were attached to the wooden beams in front of the

specimen as shown in Figure 6-19(c) and were placed at 0.25H, 0.50H and 0.75H

along the height (H) of the two central studs and only at mid-height for the outer studs

as shown in Figure 6-18.


Test Specimen

LVDT

JACK

(a) LVDTs Measuring Axial Shortening

(b) LVDTs (c) LVDTs Measuring Measuring Axial Shortening out-of-plane Deflection

Figure 6-19: LVDTs Used in the Measurement of Axial Shortening and Out-of-plane Deflection of Test Specimen Wall

6.4.5: Elevated Temperature Test Procedure

Test specimens were installed in the loading frame in the same manner as for the

ambient temperature test specimen. The furnace was then rolled forward towards the

wall specimen to close the gap between them and thus complete the combustion

chamber of the furnace with the wall forming the fourth side of the chamber facing


the burners. This arrangement ensured that only one face of the test specimen was

exposed to elevated temperatures at the start of testing. The width of the wall was

designed to be less than the width of the furnace opening by 20 mm such that there

would be a gap of 10 mm on either side of the wall, to ensure free vertical edges. The

gap was then packed with Iso wool, a non-restraining and non-combustible mineral

fibre such that the lateral displacement of the wall was not restricted due to frictional

forces. An axial compression load of 15 kN was applied gradually to each stud at a

constant rate by the hydraulic jacks. This load was based on a load ratio of 0.2, i.e. 0.2

times the ultimate capacity of each stud at ambient temperature obtained from Test 1.

The load was held constant at room temperature for about 45 minutes before the

furnace was started. The load of 15 kN per stud was maintained throughout the fire

endurance test. This allowed free vertical expansion of the wall when exposed to

elevated temperatures. During the fire test, the furnace temperature was regulated

such that the average temperature recorded by the control thermocouples inside the

furnace followed the standard cellulosic temperature-time curve in accordance with

AS 1530.4 (SA, 2005). During the fire test the vertical and lateral displacements of

the wall, the temperature readings from all the thermocouples and the furnace pressure

readings were taken at intervals of 1 minute. The test was stopped immediately

following the failure of the wall. The time to failure was then recorded.

6.4.6: Instrumentation for Elevated Temperature Tests

A) To measure displacements

To measure the axial shortening of the studs the location of the transducers was

identical to that adopted for the ambient temperature test. For the out-of-plane

movements of the wall specimen the transducers were placed at 0.25H, 0.50H and

0.75H along the height (H) of all the four studs.

B) To measure temperatures

K type thermocouples were used to measure the temperature development across the

wall specimens. The stud temperatures were measured at three levels, namely at 0.25

H, 0.50 H and 0.75 H (where ‘H’ is the height of the Test Specimen).


Fire Exposed Side

Unexposed Side

(a) Thermocouple Locations for 1x1 LBW Specimen

Fire Exposed Side

Unexposed Side

(b) Thermocouple Locations for 2x2 LBW Specimens with and without Cavity Insulation

Fire Exposed Side

Unexposed Side

(c) Thermocouple Locations for 2x2 Wall Specimens with External Insulation

Studs

Horizontal joint of face layer plasterboard

Thermocouples

(d) Positions of Thermocouples to Measure the Average Temperature Rise on the Ambient Surface of the Test Specimen

Figure 6-20: Thermocouple Locations for Load Bearing Wall Specimens


At each level three thermocouples were attached per stud to measure the temperatures

of the hot flange, web and cold flange thus giving a total of nine thermocouples per

stud (over the three levels) and 36 thermocouples per frame (as the frame is made up

of four studs). These thermocouples allowed the determination of average stud

temperature, temperature gradient across the stud cross-section and also along the stud

length.

Additional thermocouples were attached at the mid-height of the assembly between

the studs to measure the temperatures on the plasterboard surfaces facing the cavity

and also on the exposed face of the wall to measure the temperature of the

plasterboard surface subjected to fire thus giving a total of 45 thermocouples (36 + 9)

for a 1x1 LBW Specimen as shown in Figure 6-20 (a).

For 2x2 LBW Specimens with or without cavity insulation, additional thermocouples

were installed between the plasterboard surfaces on either side at mid-height as shown

in Figure 6-20 (b) giving a total of 51 thermocouples (36 + 9 + 6) to measure the

temperature variations across the test specimen. For 2x2 LBW Specimens using

external insulation six more thermocouples were installed to measure the temperature

across the insulation layers at mid-height thus giving a total of 57 thermocouples (36

+ 9 + 6 + 6) as shown in Figure 6-20 (c)

To measure the average temperature on the unexposed face of the wall, five

thermocouples were positioned on the unexposed face, 1 at the centre of the area and

one at the centre of each quarter section as mentioned in AS 1530.4 (SA, 2005). The

temperature measured by these thermocouples indicated the heat penetration across

the specimens (see Figure 6-20 (d)).

To measure the temperature at various other points on the ambient face an infrared

gun was used (Figure 6-21). Figure 6-22 shows the complete instrumentation of a test

specimen before the fire test.


Figure 6-21: Infrared Gun Used for the Measurement of Ambient Side Temperatures

.

Isowool Insulation along the periphery to close the gap between the specimen and the furnace

LVDTs

Pressure Transducer

EDCAR Data

Logger Thermocouple wires

Jacks

Figure 6-22: Test Specimen Complete with all its Instrumentation Ready for Fire Test

Eight K – type furnace thermocouples were symmetrically placed inside the furnace

chamber within a vertical plane 100 mm from the exposed surface of the specimen to

record the temperature of the furnace (see Figure 6.14 (f)).


6.5: Observations and Results


6.5.1.1) Visual Observations and Specimen Behaviour

The wall specimen showed no visible signs of deformation up to a load of 52 kN.

Beyond this load the web elements of Studs 1 and 4 which were the only visible parts

of the steel frame showed local buckling waves developing in the web portion. This

local deformation progressed with increasing load. When the load approached 79

kN/stud there was a sharp noise and the specimen failed. The load dropped rapidly as

the oil pressure could not be maintained. Lateral movement of the wall specimen was

not visible at any stage of the test. The failure was due to local buckling of the flange

and web elements at the base as shown in Figure 6-23.

(a) Local Buckling at the base of Stud 1 (b) Local Buckling at the base of Stud 4

Figure 6-23: Failure of Test Specimen 1

The gypsum plasterboards showed no damage and were seen to be successful in

effectively restraining the studs from torsional buckling and flexural buckling about

the minor axis. The screws connecting the plasterboards to the studs were seen to have

pulled through the plasterboard at the base close to the locally buckled stud sections.

Axial deformations in the range of 13 mm to 16 mm were noted in the studs as seen in

Figure 6-24.


0

10

20

30

40

50

60

70

80

90

0 2 4 6 8 10 12 14 16 18

Axial Shortening (mm)

Lo

ad (

kN)

Stud 1 Stud 2 Stud 3 Stud 4

Figure 6-24: Load Vs Axial Deformation - Profiles of Test Specimen 1 at Ambient Temperature

The wall failed at a load of 79 kN/stud (total load of 316 kN on the frame) by the local

crushing of the stud channels at the base near the loading plates.

6.5.2: Test Specimen 2 (1x1 LBW without insulation)

6.5.2.1) Visual Observations and Specimen Behaviour

The specimen was subjected to the standard time-temperature heating regime in the

furnace. During the test the Edcar software crashed for a period of 8 minutes from 9 to

17 minutes from the start of the test resulting in the loss of readings. Also it was

uncertain whether the applied load of 15 kN was maintained after 9 minutes of test.

The temperature time graphs could be plotted accurately. However the portion of the

graphs between 9th minute to 17th minute was unavailable due to the failure of the

software.

The specimen showed no signs of lateral displacement during the initial application of

the compression of 15 kN load. After 3 minutes of starting the furnace smoke was

seen coming out from the top of the wall specimen (Figure 6-25(a)). This was

probably due to the burning of the plasterboard paper on the exposed surface. At the

end of 11 minutes thick smoke and steam were seen to escape from the outer edges


from the top of the wall. The presence of steam in the mixture of escaping gases was

confirmed by the heavy condensation of steam into water on the bottom flange and

web of the top UB of the loading frame (see Figures 6-25(b) and (c)).

(a) Smoke and Steam Escaping from the Top Side of Test Specimen 2

(b) Condensation of Steam (c) Condensation of Steam on the On the Bottom Side of Top UB Web of the Top UB

Figure 6-25: Fire Performance Test of Specimen 2

By 32 minutes the lateral displacement or bowing of the wall towards the furnace was

prominently noticeable. At the end of 40 minutes soft crackling sounds were heard

from within the wall. These crackling sounds were probably the result of energy being

released by the propagation of shrinkage cracks in the plasterboard exposed to fire.


The wall failed to support the applied load (Structural Failure) at the end of 53

minutes and the test was stopped. On rolling the furnace back to expose the

plasterboard on the fire side, it was noticed that the exposed plasterboard strip over

Stud 4 (Figure 6-28(a)) had fallen off as it was attached only along one edge.

Shrinkage of the exposed plasterboard had caused it to detach from the fasteners,

opening the joints and exposing the studs as shown in Figure 6-26(b).

(a) Detachment of Plasterboards (b) Close up of Open Joint along the joint

Figure 6-26: Detachment and Opening of Plasterboard joints Caused by Shrinkage

The joints opened up from 20 mm at the base to about 35 mm at the top of the stud,

indicating the greater severity of the plasterboard shrinkage at the top, caused

probably due to the higher temperatures in the chamber at the top due to upward

movement of hot air. Due to the opening of joints the studs immediately behind the

joints not only lost their lateral support from the plasterboard on the fire side but were

also severely affected by higher temperature. As seen in Figure 6-27 Studs 2 and 4 of

the specimen were seen to be more affected than Studs 1 and 3 as they had vertical

plasterboard joints on the fire side. Time-temperature graphs of Studs 2 and 4 (see

Figures 6-30 (b) and (d)) clearly show a much higher temperature of the hot flange at

failure when compared to the hot flanges of Studs 1 and 3 (Figures 6-30 (a) and (c)).


(a) Front View of Test Specimen 2 after Removing the Exposed Plasterboards

Stud 4

Stud 3

Stud 2 Stud 1

(b) Side View Showing Detachment of Studs from the Plasterboards by Screw Pull-out

Figure 6-27: Test Specimen 2 after Removing the Exposed Plasterboard Layer


The collapse of exposed plasterboard over Stud 4, on the fire side towards the end of

the test led to a rapid increase in the temperature of the stud resulting in the plastic

deformation of the top portion of the stud as shown in Figure 6-28 (b).

Stud 4 Stud 4

(a) Detachment and Collapse of (b) Stud Failure Exposed Plasterboard

Figure 6-28: Stud Failure Initiated by Plasterboard Fall-off

Visual inspection revealed that Stud 4 was the first to fail followed by Stud 2. Studs 2

and 4 with vertical joints of plasterboard on the fire side underwent larger thermal

bowing deformations due to higher thermal gradients across the cross-section. This

caused them to separate from the ambient side plasterboard, pulling the screws

inwards as shown in Figure 6-27(b). The reduced lateral support on either side of the

studs coupled with decreasing mechanical properties at elevated temperatures caused

the studs to undergo flexural torsional buckling about the minor axis.

The friction fit connections provided at the top end of each stud, led to a greater

degree of instability in the frame as they allowed the top end to shift in the plane of

the wall during the thermal deformations. Also at the beginning of the test when the

specimen was loaded at room temperature, the gaps provided in the joints (for the

purpose of allowing free thermal expansion of the studs when exposed to fire) closed

up, thus defeating the main purpose. Figure 6-31(a) shows the axial deformation of

the studs when subjected to loading at ambient temperature. The large variations in

the profiles of the individual studs were probably on account of the closure of the

unequal gaps in the joints before the studs underwent actual axial shortening due to

the applied loads.


The plasterboard on the ambient side was seen to be in good condition with the paper

on the cavity facing surface burnt only in few locations, thus maintaining the integrity

of the wall. Heat penetration failure (Insulation failure) was also not detected as the

temperature on the ambient face of the unexposed plasterboard was much lower than

the standard failure criteria (maximum average temperature of 1400C above the

ambient or a maximum temperature of 1800C at any location on the ambient surface)

until the end of the test as recommended by AS 1530.4 (SA, 2005).

The cause of failure of the wall specimen could be attributed to the structural failure

of the frame precipitated by the opening of plasterboard joints and partial collapse of

plasterboard on the fire side.


The time-temperature profile of the furnace and the exposed face of the wall (FS)

were seen to follow the standard time-temperature curve defined by AS 1530.4 (SA,

2005)

a) Plasterboard Surfaces

1) Average temperature on the cavity facing surface of the exposed Plasterboard

(Pb1-Cav)

The temperature on this surface developed in three phases. In the first phase the

temperature was quick to rise from about 3 minutes to approximately 1000C by the

end of 5 minutes. In the second phase, which lasted from 5 minutes to 20 minutes, the

temperature on the plasterboard surface was maintained constant at about 1000C due

to the energy consumed in converting the free and chemically bound water present in

the plasterboard into steam. The data logger failed due to some glitch in the software

during the period from 9 to 17 minutes and no readings were taken in this time period.

The beginning and end of the plateau (second phase) can be observed in Figure 6-29.

In the third phase beyond 20 minutes the temperature of the plasterboard surface

reached 4000C by 40 minutes. Towards the end of the test the temperature had

reached 5000C. A sharp rise in temperature beyond 5000C indicates the breaching of

the exposed plasterboard.



2) Average temperature on the cavity facing surface of the ambient Plasterboard

(Pb2-Cav)

The temperature profile of this surface followed very closely the profile of Pb1-Cav

until the end of second phase. A maximum temperature difference of about 1100C

across the cavity was observed at the end of 30 minutes. Beyond this time the

temperature difference was seen to gradually decrease (due to the degradation of the

exposed plasterboard) with the Pb2-Cav profile almost merging with the Pb1-Cav

profile towards the end of the test.

3) Average temperature on the ambient side of unexposed plasterboard 2

The temperature on this surface was below 800C during the test.

0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 6

Time (min)

Tem

per

atu

re (

oC

)

0

AS 1530.4 Furnace FS Pb1-Cav Pb2-Cav Amb


b) Steel Surfaces

Figures 6-30 (a) to (d) show the time-temperature profiles of Studs 1 to 4,

respectively, during the fire test. Time-temperature profiles of the studs develop in

three stages in phase with the Pb1-Cav temperature, as the temperature of the hot

flanges in the studs primarily depend up on the heat conducted through the Pb1-Cav

surface. The temperature rise of Studs 2 and 4 is observed to be more rapid than that

of Studs 1 and 3 as a result of the plasterboard joints present along their lengths (see

Figure 6-3 (b) for joint locations).

Figure 6-31 (a) shows the axial deformations measured during the loading of the studs

at ambient temperature. The initial readings up to 4 kN load is primarily due to the

closure of the expansion gap provided in the friction fit end connections of the studs

during the construction of the wall. Axial shortening of approximately 1 to 1.5 mm

was observed when the studs were fully loaded. Figure 6-31 (b) shows the axial

deformations measured during the fire test. Axial deformation profiles of Studs 2 and

4 reverse suddenly at about 51 minutes, suggesting the sudden buckling of these studs

at this stage. Figure 6-28 (b) shows the plastic deformation and local buckling in Stud

4 and Figure 6-27 (b) shows the global buckling of Stud 2, with the plasterboard

screws pulled through the ambient side plasterboard. A maximum lateral deformation

of approximately 28 mm was observed at the mid-height of the wall in line with Stud

2 as can be seen from Figure 6-32.

0

100

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30 35 40 45 50 55 60

Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 Furnace Hot FlangeWeb Cold Flange Failure Time

(a) Time-Temperature Profiles across Stud 1

Figure 6-30: Time-Temperature Profiles across Studs 1 to 4 of Test Specimen 2


0

100

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30 35 40 45 50 55 60

Time (min)

Tem

per

atu

re (

oC

)

AS 1350.4 Furnace Hot Flange

Web Cold Flange Failure Time

(b) Time-Temperature Profiles across Stud 2 (A Vertical Plasterboard Joint Runs along the Length of This Stud)

0

100

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30 35 40 45 50 55 60

Time (min)

Te

mp

era

ture

(oC

)



(c) Time-Temperature Profiles across Stud 3



0

100

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30 35 40 45 50 55 60

Time (min)

Tem

per

atu

re (

oC

)



(d) Time-Temperature Profiles across Stud 4 (A Vertical Plasterboard Joint Runs along the Length of This Stud)


-14

-12

-10

-8

-6

-4

-2

0

2

0 2 4 6 8 10 12 14 16Load (kN)

Def

orm

atio

n (

mm

)


(a) Axial Deformation -Load Profiles of Studs at Ambient Temperature

Figure 6-31: Axial Deformation Plots for Studs of Test Specimen 2



-16

-14

-12

-10

-8

-6

-4

-2

0

0 10 20 30 40 50 6

Time (min)

Def

orm

atio

n (

mm

)

0


(b) Axial Deformation -Time Profiles of Studs at Elevated Temperatures


-35

-30

-25

-20

-15

-10

-5

0

0 10 20 30 40 50 6Time (min)

De

fle

cti

on

(m

m)

0

Stud 2 Stud 3 Stud 4

Figure 6-32: Lateral Deflection -Time Profiles of Test Specimen 2 at Mid-Height

6.5.2.3) Specimen Report

a) Test Specimen Number: 2

b) Date of Test: 23/06/07

c) Description of Specimen: Large scale load bearing wall comprising lipped steel

channels (90 x 40 x 15 x 1.15 mm) lined on both sides by single layer Gypsum

(FireSTOP) plasterboard 16 mm thick.

d) Overall Thickness of Wall: 122 mm

e) Severity of Test: 100%

The severity of fire exposure in a test is determined by comparison of the area under

the curve of the mean measured furnace temperature with the area under the standard

ISO 834 curve for the same period.

f) Specimen Temperature:

The average temperature of the unexposed surface of the test specimen towards the

end of the test was 700C indicating a rise of 570C above the ambient temperature of

130C. The maximum temperature of the unexposed surface at that time was 850C. The

maximum temperature of the stud hot flange was in the range of 500 to 6000C with a

temperature difference of 100 to 2000C across the stud depth.

g) Specimen Behaviour:

Most of the fire side plasterboard 1 remained attached to the steel frame until the

failure of test specimen. The lateral deflection of the test specimen was towards the

furnace and the maximum deflection at mid-height of the wall just prior to failure was

28 mm. The total thermal expansion of the studs during the fire test was in the range

of 5 to 7.5 mm.

h) Failure Criterion:

The test specimen was deemed to have failed at approximately 53 minutes from the

start of the test when the specimen could no longer sustain the applied load.


The failure of studs occurred due to the opening of plasterboard joints and partial

collapse of fireside plasterboard. Higher stud temperatures (i.e. reduced mechanical

properties) and reduced lateral support led to the studs failing by local and overall

buckling modes. The presence of plasterboard joints along the stud height affected the

behavior and failure of studs.

Stud 4 failed by local buckling at a hot flange temperature of 5500C

The value of the actual load acting on the wall specimen cannot be stated as the load

counter was reset during the experiment when the software failed temporarily.

i) Performance:

Performance observed in respect of the following criteria:

Structural adequacy - Failure at 53 minutes

Integrity - No failure at 53 minutes

Insulation - No failure at 53 minutes


Test Specimens 3 (without insulation), 4, 5, 6 (cavity insulated) and 7, 8, 9

(Externally insulated)

All the load bearing wall specimens tested for fire resistance showed similar initial

response during the test. They showed hardly any signs of stress when subjected to a

total axial load of 60 kN (15 kN/stud, giving a load ratio of 0.2). None of the

specimens showed any lateral deformation. Upon starting the furnace, smoke and

steam was seen to come out from the periphery of the specimen at the end of three to

four minutes. The smoke would indicate the burning of the paper on the exposed face

of the external plasterboard, and the steam was due to the escape of moisture (both

free and chemically bound) from the plasterboards. The presence of steam could be

easily noted as it would condense on the inner sides of the loading frame producing

streaks of water running down the column faces. The specimens would display

periods of steady burning with little or no smoke or steam. This would happen after

the complete burning of the paper and the complete conversion of water into steam

from the plasterboard. The smoke and steam would reappear with subsequent layers

of plasterboard heating up. There were periods of thick smoke ensuing continuously

from the specimens for almost 30 to 45 minutes. This would probably indicate the

burning of the insulating material used in the walls. Amongst the three insulations

used, cellulose fibre was seen to produce the maximum smoke and rock fibre the

minimum.

Lateral deflections were visible in the cavity insulated specimens after about 70

minutes of fire test. In the case of externally insulated specimens, the lateral

deflections became noticeable only towards the end of the test. In both cases, the

deflection of the wall was initially towards the furnace. Near the end of the test, a

reversal of lateral deflection was observed for all the specimens forcing the wall to

bow in the outward direction. Maintaining the applied load on the wall was difficult at

this stage with the hand pump controlling the jacks being operated more frequently.

The failure was sudden in all the specimens with the load quickly dropping off with

the studs buckling in the outward direction cracking the plasterboards on the ambient

side.


The ambient surface of each wall specimen recorded values well below the insulation

failure temperature throughout the test. The failure of the walls in every test was due

to the structural failure of the studs.


6.5.3.1) Visual Observations

The specimen was exposed to the furnace heat for a period of 112 minutes. The test

was stopped when the specimen could no longer maintain the applied load. Visual

inspection of the fire tested specimen reveled that the plasterboards 1 and 2 (exposed

plasterboards) though severely calcined were still intact offering protection to the

studs. The screws connecting the plasterboards to the studs were seen to have been

pulled in through the plasterboard thickness at the top 1/3rd portion of the wall.

On the removal of exposed plasterboards (Pb1 and Pb2) it was noticed that the studs

had been laterally displaced at the top end. The friction fit joints provided at the top

end of each stud had failed to prevent the slipping of the studs in the lateral direction

at elevated temperatures. This bending of the studs about the minor axis near the top

portion of the wall caused the screws to pull out from the plasterboard body.

The central studs also displayed distortional buckling in the top 1/3rd portion of their

lengths. The ambient side plasterboards (Pb3 and Pb4) were seen to be in a fairly

good condition. The unexposed surface of the specimen showed no visible signs of

wall failure.


a) Plasterboard Surfaces



The temperature on this surface was seen to follow three phases of development. The

first phase was highlighted by a quick rise in temperature from the ambient

temperature at about 2 minutes to approximately 1000C by the end of 5 minutes. This

was followed by a second phase, a plateau during which with the temperature was


constant at 1000C until 16 minutes. During the second phase the free and chemically

bound water in the plasterboard was expelled out in the form of steam. Beyond 16

minutes the third phase started with the temperature rising sharply from 1000C to

8000C by the end of 83 minutes. The temperature gradient was seen to be almost

constant until this point. Beyond 83 minutes the temperature gradient decreased

slightly and the curve seemed to run parallel to the fire curve maintaining a difference

of approximately 1500C to 2000C until the end of the test.

Plasterboard 3 intact although slightly damaged at screw locations due to pull out caused by stud distortion in the top portion

Distortional and flexural buckling @ minor axis in the top third portion for studs 1, 2 and 3

Stud 1

Stud 2

Stud 3

Stud 4

(a) Side View of Test Specimen 3 after Removing the Exposed Plasterboards

Figure 6-33: Close up of Test Specimen 3 after Removing the Exposed Plasterboards


Stud 3 Stud 2

(b)

Figure 6-33: Test Specimen 3 after Removing the Exposed Plasterboards

(a) Stud 1 (b) Stud 2

(c) Stud 3 (d) Stud 4

Figure 6-34: Studs of Test Specimen 3 after the Fire Test


Feng et al. (2005) observed in his experiments on load bearing wall panels that about

25% of the chemically bound water in the plasterboard which was not released during

the second phase of the graph was released at about 7000C giving a second plateau in

the time-temperature profile of the plasterboard. However, this was not observed in

the current experiment. The temperature of the interface (Pb1-Pb2) showed no change

in gradient as it crossed the 7000C mark. Also the slight drop in gradient beyond 83

minutes could be the result of a balance being achieved between the inflow and

outflow of heat across the thickness of the plasterboard, thus maintaining a

temperature difference of approximately 2000C across the 16 mm thick FireSTOP

gypsum plasterboard.

The external plasterboard appeared to be intact until the end of the fire test as the

temperature (between the fire curve and the Pb1-Pb2 interface) was maintained until

the failure of the frame.

ii) Average temperature on the cavity facing surface of the exposed plasterboard

2 (Pb2-Cav)

The temperature on this surface was seen to increase from the ambient at about 5

minutes to approximately 800C by the end of 13 minutes representing the end of the

first phase of the curve. The second phase constituting the plateau lasted until 55

minutes with the temperature hovering about 1000C, beyond which the third phase

started with a sharp rise in temperature. The temperature crossed 4000C by about 90

minutes and 5000C by about 110 minutes. During the third phase the temperature

gradient was fairly constant and maintained a temperature difference of approximately

4000C with the Pb1-Pb2 interface temperature graph which served as the fire curve for

the Pb2-Cav surface. This meant a temperature difference of 4000C across the

thickness of plasterboard 2 (i.e. the second layer on the fire side). The maintenance of

this temperature difference until the end of the test indicates the continued integrity of

plasterboard 2 until the failure of the specimen.


plasterboard 3 (Pb3-Cav)

The temperature profile of this surface coincided almost identically with that of the

Pb2-Cav surface for about 60 minutes. This was probably due to the almost


instantaneous transmission of heat across the cavity by radiation. Beyond 60 minutes

the temperature increased rapidly to 4000C in 102 minutes. The temperature growth

rate was seen to be linear and lagged behind the Pb2-Cav surface profile by

approximately 500C -1000C until the end of the test.

iv) Average temperature on the ambient side of unexposed plasterboard 3

(Pb3-Pb4)

The temperature of this interface remained below 1000C until the end of the test.

v) Average temperature on the ambient side of unexposed plasterboard 4

The temperature on the unexposed face of the wall was below 700C until the end of

the test.

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Tem

per

atu

re (

oC

)


(a): Average Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 3



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Tem

per

atu

re (

oC

)

FS-L FS-M FS-R Pb1-Pb2-L Pb1-Pb2-M

Pb1-Pb2-R Pb2-Cav-L Pb2-Cav-M Pb2-Cav-R Pb3-Cav-L

Pb3-Cav-M Pb3-Cav-R Pb3-Pb4-L Pb3-Pb4-M Pb3-Pb4-R

(b) Time-Temperature Profiles across the left, middle and right sections of

Plasterboard Surfaces in Test Specimen 3


b) Steel Surfaces

The studs were well protected from the furnace heat by two layers of plasterboard on

the fire side. Figure 6-36(a) shows the temperature profiles of the 6 thermocouples

reading the hot flange temperatures of the central studs. The temperatures of the

middle portion of the studs are seen to be higher than the top and bottom level

temperatures. This could be due to the thermal bowing of the wall panels towards the

furnace bringing their central portions closer to the furnace burners, thus causing the

central portion of the wall to heat up faster than the top and bottom levels. All

thermocouples except the one at the middle of Stud 2 are seen to fall in a very narrow

band signifying an almost constant temperature distribution over the length of the

studs.

The temperature profiles were seen to develop in three phases. The first phase saw the

temperature increase from the ambient at about 5 minutes to approximately 900C by

20 minutes, which was followed by the second phase of almost constant temperature

of 1000C for about 40 minutes. In the third phase the temperatures increased steadily

until the failure of the specimen. The temperature growth in the hot flanges


corresponded well with that of the Pb2-Cav surface profile with the only exception

being the thermocouple located at the centre of Stud 2, which recorded higher

temperatures. Towards the end of the test at about 112 minutes the lower and middle

portions of Stud 2 recorded a steep increase in temperature, which coincided with the

suspected breach of plasterboard 2 on the fire side at about the same time.

Figures 6-36 (b) and (c) show the temperature profiles of the webs and cold flanges of

the central studs measured at six locations. The plateau in the second phase extended

until 50 minutes for the webs and 60 minutes for the cold flanges, beyond which the

third phase included a steady temperature rise at almost the same rate (only the lower

cold flange thermocouple on Stud 2 deviated from the group recording higher

temperatures throughout).

Figure 6-37 shows the temperature profiles of the six thermocouples located at mid-

height of Studs 2 and 3 measuring the hot flange, web and cold flange temperatures

for each stud thus giving the temperature variation across the depth of the central

studs during the test. The first two phases demonstrate an almost uniform temperature

gradient across the central studs. The third phase shows a temperature difference

ranging from 1000C to 1700C (i.e. hot flange temperature-cold flange temperature)

across Stud 2 and about 800C to 1000C across that of Stud 3 until the end of the test.

0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Tem

per

atu

re (

oC

)

S2-T-HF S3-T-HF S2-M-HF S3-M-HF S2-L-HF S3-L-HF

(a) Time-Temperature Profiles on Hot Flange Surfaces of Central Studs in Test Specimen 3



0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Te

mp

era

ture

(oC

)

S2-T-W S3-T-W S2-M-W S3-M-W S2-L-W S3-L-W

(b) Time-Temperature Profiles on Web Surfaces of Central Studs in Test Specimen 3

0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Tem

per

atu

re (

oC

)

S2-T-CF S3-T-CF S2-M-CF S3-M-CF S2-L-CF S3-L-CF

(d) Time-Temperature Profiles on Cold Flange Surfaces of Central Studs in Test Specimen 3



0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Te

mp

era

ture

(oC

)

S2-M-HF S3-M-HF S2-M-W S3-M-W S2-M-CF S3-M-CF

Figure 6-37: Time-Temperature Profiles across Central Studs at Mid-height in Test Specimen 3

Note:

S2/3-T-HF/W/CF: Time-temperature profile followed by the hot flange/web/cold flange of Stud No.2/3 at the top level

S2/3-M-HF/W/CF: Time-temperature profile followed by the hot flange/web/cold flange of Stud No.2/3 at mid-height

S2/3-L-HF/W/CF: Time-temperature profile followed by the hot flange/web/cold flange of Stud No.2/3 at lower level

6.5.3.3) Behaviour of Specimen

The loading of the specimen at ambient temperature resulted in an initial displacement

of the studs in the axial direction caused by the closure of the 15 mm gap in the

friction fit connection between the studs and the top track. As the studs were not

screw connected with the top track, the gap closed immediately on the initial

application of the load. On further application of load to 15 kN/stud (giving a load

ratio of 0.2) an axial shortening of about 10 mm was observed for each stud. Figure 6-

38 (a) shows the graph of axial deformation versus load, in which the axial

displacement includes the 15 mm rigid body displacement of the studs in the axial

direction with the rest being the axial deformation for each stud at ambient conditions

when subjected to a load of 15 kN/stud. The wall was able to take the load without


any sign of visible lateral deformations. On starting the furnace, with the applied load

held constant, the studs were seen to elongate due to thermal expansion.

Figure 6-38 (b) shows the axial deformations of the studs under sustained loading

when exposed to heat. The maximum axial expansion of almost 9 mm was noticed in

Stud 2 by 80 minutes.

Figures 6-39 (a), (b) and (c) show the lateral deformations of the wall during the fire

test at the top, middle and bottom level, respectively (middle and quarter points). A

reversal in the direction of lateral deflection of the wall is seen in Figure 6-39 (a) by

67 minutes allowing the wall to straighten out. About this time the average

temperatures across the hot flanges, webs and cold flanges were 2500C, 1750C and

1400C, respectively. The wall continued to maintain the applied load of 60 kN until

111 minutes at which time the wall suddenly failed. The sudden increase in lateral

deflection as seen in Figure 6-39 (a) and the drop in load in Figure 6-40 confirm the

time of failure.

-30

-25

-20

-15

-10

-5

0

0 2 4 6 8 10 12 14 16

Load (kN)

Dis

pla

cem

ent

(mm

)


(a) Axial Deformation -Load Profiles of Test Specimen 3 at Ambient Temperature



-35

-30

-25

-20

-15

-10

-5

0

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Def

orm

atio

n (

mm

)


(a) Axial Deformation -Time Profiles of Test Specimen 3 at Elevated Temperatures


-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Def

lect

ion

(m

m)


(a) Lateral Deflection -Time Profiles of Test Specimen 3 at Upper Level

Figure 6-39: Lateral Deflection-Time Plots of Test Specimen 3


-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

De

fle

cti

on

(m

m)


(b) Lateral Deflection -Time Profiles of Test Specimen 3 at Middle Level

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Def

lect

ion

(m

m)


(c) Lateral Deflection -Time Profiles of Test Specimen 3 at Lower Level



10

11

12

13

14

15

16

17

18

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Lo

ad (

kN)

Figure 6-40: Axial Load -Time Profile of Test Specimen 3 during Fire Test


a) Date of Test: 27/08/07

b) Severity of Test: 100%

c) Specimen Temperature



200C. The maximum temperature of the unexposed surface at that time was 720C.

d) Specimen Behaviour

The fire side plasterboards 1 and 2 remained attached to the steel frame until the

failure of the test specimen. The lateral deflection of the specimen was initially

towards the furnace and then reversed in direction at the end of 67 minutes (more

pronounced at the top) from the commencement of the test. The maximum deflection

at mid-height of the wall at that time was 16 mm.


e) Failure Criterion



Partial fall off of exposed plasterboards near Studs 2 and --- occurred suddenly that

led to overall buckling failure of those studs and the failure of test specimen.

f) Performance


Structural adequacy - Failure at 111 minutes.




6.5.4: Test Specimen 4 (LBW-Cavity Insulation-GF)

6.5.4.1) Visual Observations:

Test Specimen 4 was subjected to heat in the furnace for 101 minutes. Visual

inspection soon after the test showed that plasterboards 1 and 2 (Fire side

plasterboards) had fallen off in the lower right hand portion of the wall. In the

remaining area of the wall plasterboard 1 had fallen off in some portions of the wall

whereas plasterboard 2 was still intact though severely damaged. The external

plasterboards fully collapsed under their own self weight when the specimen was

removed from the loading frame for further inspection (see Figure 6-41 (b)).

The cavity insulation was totally burnt out at the lower right hand portion whereas in

the remaining portion of the wall the cavity insulation was still intact though it had

warped and shrunk to some extent exposing certain parts of the cavity facing surface

of plasterboard 3. The glass fibres had melted on the exposed surface of the

insulation. The inner layers of insulation were still in good condition (see Figure 6-41

g). On removing the remaining pieces of exposed plasterboard and the cavity

insulation the paper on the cavity facing surface of plasterboard 3 was found to be

more or less intact although burnt out at the lower right hand portion (Figure 6-41

(c)).

The front view of the studs clearly shows that torsional failure along with flexural

buckling of the steel channels about the minor axis was fully prevented by the lateral

support offered by the plasterboards on both sides. The ambient surface of the wall

was not affected by the heat of the furnace although it had cracked up horizontally at

the centre when the wall failed by bowing in the outward direction. Studs 1, 2 and 3

from the right side of the wall had failed by local buckling (compressive failure) of

the hot flange close to the mid-height of the wall (see Figure 6-41 (d)) resulting in the

reversal of lateral displacement and causing the outward movement of the wall. Stud 4

was seen to be relatively undamaged. Stud 1 although placed symmetrically to Stud 4

showed greater damage, probably because the base layer plasterboard on the fire side

had a vertical joint running over its length, whereas the plasterboard was continuous

over Stud 4. Local buckling of the hot flange of Stud 3 was seen to occur between the

screws connecting the hot flange to the plasterboards on the fire side indicating a good

support offered by the fire side plasterboards to Stud 3 until its failure. Stud 2 (see


Figure 6-41 (e)) had its hot flange buckle locally at the screw location after the screw

got pulled out from the plasterboard thus doubling the effective length for the flange

buckling. The upper and lower tracks supporting the studs were relatively undamaged

and were seen holding the studs firmly in place (see Figure 6-41 (f)). The

plasterboards on the ambient side were intact giving good lateral support to the studs

throughout the test.

(a) View of Test Specimen 4 after the Fire Test

(b) Front View Showing Cavity Insulation (glass fibre) Burnt Out in the Lower Right Region

Figure 6-41: Test Specimen 4 after the Fire Test



(b) Front View After Removing Exposed Plasterboards and Cavity Insulation

Stud 3 Stud 4

Stud 2

Stud 1

(c) Side View Showing the Overall Buckling away from Furnace



(e) Local Buckling of Hot Flange (Stud 2)

(f) Top Track

(g) Glass Fibre Mat Used as Cavity Insulation





a) Plasterboard Surfaces (see Figures 6-42 (a) and (b))

Figures 6-42 (a) and (b) show the time-temperature profiles of the plasterboard

surfaces as observed during the fire test. Figure 6-42 (b) shows the detailed time-

temperature profiles as recorded by all the individual thermocouples installed across

the left, middle and right sections of the wall, the average of which gives the profiles

in Figure 6-42 (a).


plasterboards 1 and 2 (Pb1-Pb2)

The temperature on this surface was seen to develop in three phases. The first phase

involved a quick rise in temperature from the ambient at about 3 minutes to about

800C by the end of 5 minutes. Beyond this time the second phase started with the

temperature being held constant at about 1000C for about 20 minutes. Beyond 20

minutes the third phase started with the temperature rising quickly from about 1000C

to 8500C after 90 minutes. The temperature gradient was almost constant until this

point (similar to specimen 4, this graph too did not show any change in gradient upon

crossing the 7000C mark). Beyond 90 minutes the curve flattened out and ran with an

almost constant temperature difference of approximately 1500C with the fire curve

until the end of the test (indicating a temperature difference of 1500C across the

thickness of the fire side plasterboard). This constant temperature difference across

the plasterboard thickness also suggests no loss in the integrity of the exposed

plasterboard until the failure of the frame.

ii) Average temperature on the cavity facing surface of the exposed plasterboard

2 (Pb2-Cav)

The thermocouples on this surface responded at about 4 minutes and recorded a

temperature of about 900C by the end of 8 minutes. The second phase (plateau) lasted

until 55 minutes beyond which the third phase started with the temperature rising very

rapidly to 4000C by 65 minutes and 5000C by 70 minutes. The rate of temperature rise

for this surface decreased beyond 70 minutes and seemed to have reached a steady

state of heat flow maintaining a temperature difference of 2000C across the thickness

of the plasterboard until the end of the test. This also indicated that plasterboard 2

maintained its integrity and continued to offer protection to the steel frame until its

failure.

ii) Average temperature on the cavity facing surface of the ambient side

plasterboard 3 (Pb3-Cav)

The cavity insulation kept the temperature of this surface below 900C until 70 minutes

from the start of the test. A linear growth rate in temperature rise was observed

beyond this time. The temperature on this surface was below 2000C at the end of the

test. The temperature across the thickness of glass fibre insulation in the cavity was

about 5000C when the test was terminated. This indicated that the glass fibre

insulation maintained its integrity and was not burnt through until the failure of the

specimen.

iv) Average temperature on the ambient side of unexposed plasterboard 3

(Pb3-Pb4)


0

100

200

300

400

500

600

700

800

900

1000

1100

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 Furnace FS Pb1-Pb2

Pb2-Cav Pb3-Cav Pb3-Pb4 Amb

(a) Average Time-Temperature Profiles of Plasterboard Surfaces in Test Specimen 4



0

100

200

300

400

500

600

700

800

900

1000

1100

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 Furnace Pb1-Pb2-L Pb1-Pb2-M Pb1-Pb2-RPb2-Cav-L Pb2-Cav-M Pb2-Cav-R Pb3-Cav-L Pb3-Cav-MPb3-Cav-R Pb3-Pb4-L Pb3-Pb4-M Pb3-Pb4-R

(b) Time-Temperature Profiles across the left, middle and right sections of

Plasterboard Surfaces in Test Specimen 4


b) Steel Surfaces

Figures 6-43 (a), (b) and (c) show the temperature profiles of the hot flanges, webs

and cold flanges, respectively for the central studs. The rate of temperature increase in

the first phase was seen to decrease with the increase in the distance of the

thermocouples from the Pb2-Cav surface, giving the fastest temperature growths in

the hot flanges and the lowest in the cold flanges. The second phase extended until 50

minutes in the case of hot flanges, whereas it was seen to last until 56 minutes and 62

minutes for the webs and cold flanges, respectively. The third phase was marked by a

rapid rise in stud temperatures with the temperature growth rate of the hot flanges

being the maximum with the central hot flange temperature of Stud 2 following the

Pb2-Cav profile closely. A temperature difference was noticed along the length of the

studs. Maximum temperatures were recorded at mid-height and minimum at the top.

A temperature difference ranging from 1300C to 2300C was noticed along the length

of the studs in the hot flanges towards the end of the test. Similar patterns were

observed for web and cold flange, with the temperature differences along the stud

lengths ranging from 700C to 1800C in Figure 6-43 (b) and 1000C to 2600C in Figure

6-43 (c) at the end of the test.


0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110Time (min)

Te

mp

era

ture

(oC

)


(a) Time-Temperature Profiles on Hot Flange Surfaces of Central Studs in Test Specimen 4

0

100

200

300

400

500

600

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

Tem

per

atu

re (

oC

)


(b) Time-Temperature Profiles on Web Surfaces of Central Studs in Test Specimen 4



0

50

100

150

200

250

300

350

400

0 10 20 30 40 50 60 70 80 90 100 110Time (min)

Tem

per

atu

re (

oC

)


(c) Time-Temperature Profiles on Cold Flange Surfaces of Central Studs in Test Specimen 4


0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110Time (min)

Te

mp

era

ture

(oC

)




Figure 6-44 shows the temperature profiles across the depth of the central studs

measured at mid-height. In the first two phases the temperature distribution across the

studs was seen to be almost uniform. After the calcination of the exposed

plasterboards, the temperatures started rising rapidly (third phase) in the studs. The

presence of cavity insulation shielded the cold flanges from direct heat introducing a

large temperature variation across the depth. Maximum temperature differences of

3600C and 3300C were observed across Studs 2 and 3, respectively, at the end of the

test.


Figure 6-46 (a) shows the axial deformations of each stud loaded to 15 kN at ambient

temperature. Rigid body displacements of the studs as witnessed in Test Specimen 3

were not observed as the studs were screw connected to the upper and lower tracks.

The studs were seen to deform from 4mm to 6 mm towards the end of the loading. At

this stage, the wall had no visible lateral deformations.

Figure 6-46 (b) shows the thermal expansions of the individual studs from the time

the furnace was started until the end of the fire test. A total expansion of

approximately 10 mm was observed in the studs at the end of the test.

Figures 6-47 (a), (b) and (c) show the lateral deformations of studs with respect to

time taken at three different levels. Lateral deformations towards the furnace could be

noticed from the start with the central studs deflecting more than the end studs making

them the critical studs. Until 55 minutes the lateral deformations were seen to develop

slowly. Beyond this time the deformations became more rapid, with Stud 2 recording

the maximum deflection of approximately 32 mm at the centre by the end of 85

minutes with its hot flange recording a temperature of 5700C and with a temperature

difference of approximately 3300C across the cross-section. About the same time the

lateral deflection of Stud 3 was approximately 30 mm at the centre, with its hot flange

measuring a temperature of about 4700C and a temperature difference of 2500C across

its depth. Beyond 85 minutes, Stud 2 reversed its direction of lateral deformation and

started to straighten out. Stud 3 maintained its deflection profile from 85 to

approximately 95 minutes before reversing its lateral deformation and begin to

straighten out. By this time its hot flange had reached a temperature of 5700C


(temperature at which Stud 2 had started to deform in the outward direction) and had a

temperature difference of approximately 3100C across its depth. Meanwhile Stud 2

had deflected by 12 mm in the reverse direction and had a hot flange temperature of

6600C. Beyond 95 minutes the lateral deformations progressed rapidly in both the

studs leading to failure at 101 minutes as seen from Figure 6-48 which gives the exact

time of failure of the wall. The ambient side plasterboards were cracked open by the

outward thrust offered by Stud 2 at the centre (Figure 6-45).

Figure 6-45: Outward Lateral Deflection of Test Specimen 4 at Failure

The sudden deformation of the wall at about 101 minutes must have caused the partial

collapse the exposed plasterboards and led to the burning out of the cavity insulation.

The presence of the external plasterboards and the insulation can be verified from the

time-temperature profiles as seen in Figures 6-42 (a) and (b).


-8

-6

-4

-2

0

2

4

6

8

10

12

0 2 4 6 8 10 12 14 16

Load (kN)

De

form

ati

on

(m

m)



-8

-6

-4

-2

0

2

4

6

8

10

12

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

Def

orm

atio

ns

(mm

)


(b) Axial Deformation -Time Profiles of Test Specimen 4 at Elevated Temperatures



-35

-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

Def

lect

ion

(m

m)



-35

-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110Time (min)

Def

lect

ion

(m

m)





-35

-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

Def

lect

ion

(m

m)




10

11

12

13

14

15

16

17

18

0 10 20 30 40 50 60 70 80 90 100 110

Time (min)

Lo

ad (

kN)












failure of the test specimen. Portions of plasterboard in the lower right hand area of

the wall fell off after the structural failure of the wall.The lateral deflection of the test

specimen was initially towards the furnace and then reversed in direction at the end of

85 minutes from the commencement of the test. The maximum deflection at mid-

height of the wall at that time was 32 mm.




Fire side plasterboards appeared to have provided sufficient restraint to the studs and

hence the studs did not undergo any buckling failures associated with twisting. The

main failure mode was due to local buckling / crushing of hot flanges and flexural

deformations about the major axis.

f) Performance






6.5.5: Test Specimen 5 (LBW-Cavity Insulation-RF)

6.5.5.1) Visual Observations:

Test Specimen 5 was subjected to heat in the furnace for 107 minutes. Both of the fire

side plasterboards had fallen off at the end of the test as seen in Figure 6-49 (a). The

rock fibre cavity insulation was almost fully intact with only the outer layer of

insulation having lost its integrity at certain locations as seen in Figure 6-49 (b). On

stripping the cavity insulation off from the wall, it was noted that Plasterboard 3 had

remained in good condition until the end of the test. Only the paper on the cavity

facing surface of Plasterboard 3 was burnt in certain locations as seen in Figure 6-49

(c).

The front view (see Figure 6-49 (c)) shows that torsional failure was prevented in

Studs 2, 3 and 4. However, Stud 1 displayed a combination of local compressive

failure and torsional buckling of the hot flange. The torsional buckling of the hot

flange probably occurred as the exposed plasterboard 1 had partially collapsed in that

region and the severely calcined base layer plasterboard (Pb2) was unable to provide

sufficient lateral restraint all by itself. Figure 6-49 (d) shows the local compressive

failure of the hot flanges of Studs 1, 2 and 3. Figures 6-49 (e) to (g) show the close up

views of the individual studs.

The ambient side plasterboards, although in good condition until the end of the test,

had cracked up when the wall failed by bowing in the outward direction. Studs 2 and

3 from the right side of the wall failed by local compressive failure of the hot flange at

the mid-height between the screws, whereas Stud 1 failed by local buckling of hot

flange initiated by screw pull out. Stud 4 was seen to be in good condition. The tracks

were seen to be in good condition and maintained good contact with the studs

throughout the fire test (Figures 6-49 (h) and (i)).



(a) Front View Showing Cavity Insulation (Rock Fibre) after the Fire Test

(b)Front View Showing Cavity Insulation (Rock Fibre) after the Fire Test



Studs are numbered 1 to 4 from right to left

Stud 1 displaying local compressive failure and torsional buckling of the hot flange

Burn-out of paper on the cavity facing surface of Pb3

(c) Front View after Removing Exposed Plasterboard Pieces and Cavity Insulation (Rock Fibre) after the Fire Test


(d) Side View Showing Overall Buckling of Wall Away From Furnace



(e) Local Compressive Failure (f) Local Compressive Failure of Hot Flange (Stud 1) of Hot Flange (Stud 2)

(g) Local Compressive Failure (h) View of Top Track of Hot Flange (Stud 3)

(i) View of Bottom Track (j) Rock fibre mats after Fire Test



6.5.5.2) Time-Temperature Profiles:




The first two phases of the time-temperature profile of this interface was almost

identical to that of Specimens 3 and 4. The third phase started from about 19 minutes

with the temperature rising sharply from 1000C to 9000C by 90 minutes while the

temperature gradient was maintained almost constant over the entire period. Beyond

90 minutes the temperature was almost constant with the fire curve giving a

temperature difference of approximately 500-1000C across the plasterboard thickness

until the end of the test. Figure 6-50 (b) shows the Pb1-Pb2-L temperature profile

merging with the fire side curve at about 88 minutes signifying gradual collapse of

parts of exposed plasterboard in the left section of the wall i.e. between studs 1 and 2.

ii) Average temperature of the cavity facing surface of the exposed Plasterboard

2 (Pb2-Cav)

The first phase started at approximately 4 minutes from the ignition of the furnace

taking the temperature from the ambient to about 500C by the end of 6 minutes

beyond which the profile with a very gentle gradient representing the second phase

started. The temperature in the second phase reached 1100C by the end of 55 minutes,

beyond which the third phase started with the temperature rising sharply and crossing

4000C by 70 minutes and 5000C by 78 minutes. The gradient of the curve beyond 65

minutes reduced with the graph stabilizing itself and maintaining a temperature

difference of 2500C – 3000C across the thickness of the second layer of exposed

plasterboard until the end of the test indicating the physical presence and the

protection offered by the plasterboard until the failure of the specimen.

The absence of any sharp temperature rise in the profiles of the Pb2-Cav as observed

in Figure 6-50 (b) suggests that the exposed base layer plasterboard (Pb2) was intact

in all the three wall sections until the end of the test. Pb2-Cav-L showed a small

uptrend as compared to Pb2-Cav-M and Pb2-Cav-R probably due to the partial

collapse of Pb1 in that region at about 88 minutes. The base layer plasterboard in this

region must have got more severely calcinated as compared to other regions as it


collapsed at about 106 minutes following the structural failure of the wall as can be

seen by the sudden temperature rise of Pb2-Cav-L profile at 106 minutes in Figure 6-

50 (b).



Similar to Specimen 4, the temperature on this surface was kept below 900C until 70

minutes by the protection offered by the rock fibre insulation in the cavity. Beyond 70

minutes the temperature growth rate was linear reaching 2000C by the end of the test.

The rock fibre insulation seemed to have maintained its integrity throughout the test

as no sudden increment in temperature was noted in the Pb3-Cav surface temperature.

The temperature across the insulation was approximately 5500C at the end of the test.

The Pb3-Cav-L profile in Figure 6-50 (b) shows the maximum gradient as compared

to Pb3-Cav-M and Pb3-Cav-R. This is in confirmation to the partial collapse of

plasterboard 1 in this region.

0

100

200

300

400

500

600

700

800

900

1000

1100

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Te

mp

era

ture

(oC

)





0

100

200

300

400

500

600

700

800

900

1000

1100

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Te

mp

era

ture

(oC

)


(b) Time-Temperature Profiles across the left, middle and right sections of Plasterboard Surfaces in Test Specimen 5



(Pb3-Pb4)



The temperature on the unexposed face of the wall remained below 550C (well below

the insulation failure temperature) during the fire test.

b) Steel Surfaces

The time-temperatures profiles of the hot flanges, webs and cold flanges taken along

the central studs are shown in Figures 6-51 (a), (b) and (c). The temperatures along

the stud lengths were seen to be more uniform than Test Specimen 4, with the

maximum temperature difference along the stud length being less than 1000C.

The sudden rise in temperature of S2-L-HF (see Figure 6-51 (a)) at about 92 minutes

confirms with the partial collapse of plasterboard 1 in that region.


The second phase of the time-temperature profile was seen to extend until 50 minutes

for the hot flanges, 60 minutes for the webs and 65 minutes in the case of cold

flanges.

0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Te

mp

era

ture

(oC

)


(a) Time-Temperature Profiles of Hot Flange Surfaces of Central Studs in Test Specimen 5

0

50

100

150

200

250

300

350

400

450

500

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Tem

per

atu

re (

oC

)


(b) Time-Temperature Profiles of Web Surfaces of Central Studs in Test Specimen 5



0

50

100

150

200

250

300

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Te

mp

era

ture

(oC

)


(c) Time-Temperature Profiles of Cold Flange Surfaces of Central Studs in Test Specimen 5


0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Tem

per

atu

re (

oC

)




Note:

S2/3-T/M/L-HF: Time-temperature profile followed by the hot flange at top/middle/lower level of Stud No. 2/3

S2/3-T/M/L-W: Time-temperature profile followed by the web at top/middle/lower level of Stud No. 2/3

S2/3-T/M/L-CF: Time-temperature profile followed by the cold flange at top/middle/lower level of Stud No. 2/3

Figure 6-52 shows the temperature profiles of the thermocouples placed across the

depth of the central studs at mid-height. Similar to Test Specimen 4, a large

temperature variation is noted across the stud cross-sections. Studs 2 and 3 were seen

to have temperature differences of 3800C and 3300C, respectively, over their depths at

the end of the test.


The specimen was seen to deflect laterally towards the furnace during the early stages

of the test and then outwards during the final stages (see Figure 6-53 (a) and (b))

finally leading to the cracking of the plasterboard on the ambient side of the wall( see

Figure 6-54).

Figure 6-55 (a) shows the axial deformations suffered by the studs during the loading

of each stud to 15 kN at ambient temperatures. The studs were seen to deform on an

average by approximately 6 mm. Figure 6-55 (b) shows the axial expansion of the

studs brought about by increasing the stud temperatures under constant load. A total

expansion of approximately 10 mm was observed in the studs by the end of the test.

Figures 6-56 (a), (b) and (c) give the lateral deflections of the studs taken at the top,

middle and lower levels. The wall started bowing slowly towards the furnace from the

start of the test and was seen to deform rapidly beyond 60 minutes. The deformations

in stud reached a maximum of 36 mm by 92 minutes at which time the hot flange

temperature of the stud had reached 5500C at the centre and the temperature

difference across the stud was 3300C. Beyond 92 minutes Stud 2 underwent a quick

reversal in lateral deformation and started to straighten out. This affected the axial

deformation plot.

A sharp rise in the expansion rate of Stud 2 can be observed from about 94 minutes

which was caused by the release of the excess oil pressure in the jacks to maintain the


load constant, thus allowing Stud 2 to reverse its lateral deformation and straighten

out without accumulating temperature stresses. The release of the oil pressure did not

affect the load carried by the other studs as only the excess oil pressure developed by

Stud 2 was released. This could be verified by the smooth continuation of the load

versus time graph (Figure 6-57) until failure.

Stud 3 continued to deflect laterally until 103 minutes reaching a maximum of

approximately 34 mm. At this time its hot flange temperature at the centre had

reached 5500C and the temperature across the stud was 3300C (same as that observed

in Stud 2 at the end of 92 minutes) and started to reverse laterally in the outward

direction. A response similar to the one observed for Stud 2 at 94 minutes can be seen

for Stud 3 at 104 minutes in the axial deformation graph (Figure 6-55 (b)). By 102

minutes Stud 2 had moved in the reverse direction by 24 mm and continued to deform

rapidly outwards. This was simultaneously followed by a rapid increase in the lateral

deformation in the outward direction by Stud 3 leading to the failure of the wall at 107

minutes as seen from Figure 6-57.

(a) Inward Bowing of Specimen during Early Stages of Test

Figure 6-53: Specimen Behaviour during the Test


(b) Outward Bowing of Specimen during Final Stages of Test Figure 6-53: Specimen Behaviour During the Test

Figure 6-54: Ambient Side of Wall after Fire Test



-7

-6

-5

-4

-3

-2

-1

0

0 2 4 6 8 10 12 14 1

Load (kN)

Def

orm

atio

n (

mm

)

6



-8

-6

-4

-2

0

2

4

6

8

10

12

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (mm)

Def

orm

atio

ns

(mm

)




-40

-35

-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

De

fle

cti

on

(m

m)

Stud 2 Stud 3


-40

-35

-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Def

lect

ion

(m

m)





-40

-35

-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

De

fle

cti

on

(m

m)




10

11

12

13

14

15

16

17

18

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Lo

ad (

kN)












failure of the test specimen. The lateral deflection of the test specimen was initially

towards the furnace and reversed in direction at the end of 92 minutes from the

commencement of the test. The maximum deflection at mid-height of the wall at that

time was 35 mm.


The specimen was seen to fail deflecting laterally in the outward direction (away from

the furnace) due to the local compressive failure of the exposed flanges under a

flexural bending action about the major axis.

The Test Specimen was deemed to have failed at approximately 107 minutes from the


f) Performance






6.5.6: Test Specimen 6 (LBW-Cavity Insulation-CF)


The fire test was stopped at 112 minutes when the specimen failed to maintain the

applied load and it was observed that the fire side plasterboards had fallen off from

the central and lower portion as seen in Figure 6-58 (a). Cellulose fibre used in the

cavity had completely burnt out leaving behind a black residue. The paper on the

cavity facing surface of Plasterboard 3 had burnt out completely. The surface had also

undergone deep calcination showing multiple shrinkage cracks (Figure 6-58 (b)).

The studs stayed connected to the ambient side plasterboards over the entire height.

The ambient surface of Plasterboard 4 did not show any temperature effects although

it had lost its integrity due to frame failure. The upper track connecting the studs was

heat affected and showed distortional buckling along its length (Figure 6-58 (g)). The

lower track was still in good condition probably because the convection currents carry

the heat in the upward direction keeping the top track at a temperature higher than the

bottom track.

The central studs were seen to be severely damaged. Stud 1 showed a small amount of

local buckling of the hot flange at 370 mm from the bottom (Figure 6-58 (c)). Stud 4

was undamaged. In Studs 2 and 3 the local buckling of the hot flange involved screw

pull out through the plasterboard thus doubling the buckling length of the flange

leading to rapid outward movement of the studs causing the punching of the

plasterboard on the ambient side (Figure 6-58 d). Stud 3 also displayed local crushing

of the cross-section near the top support in the frame (Figure 6-58 (e)).


(a) View of Test Specimen 6 after the Fire Test

(b) Front View of Test Specimen 6

Note: In all the test specimens the studs are numbered 1 to 4 from right to left



(c) Local Buckling of Hot Flange in Stud 1 at 370 mm from base

(d) Local Buckling of Hot Flange in Stud 2 at 350 mm from base

(e) Local Crushing Near the Top Support



(f) Side View of Specimen (g) View of the Top Track


6.5.6.2) Time-Temperature Profiles a) Plasterboard Surfaces (see Figures 6-59 (a) and (b))



The first two phases of this interface was seen to follow the same profile as that of

Specimens 3, 4 and 5. The third phase started from about 20 minutes with the

temperature rising sharply from 1000C to 9000C by about 85 minutes from the start of

the test. The temperature gradient in the third phase was fairly constant until this time.

Beyond 85 minutes the graphs started to flatten out, but continued to converge with

the fire side curve (suggesting the rapid deterioration of the fire side plasterboard) and

indicating a breach or partial collapse of the plasterboard (see Figure 6-59 (a)).

Figure 6-59 (b) gives the temperature profiles of the individual thermocouples placed

in the three regions of the wall (left, middle and right). Temperature profile of Pb1-

Pb2-L meets the fire side curve at about 79 minutes indicating a partial collapse of

Plasterboard 1 in this region (i.e. between Studs 1 and 2).

ii) Average temperature on the cavity facing surface of the exposed Plasterboard

2 (Pb2-Cav)

The first phase started from about 5 minutes taking the temperature up to 900C by 11

minutes from the start of the test. Beyond this time in the second phase the

temperature was maintained at about 1000C until about 50 minutes beyond which the

third phase started with the temperature rising quickly and crossing 4000C by 68


minutes and 5000C by 72 minutes. Beyond 70 minutes the curvature of the

temperature gradient started to decrease with the graph stabilizing itself and

maintaining a temperature difference of approximately 2500C across the thickness of

the second layer of the exposed plasterboard until the end of the test. Figure 6-59 (b)

shows a sudden rise in the temperature of Pb2-Cav-R profile close to the end of the

test at about 106 minutes signifying the breaching of Plasterboard 2 in this region.



Similar to the other cavity insulated specimens, the temperature on this surface was

maintained below 1000C until 70 minutes. The temperature started rising fast beyond

70 minutes crossing 2000C by 88 minutes, 3000C by 97 minutes and 4000C by 110

minutes. The temperature difference across the thickness of the insulation was close to

2000C at the end of the test.

A sharp rise in the temperature of the Pb3-Cav surface after 110 minutes indicates the

quick deterioration of the cellulose fibre in the cavity of the wall. A steep rise of Pb3-

Cav-L and Pb3-Cav-R in Figure 6-59 (b) indicates the burn-out of the cellulose

insulation in these regions at the end of the test.


(Pb3-Pb4)

The temperature of this interface remained below 850C until end of the test.


The temperature of the unexposed face of the wall remained under 600C (well below


Figure 6-59 (b) shows the detailed time-temperature profiles as recorded by all the

individual thermocouples installed across the left, middle and right sections of the

wall.


0100200300400500600700800900

100011001200

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Tem

per

atu

re (

oC

)



0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Tem

per

atu

re (

oC

)


(b) Time-Temperature Profiles across the left, middle and right sections of Plasterboard Surfaces in Test Specimen 6



b) Steel Surfaces

Similar to the other cavity insulated specimens the temperatures along the length of

the central studs in the web and cold flange portions were seen to be almost uniform

(see Figures 6-60 (b) and (c)). Larger temperature variations along the length were

observed in the hot flanges with the maximum temperature occurring at the centre (6-

60 (a)).

Figure 6-61 shows the temperature variations across the depth of the central studs at

mid-span. The temperature difference between the hot and cold flanges can be seen to

increase from about 1000C at 60 minutes to about 3500 to 3700C at the end of the test.

0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Tem

per

atu

re (

oC

)





0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Tem

per

atu

re (

oC

)



0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120Time (min)

Te

mp

era

ture

(oC

)


(c)Time-Temperature Profiles of Cold Flange Surfaces of Central Studs in Test Specimen 6



0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Tem

per

atu

re (

oC

)




The Test Specimen on loading until 60 kN (i.e. 15 kN/Stud) showed no visible signs

of lateral deformation at ambient temperature. However, the studs showed an average

axial shortening of about 7 mm on reaching the applied load (Figure 6-63 (a)). On

starting the furnace the studs started expanding due to heat and by the end of the test

the average expansion of the studs was seen to be about 8 mm (Figure 6-63 (b)). The

wall was observed to bow towards the furnace from the start. With the passage of time

the lateral deformations continued to increase (see Figure 6-62) as the temperature

difference across the depth of the wall increased. Interface temperature Pb1-Pb2-L

(Figure 6-59 (b)) shows a jump in temperature at about 78 minutes, indicating a

partial breach or collapse of Plasterboard 1 between Studs 1 and 2. By 96 minutes

Stud 2 was seen to have laterally deformed by 35 mm towards the furnace beyond

which its direction was seen to reverse (Figure 6-64 b). The rest of the studs continued

to bow towards the furnace. About 107 minutes, a step in the temperature profile of

interface Pb1-Pb2-R indicated the cracking up of Plasterboard 1 between Studs 3 and

4 (Figure 6-59 (b)). Within a couple of minutes the exposed base layer plasterboard in

the same place cracked up as shown by Pb2-Cav-R profile exposing the studs and

cavity insulation (cellulose fibre) to direct furnace heat in this region. A temperature


difference of about 3300C across the cavity at this time (Pb2-Cav R – Pb3-Cav R)

indicates the physical presence of cellulose fibre in the cavity. By about 107 minutes,

Stud 3 had deformed laterally towards the furnace by 44 mm (Figure 6-64 (b)). Its

direction was seen to reverse abruptly beyond this time with the deformations

progressing rapidly and finally leading to the failure of the wall at about 110 minutes

as confirmed by Figure 6-65.

Similar to Test Specimen 5, the structural failure in this specimen also occurred

within 15 minutes of the first stud reversal. In this period large lateral deformations

were responsible for the braking of the fragile and calcinated exposed plasterboards

Figure 6-62: Inward Thermal Bowing of Test Specimen in the Initial Period of the Test



-15

-10

-5

0

5

10

15

0 2 4 6 8 10 12 14 1Load (kN)

De

form

ati

on

(m

m)

6



-15

-10

-5

0

5

10

15

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

De

form

ati

on

(m

m)




-45

-40

-35

-30

-25

-20

-15

-10

-5

0

5

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (mm)

Def

lect

ion

(m

m)



-45

-40

-35

-30

-25

-20

-15

-10

-5

0

5

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Def

lect

ion

(m

m)





-45

-40

-35

-30

-25

-20

-15

-10

-5

0

5

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Def

lect

ion

(m

m)




10

11

12

13

14

15

16

17

18

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Lo

ad

(k

N)











The fire side Plasterboard 1 breached partially at about 78 minutes between Studs 1

and 2, whereas Stud 2 remained attached to the steel frame until the failure of the Test

Specimen. The lateral deflection of the test specimen was initially towards the furnace

and reversed in direction at the end of 96 minutes from the commencement of the test.

The deflection at mid-height of the wall at that time was 35 mm.




f) Performance






6.5.7: Test Specimen 7 (LBW-External Insulation-GF)


At the end of fire test, the exposed plasterboards were seen to have fallen off near the

centre of the wall exposing the Pb3-Cav surface (Figure 6-66 (a)). The glass fibre

insulation used between Plasterboards 1 and 2 had completely disappeared leaving

only some molten glass traces along the periphery of the specimen. The base layer

plasterboard on the fire side was seen to have shrunk, opening the plasterboard joints

by 10 -15 mm thus exposing the studs to fire (Figure 6-66 (d)).

The exposed plasterboards were stripped off and the debris removed to expose the

frame (Figure 6-66 (b)). A local buckling wave in the web and flanges was visible

over the middle third length of Stud 3 (Figure 6-66 (c)). Stud 2 showed local

compressive failure of the entire cross-section close to the mid-span suggesting the

complete mobilization of the capacity of the cross-section (Figure 6-66 (f)). Figure 6-

66 (e) shows the local buckling wave in the flange and web near mid-height of Stud 2.

Stud 1 was seen to display local buckling of hot flange and web near mid-height. Stud

4 was seen to be undamaged (Figure 6-66 (c)).

(a) Front View Showing the Partial Collapse of Exposed Plasterboards



The front view (Figure 6-66 (b)) shows that torsional buckling and flexural buckling

about minor axis were fully prevented by the lateral support provided by the

plasterboards. From the side view (Figure 6-66 (c)) one can see that even the flexural

buckling about the major axis was almost negligible for all the studs.


(b) Front View after Removing the Exposed Plasterboards and Remains of External Insulation

Stud 4 Stud 3

Stud 2 Stud 1

(c) Side View Showing the Slight Outward Buckling of Wall



(d) 10-15 mm Wide Opening (e) Local Buckling Wave in of Plasterboard Joints Flanges and Web at mid-height of Stud 2

(f) Local Buckling of Flanges

and Web in Stud 2



a) Plasterboard Surfaces (see Figure 6-67)



As for the cavity insulated specimens, the temperature profile of this surface showed a

rapid rise in temperature within three minutes of starting the test (first phase) reaching

a temperature of about 900C by the end of 5 minutes. Beyond 5 minutes the second

phase started with the temperature increasing gradually to about 1200C by the end of

15 minutes. This was followed by a very rapid rise in temperature (third phase)

crossing 4000C in 25 minutes and 5500C in 30 minutes. In comparison for Specimen 4

using glass fibre as cavity insulation, the ambient surface of Plasterboard 1 recorded


the temperature of 4000C after 42 minutes and crossed 5500C by 59 minutes from the

start of the test. The sudden increase in the temperature of the interface of Specimen 7

in contrast to that of Specimen 4 in the early stages of Phase 3 was probably caused

by the heat blocked and redirected by the adjoining layer of insulation. Beyond 30

minutes the temperature gradient was reduced with the temperature crossing 6500C by

about 40 minutes. From this point onwards a difference in temperature of

approximately 2500C was maintained across the thickness of Plasterboard 1 until 72

minutes at which time the furnace malfunctioned causing the fire curve to drop as

seen in Figure 6-67. This caused the temperature of the Pb1-Ins interface to fall

simultaneously and follow the fire curve with a small lag in temperature. The fire

curve was stepped up by manually operating the furnace from 150 minutes onwards

which caused the interface temperature to also rise. The temperature of the interface

was about 9500C by the end of the test.



This interface responded to the initial rise in temperature (phase 1) in under 4 minutes

from the start of the furnace and reached a temperature of approximately 800C rapidly

by about 6 minutes and then remained constant (second phase) until about 25 minutes.

Beyond this time the third phase started with the temperature rising almost linearly

with respect to time touching 7000C by about 72 minutes. A closer look at the

individual thermocouples revealed that Ins-Pb2-L thermocouple recorded very rapid

temperature gains and intersected the Pb1-Ins-L curve at 55 minutes indicating the

complete burning of the insulation at mid-height between Studs 1 and 2. The

intersection of curves Pb1-Ins-M and Ins-Pb2-M at about 70 minutes indicates the

burning out of the insulation at mid-height of the wall between Studs 2 and 3. This

was followed by the disintegration of the glass fibre insulation at mid-height of wall

between Studs 3 and 4 at about 78 minutes.


2 (Pb2-Cav)

The initial rise in temperature of this surface was followed by a plateau extending

until 80 minutes (compared with the 55 minutes observed in Specimen 4, using glass


fibre as cavity insulation). The third phase started beyond 80 minutes with the

temperature increasing gradually and crossing 5500C towards the end of the test. A

temperature difference of approximately 5000C was present across the thickness of

Plasterboard 2 towards the end of the test indicating the sustained integrity of the

exposed base plasterboard layer until the failure of the specimen.


3 (Pb3-Cav)

The transmission of heat across the cavity was almost instantaneous due to radiation

making the temperature profile of Pb3-Cav surface follow very closely but just on the

underside of the Pb2-Cav surface profile.



The temperature of this surface increased gradually reaching about 1000C and

remained almost constant until 125 minutes. The third phase started beyond this time

with the temperature increasing linearly with respect to time reaching 4000C by 180

minutes. Beyond 160 minutes it was observed that the temperature difference across

the plasterboard was approximately 750C until the failure of the specimen showing the

integrity of the ambient base layer plasterboard during the test.



This surface was well protected from the fire and remained under 1000C until 170

minutes from the start of the test. The temperature was just below 1100C towards the

end of the fire test.


The temperature on the unexposed face of the wall remained under 500C (well below



0

100

200300

400

500

600

700

800900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190

Time (min)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190

Time (min)

Tem

per

atu

re (

oC

)

AS 1530.4 Furnace FS-L FS-M FS-RPb1-Ins-L Pb1-Ins-M Pb1-Ins-R Ins-Pb2-L Ins-Pb2-MIns-Pb2-R Pb2-Cav-L Pb2-Cav-M Pb2-Cav-R Pb3-Cav-LPb3-Cav-M Pb3-Cav-R Pb3-Ins-L Pb3-Ins-M Pb3-Ins-R

(b) Time-Temperature Profiles across the Left, Middle and Right Sections of Plasterboard Surfaces in Test Specimen 7


b) Steel Surfaces

Figures 6-68 (a), (b) and (c) show the time-temperature profiles of hot flanges, webs

and cold flanges of the central studs. The temperatures along the length are seen to be


almost uniform in all three cases. Figure 6-69 gives the temperature variation across

the central stud cross-sections at mid-height. The temperature difference is observed

to be under 800C throughout.

0

100

200

300

400

500

600

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Te

mp

era

ture

(oC

)

S2-T-HF S3-T-HF S3-M-HF S2-L-HF S3-L-HF


0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)





0

100

200

300

400

500

600

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200Time (min)

Te

mp

era

ture

(oC

)

S3-M-HF S2-M-W S3-M-W S2-M-CF S3-M-CF



The studs when loaded at ambient temperature up to the applied load (15 kN/stud)

showed an average axial shortening of 5 mm (Figure 6-70 (b)). There were no signs of



lateral deformations. On starting the furnace the wall was observed to start bowing

towards the furnace with Studs 2 and 3 reaching a maximum central deflection of 16

mm at 172 and 180 minutes, respectively (Figure 6-71 b)). Near the end of the test

this was reversed and both studs deformed very rapidly in the outward direction

leading to the wall specimen failure at 181 minutes (Figure 6-72).

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

0 2 4 6 8 10 12 14 1

Load (kN)

Def

orm

atio

n (

mm

)

6



-8.0

-6.0

-4.0

-2.0

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Time (min)

Def

orm

atio

n (

mm

)




-20

-15

-10

-5

0

5

10

0 20 40 60 80 100 120 140 160 180 200

Time (min)

Def

lect

ion

(m

m)



-20

-15

-10

-5

0

5

10

0 20 40 60 80 100 120 140 160 180 200Time (min)

De

fle

cti

on

s (

mm

)





-15

-10

-5

0

5

10

0 20 40 60 80 100 120 140 160 180 200

Time (min)

Def

lect

ion

(m

m)




10

11

12

13

14

15

16

17

18

0 20 40 60 80 100 120 140 160 180 200

Time (min)

Lo

ad (

kN)





b) Severity of Test: Less than 80 %




150C. The maximum temperature of the unexposed surface at that time was 570C


The fire side Plasterboards 1 and 2 remained attached to the steel frame until the




time was 15 mm.




f) Performance

Fire performance could not be specified as the test specimen was not subjected to the

standard time temperature curve due to the malfunctioning of the furnace during the

test.




The specimen was subjected to 112 minutes of furnace exposure after which the test

was stopped following the failure of the specimen. Figure 6-73 (a) shows the

specimen after the fire test. Exposed Plasterboard 1 (Pb1) had completely fallen off.

Rock fibre insulation had disintegrated near the lower right corner of the wall.

Exposed base layer plasterboard (Pb2) had also collapsed in this area exposing the

cavity surface of the ambient side plasterboard (Pb3).

The external insulation had undergone overall shrinking leading to the opening of the

joints and exposing Plasterboard 2 (Figure 6-73 (b)). The base layer plasterboard had

also undergone extensive calcination with the joints opening up to expose the hot

flanges of the studs (Figure 6-73 (c)).

The front view of the specimen (Figure 6-73 (d)) after removing the external

plasterboards and insulation shows the studs without any torsional or flexural

buckling about the minor axis as the studs were laterally supported adequately.

Figures 6-73 (e) and (h) show the local buckling wave in the central part of Stud 1.

The local buckling wave can be clearly seen in the hot flange and web elements. Studs

2 and 3 also exhibited local buckling of flange and web elements (Figure 6-73 (i)).

Severe local buckling was seen near the mid-height, that led to an outward movement

of the studs and thus breaking of the ambient side plasterboards (Figure 6-73 (g) and

(h)). The tracks were seen to maintain good contact and connection with the studs.

(Figure 6-73 (k)).


Rock fibre insulation intact in the top half of the specimen

Collapse of external plasterboards and insulation in the lower right portion of the specimen

Pb3-cav surface

(a) View Showing Partial Collapse of Exposed Plasterboards and External Insulation

(b) Shrinkage Gap (c) Shrinkage Gap in Base in Insulation Joint Exposing Plasterboard Exposing Steel Stud

Base Plasterboard




(d) Front View after Removing Exposed Plasterboards and External Insulation (Rock Fibre)

Stud 4

Stud 3

Stud 2 Stud 1

(e) Side View Showing the Failure Modes of Studs 1, 2 and 3



(f) Local Buckling Wave (g) Local Buckling of in Flanges and Web Complete cross-section in Stud 2

Stud 3 Stud 2

(h) Local Buckling of (i) Side View Showing Local Buckling Complete cross-section in Stud 3 Wave near Failure Points in the

Studs

(j) View of Top Track (k) Joint between Stud and Track



6.5.8.2) Time-Temperature Profiles a) Plasterboard Surfaces (see Figure 6-74)



The temperature on this surface developed in three phases as in Specimen 7 with the

first phase involving a quick rise in temperature from the ambient at about 2 minutes

to 900C by the end of 5 minutes (see Figure 6-74 (a)). The second phase involving the

plateau lasted until 18 minutes by which time the temperature had gradually risen to

about 1100C. Beyond 18 minutes the third phase took off with a sharp rise in

temperature crossing 4000C by 24 minutes and 5500C by 30 minutes (almost identical

with Specimen 7). Beyond 30 minutes the curve seemed to flatten out with the

temperature of the interface increasing gradually and reaching 9000C by about 105

minutes. A temperature difference of 100 - 1500C was noticed at about this time

across the thickness of the plasterboard which was maintained until the end of the test.

A temperature difference of 1000C across Plasterboard 1 towards the end of the test

suggested that the integrity of the fire side plasterboard was maintained until the

failure of the specimen. Figure 6-74 (b) also shows gentle gradients for the Pb1-Ins

profiles signifying the presence of Plasterboard 1 until the end of the test.



The temperature of this interface gained rapidly (first phase) from the ambient

temperature at 4 minutes and climbed rapidly to about 900C under 5 minutes (see

Figure 6-74 (a)). This was followed by a plateau (second phase) extending until 25

minutes, beyond which the temperature gained almost linearly with respect to time

(third phase) reaching 4000C by 60 minutes and 8000C by 110 minutes. After this time

a constant temperature difference of about 1000C was maintained with the Pb1-Ins

curve, which signifies a temperature difference of 1000C across the thickness of the

insulation until the end of the test. This indicated the intactness of the rock fibre

insulation used as external insulation between Plasterboard 1 and Plasterboard 2 until

the failure of the specimen.


Figure 6-74 (b) shows a temperature difference between the profiles of Pb1-Ins and

that of Ins-Pb2 at any time during the test signifying the presence of insulation which

caused the drop in temperature until the test was stopped.


2 (Pb2-Cav)

The temperature profile of this surface was almost identical with that of Specimen 7

until 80 minutes, beyond which the third phase started with an almost linear

temperature growth rate crossing 3000C by 105 minutes (compared to 7200C by the

same time in Specimen 5 using the same insulation in the cavity). The surface

recorded a temperature of 4500C by 135 minutes near the end of the test which also

happened to be the temperature difference across the thickness of the plasterboard

demonstrating its integrity until the failure of the specimen.


3 (Pb3-Cav)

In the absence of insulation, the transmission of heat across the cavity by radiation

was very quick, forcing the Pb3-Cav surface to heat up almost instantaneously and

trace very closely on the underside of the time-temperature profile of Pb2-Cav surface

with the maximum temperature difference between the two surfaces not exceeding

500C until the end of test.



The temperature of this interface remained at about 1000C until almost 120 minutes,

beyond which it started rising gently. A temperature difference of 2500C was

noticeable across the thickness of the Plasterboard 3 near the end of the test signifying

the continued integrity of the plasterboard until the failure of the specimen.



The ambient side base layer plasterboard and the rock fibre insulation continued to

give very good protection to this surface keeping its temperature below 1000C during

the test.


0100200300400500600700800900

100011001200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Te

mp

era

ture

(oC

)

AS 1530 Furnace FS-L FS-M FS-RPb1-Ins-L Pb1-Ins-M Pb1-Ins-R Ins-Pb2-L Ins-Pb2-MIns-Pb2-R Pb2-Cav-L Pb2-Cav-M Pb2-Cav-R Pb3-Cav-LPb3-Cav-M Pb3-Cav-R Pb3-Ins-L Pb3-Ins-M Pb3-Ins-R





The temperature on this surface remained below 700C until the end of the test thus

excluding the possibility of insulation failure at any stage of the fire test.

b) Steel Surfaces

The time-temperature profiles of hot flanges, webs and cold flanges of the central

studs are shown in Figures 6-75 (a) to (c) with the second phases extending until 70,

80 and 85 minutes respectively (compared to 50, 60 and 65 minutes in Test Specimen

5 using rock fibre as cavity insulation). As in the case of other specimens, the

temperature at the mid-height of the wall was found to be the maximum. Beyond 80

minutes and until the end of the test the difference in temperature in the hot flanges

along the stud lengths was found to be about 1000 – 1300C (Figure 6-75 (a))

(compared to the temperature difference of 700 – 1000C along the hot flanges of the

central studs in Test Specimen 4 beyond 70 minutes and until end of the test (Figure

6-43 (a)). The maximum recorded hot flange temperatures at 60 and 90 minutes were

1200C and 2500C, respectively (compared to 2100C and 5400C in the cavity insulated

specimen at the same times).

0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Tem

per

atu

re (

oC

)





0

100

200

300

400

500

600

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Tem

per

atu

re (

oC

)



0

100

200

300

400

500

600

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Tem

per

atu

re (

oC

)





0

100

200

300

400

500

600

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Tem

per

atu

re (

oC

)



Figure 6-76 shows the temperature profiles across the depth of the central studs

measured at mid-height. The maximum temperature differences (HF temperature – CF

temperature) across the central studs at 60 and 90 minutes were observed to be 200C

and 1200C, respectively, compared to 1000C and 3150C in the cavity insulated Test

Specimen 5 at the same times). Towards the end of the test the temperature difference

across the central studs had dropped below 1000C.


The specimen was loaded to 15 kN at ambient temperature when there were no signs

of lateral deformation. Axial deformations of the studs were seen to be about 3 mm as

shown in Figure 6-78 (a). During the fire test the studs expanded due to heat. The total

average elongation of the studs towards the end of the test was about 15 mm.

Figures 6-79 (a) to (c) give the lateral deflections of the studs at the top, middle and

lower levels. The wall was found to bow towards the furnace within a few minutes of

starting the fire test. The deflections were maximum at mid-span and started

developing faster beyond 90 minutes. Stud 2 underwent the maximum lateral

deformation of 22 mm by 114 minutes (Figure 6-79 (b)), beyond which it stopped

bowing any further until 124 minutes at which time the hot flange temperature of the


stud had reached 4700C at the centre (Figure 6-75 (a)) and the temperature difference

across the stud was 1200C (Figure 6-76). At this time, Stud 3 had reached its

maximum lateral deflection of 22 mm (Figure 6-79 (b)).

Stud 2 reversed its direction of deformation and started to straighten out from about

124 minutes. This was soon followed by Stud 3 straightening out at 127 minutes at

which time its hot flange temperature was about 4600C at centre (Figure 6-75 (a)) and

the temperature difference across the depth was 1300C (Figure 6-76). Both central

studs continued to deform rapidly in the outward direction. By 132 minutes Stud 2

had exhausted the free expansion scope provided for it in the loading frame as the

loading plate of Stud 2 came into contact with the lower beam rendering the jack non-

functional (see Figure 6-77). This meant that beyond 132 minutes and until the failure

of the specimen, the load on Stud 2 could have gone up well beyond 15 kN due to the

temperature stresses induced in the stud on the prevention of its thermal expansion.

This was soon followed by the other studs also running out of free thermal expansion

scope leading to the collapse of the frame due to increased thermal stresses by about

136 minutes. All the plasterboards were seen to be intact throughout the fire test

(Figure 6-74). As observed for the previous specimens the structural failure occurred

within 15 minutes of the first observed reversal in lateral deformation of the central

studs.

(a)

Figure 6-77: View of Loading Arrangement



(b)

Figure 6-77: View of Loading Arrangement

-4.0

-3.5

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0 2 4 6 8 10 12 14 1Load (kN)

Def

orm

atio

n (

mm

)

6




Allowance for free thermal expansion

Jack

Loading Plate

-5

0

5

10

15

20

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Def

orm

atio

n (

mm

)




-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Def

lect

ion

(m

m)





-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

De

fle

cti

on

(m

m)



-30

-25

-20

-15

-10

-5

0

5

10

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Def

lect

ion

(m

m)





10

11

12

13

14

15

16

17

18

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Lo

ad (

kN)


4) Specimen Report




The average temperature of the unexposed surface of the test specimen at the end of

120 minutes from the commencement of the test was 570C indicating a rise of 410C

above the ambient temperature of 160C. The maximum temperature of the unexposed

surface at that time was 590C.


The fire side Plasterboards 1 and 2 remained attached to the steel frame until the




time was 23 mm.





f) Performance








The specimen was subjected to fire for 124 minutes until the wall specimen failed.

Figure 6-81 (a) shows the specimen after the fire test. Plasterboards 1 and 2 had fully

collapsed and the cellulose fibre between them was totally burnt out (Figure 6-81 (a)).

Figure 6-81 (b) shows the outward movement of the studs caused by the local

compressive failure of the hot flanges of the studs. Studs 1, 2 and 3 were severely

damaged by the local buckling of the hot flange and web whereas Stud 4 was seen to

be unaffected (see Figure 6-81 (c)). Out of plane deformations of Studs 2 and 3 had

also ruptured the ambient side plasterboards (Figures 6-81 (d) and (e)). The tracks

were in good condition providing good support to the studs (Figure 6-81 (f)).

Remains of the external cellulose fibre insulation

Plasterboard 1

Plasterboard 2

Pb3-Cav surface

(a) Front View Showing the Collapse of External Plasterboard and Burn-out of External Insulation

(Cellulose fibre)




(b) Side View of Wall after Removing the External Plasterboards

Stud 4 Stud 3

Stud 2 Stud 1

(b) Failure Modes of Studs 1, 2 and 3



(d) Local Compressive Failure of Flanges and Web in Stud 2

(e) Local Compressive Failure (f) Top Track in Good of Hot Flange and Web in Stud 2 Condition





Plasterboards 1 and Insulation (Pb1-Ins)

The initial response of this interface was very much similar to Specimens 7 and 8. The

first phase displayed a quick temperature rise up to 800C by the end of 5 minutes (see

Figure 6-82 (a)). The second phase of near constant temperature lasted until 18

minutes, beyond which the temperature gained sharply in the third phase reaching

4000C by 29 minutes and 5500C by 48 minutes. The rate of temperature growth was

reduced after 30 minutes. The thermocouple positioned at the centre along mid-height

between Studs 1 and 2 in this interface showed a continuous increase in its

temperature finally merging with the fire curve at about 88 minutes indicating a


breach at that time of the external plasterboard in this area (refer Figure 6-82 (b)). The

other two thermocouples measuring the interface temperature and placed at the mid-

height of the wall between the remaining studs merged with the fire curve at 97

minutes and 125 minutes, respectively.



The temperature profile of this interface almost coincided with that of Pb1-Ins curve

until about 20 minutes. The temperature gained beyond 20 minutes, reaching 4000C

by 53 minutes. Figure 6-82 (b) shows the merging of Ins-Pb2-L curve with Pb1-Ins-L

at about 84 minutes indicating the total disintegration of the insulation at mid-height

region between Studs 1 and 2 at this time. The temperature profiles of Pb1-Ins-R and

Ins-Pb2-R were merged from the start of the test until the end of the test. The only

possible explanation for this would be that somehow, the two thermocouple wires

measuring the two interface temperatures must have got entangled with the hot

junctions of both touching each other resulting in identical readings. The Ins-Pb2-M

curve intersected the Pb1-Ins-M curve at about 106 minutes indicating the insulation

burn out at the mid-height of wall between Studs 2 and 3.


2 (Pb2-Cav)

The temperature of this surface increased very slowly from the ambient and reached

approximately 1000C by 70 minutes from the start of the test (compared with 55

minutes observed in Specimen 6 using cellulose insulation in the cavity). The third

phase started beyond 70 minutes with an almost linear temperature growth rate and

reached 4000C by 108 minutes and 5000C by 120 minutes (compared with 68 minutes

and 72 minutes in Specimen 6). After 125 minutes the temperature difference across

the plasterboard thickness was approximately 4750C, which reduced sharply soon

after indicating it’s partial or full collapse.


3 (Pb3-Cav)

As seen in Specimens 7 and 8 the temperature profile of Pb3-Cav surface traced the

Pb2-Cav profile with the lag not exceeding 750C during the test. The temperature of


the surface increased suddenly near the end of the test (125 minutes) indicating the

breaching of Plasterboard 2 (i.e. the exposed base layer plasterboard).



The temperature of this interface remained under 1000C until the end of the test. The

temperature difference across the thickness of Plasterboard 3 was approximately

4000C towards the end of the test.



The temperature of this interface remained under 900C for the entire duration of the

test (i.e. 125 minutes).


The surface recorded temperature values less than 700C until the end of the test.

0

100200

300

400500

600

700800

900

10001100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

per

atu

re (

oC

)

AS 1530 Furnace FS Pb1-Ins Ins-Pb2Pb2-Cav Pb3-Cav Pb3-Ins Ins-Pb4 Amb




0

100

200

300

400

500

600

700

800

900

1000

1100

1200

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

per

atu

re (

oC

)

AS 1530 Furnace FS-L FS-M FS-R Pb1-Ins-L

Pb1-Ins-M Pb1-Ins-R Ins-Pb2-L Ins-Pb2-M Ins-Pb2-R Pb2-Cav-LPb2-Cav-M Pb2-Cav-R Pb3-Cav-L Pb3-Cav-M Pb3-Cav-R Pb3-Ins-L

Pb3-Ins-M Pb3-Ins-R Ins-Pb4-L Ins-Pb4-M Ins-Pb4-R



b) Steel Surfaces

Figures 6-83 (a), (b) and (c) give the temperature profiles for the hot flanges, webs

and cold flanges along the central studs with the second phase extending until 70, 75

and 80 minutes (compared to 50, 55 and 65 minutes in Test Specimen 6 using

cellulose fibre as cavity insulation). Temperature variations along the length were

more pronounced in the hot flanges than in the web and cold flange elements. The hot

flange temperature profiles very closely traced the Pb2-Cav temperature profile. A

maximum temperature difference of 1500C could be noted in the hot flanges along the

length of the central studs. Temperature variations along the length in the web and

cold flange elements appeared towards the end of the test. The maximum recorded hot

flange temperatures at 60, 80 and 100 minutes were 1000C, 2300C and 4200C,

respectively (compared to 2200C, 4600C and 6000C in the cellulose fibre cavity

insulated Test Specimen 6 at the same times as seen in Figure 6-60a). The maximum

temperature variation across the cross-section was observed to be about 2000C in Stud


2 at 100 minutes (compared to the temperature difference of 3300C in Stud 2 at 94

minutes for Test Specimen 6, Figure 6-61).

0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Te

mp

era

ture

(oC

)



0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

per

atu

re (

oC

)






0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

per

atu

re (

oC

)




0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140Time (min)

Tem

per

atu

re (

oC

)




Each stud of the specimen was loaded at the ambient temperature conditions up to 15

kN without any signs of lateral deformations. The studs however suffered an axial


shortening of about 3.5 mm (Figure 6-85 a). On starting the fire test, the specimen

was seen to gradually start bowing towards the furnace. This lateral deformation

increased rapidly beyond 80 minutes. Both the central studs deformed more than the

end studs, with Stud 2 reaching a maximum lateral deformation of 22 mm by 110

minutes (Figure 6-86 b). At this time the temperature of the hot flange at the centre

was 5000C and the temperature difference across the stud depth was 1800C (Figure 6-

84). Beyond 110 minutes Stud 2 reversed its direction of lateral deflection and started

to straighten out. Stud 3 reached a maximum of 20 mm by the end of 120 minutes

after which it also reversed its direction of lateral deformation (Figure 6-86 (b)). At

120 minutes the central hot flange temperature in Stud 3 had reached 5300C and the

temperature difference across the cross-section was 1600C (Figure 6-84). After

reversing, the lateral deformations in both the studs progressed very rapidly, leading

to the failure of the frame at 124 minutes (Figure 6-87). The exposed Plasterboard 1

had partially fallen off about the central portion of the wall after 73 minutes as

indicated by the step in the temperature profile of Pb1-Ins-M (Figure 6-82 (b)).

However Plasterboard 2 was intact throughout the fire test offering protection to the

steel frame.

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

0 2 4 6 8 10 12 14 16

Load (kN)

Def

orm

atio

n (

mm

)


(a) Axial Deformation -Load Profiles of Test Specimen 9 at Ambient

Temperature

9

Figure 6-85: Axial Deformation Plots for Studs of Test Specimen


-10

-5

0

5

10

15

20

25

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Def

orm

atio

n (

mm

)




-30

-25

-20

-15

-10

-5

0

5

10

15

20

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Def

lect

ion

(m

m)

Stud 2 Stud 3




-30

-25

-20

-15

-10

-5

0

5

10

15

20

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Def

lect

ion

(m

m)



-30

-25

-20

-15

-10

-5

0

5

10

15

20

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Def

lect

ion

(m

m)

L1 L2 L3 L4




10

11

12

13

14

15

16

17

18

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Lo

ad

(k

N)


6.5

st: 00%

re

the unexposed surface of the test specimen towards the 0 0

1 collapsed partially near the central portion of the wall at

.9.4) Specimen Report

a) Date of Test:

b) Severity of Te 1

c) Specimen Temperatu

The average temperature of

end of the test was 58 C indicating a rise of 42 C above the ambient temperature of



The fire side Plasterboard

about 73 minutes from the start of the test, whereas Plasterboard 2 remained attached

to the steel frame until the failure of the test specimen. The lateral deflection of the

Test Specimen was initially towards the furnace and reversed in direction at the end of

110 minutes from the commencement of the test. The maximum deflection at mid-

height of the wall at that time was 22 mm. The structural failure of the frame resulted

within 15 minutes of the first stud reversing its lateral deformation.


) Failure Criterion

s deemed to have failed at approximately 124 minutes from the


rved in respect of the following criteria:

s.

s

sio

nd plateau in the time-temperature graph of the plasterboards due to

the conversion into steam of the remaining portion (approximately 25%) of the

e a slight negative pressure

on the surface due to the rising hot gases favoring the moisture movement towards the

e

The test specimen wa

f) Performance

Performance obse

Structural adequacy - Failure at 124 minute

Integrity - No failure at 124 minute


6.6: Discus ns

The expected seco

chemically bound water present in the plasterboard was never observed, probably

because the conversion of the small quantity of bound water into steam was very

quick and the steam was able to escape across the plasterboard thickness very fast

without much of re-condensation or migration towards the ambient side (which was

possible during the formation of the first plateau) due to the already present numerous

shrinkage cracks over the entire body of the plasterboard.

The extreme heat on one side of the plasterboard can caus

hot surface, quickly removing the steam from the already gaping cracks caused due to

the plasterboard shrinkage. Also with the passage of time during the fire exposure the

total surface area of the plasterboard receiving heat from the furnace increases due to

the presence of numerous shrinkage cracks, which probably offsets the formation of

the second temperature plateau due to the conversion of the final portion of the

chemically bound water in the plasterboard.

Chapter 7: Discussions and Recommendations

In this research, the fire performance of non-load bearing wall test specimens was

studied with and without cavity insulation. Test specimens with insulation placed

outside the cavity and between external plasterboards were also tested and the fire

performance when subjected to cellulosic fire curve was studied and compared with

the conventional wall systems. Similar tests were carried out on large scale load

bearing wall test specimens subjected to a load ratio of 0.2. Effects of different types

of insulation material such as Glass fibre, Rock wool fibre and Cellulose fibre on the

thermal performance of non-load bearing and load bearing walls were studied and

compared in the earlier chapters of this thesis.

This chapter presents the outcomes of the tests performed on non-load bearing and

load bearing wall specimens with and without the use of insulating material.

7.1: Discussions

7.1.1 Comparison between conventional non-load bearing wall models and

proposed externally insulated wall models

The fire performance of the externally insulated wall specimens was considerably

better than the cavity insulated specimens as the steel stud frame in the former case

was well protected by the external layer of insulation. The temperature plateau in the

steel studs of externally insulated specimens lasted up to approximately 75 minutes

while it was 55 minutes for the cavity insulated specimens. The studs in the externally

insulated specimens not only enjoyed an extended period of almost constant

temperature, but also were benefitted by the lower temperature growth rates following

the plateau when compared with the rapid escalation of temperatures in the case of

cavity insulated specimens.

The absence of insulation in the cavity of externally insulated specimens helped to

equalize the temperatures across the stud cross-sections more rapidly through

radiation than in the cavity insulated specimens. The large temperature gradients

developed across the studs in the cavity insulated specimens forced the hot flanges to


expand much more than the cold flanges causing the walls to deflect laterally towards

the furnace. Externally insulated specimens on the other hand displayed minimum

lateral deflections during the tests.

Figures 7-1 to 7-3 compare the effects of external insulation to that of cavity

insulation. A clear separation of the time-temperature profiles of the studs in the two

kinds of wall specimens shows that externally insulated specimens took longer times

to reach the equivalent temperatures. This clearly demonstrates the resulting benefits

of placing the insulation outside the steel stud frame.

0100

200300

400500

600700

800900

10001100

12001300

0 20 40 60 80 100 120 140 160 180 200 220

Time (min)

Tem

per

atu

re (

oC

)

Sp4 HF Sp4 W Sp4 CF Sp7 HF Sp7 W Sp7 CF

Figure 7-1: Time-temperature Profiles for the Central Stud in Non-Load Bearing Wall Test Specimens 4 and 7 (Glass fibre insulation)

Note: Sp4/7 HF/W/CF: Time-temperature profile followed by the HF/W/CF of the central

stud in Test Specimen 4 (cavity insulations) and Test Specimen 7 (external insulation)


0100200300400500600700800900

1000110012001300

0 20 40 60 80 100 120 140 160 180 200 220 240

Time (min)

Tem

per

atu

re (

oC

)

Sp 5 HF Sp 5 W Sp 5 CF Sp 8 HF Sp 8 W Sp 8 CF

Figure 7-2: Time-temperature Profiles for the Central Stud in Non-load Bearing Wall

Test Specimens 5 and 8 (Rock fibre insulation)

Note: Sp5/8 HF/W/CF: Time-temperature profile followed by the HF/W/CF of the central stud in Test Specimen 5 (cavity insulations) and Test Specimen 8 (external insulation)

0100200300400500600700800900

1000110012001300

0 20 40 60 80 100 120 140 160 180

Time (min)

Tem

per

atu

re (

oC

)

Sp 6 HF Sp 6 W Sp 6 CF Sp 9 HF Sp 9 W Sp 9 CF

Figure 7-3: Time-temperature Profiles for the Central Stud in Non-Load Bearing Wall

Test Specimens 6 and 9 (Cellulose fibre insulation)

Note: Sp6/9 HF/W/CF: Time-temperature profile followed by the HF/W/CF of the central stud in Test Specimen 6 (cavity insulations) and Test Specimen 9 (external insulation)


0100200300400500600700800900

1000110012001300

0 20 40 60 80 100 120 140 160 180 200 220Time (min)

Tem

per

atu

re (

oC

)

CI-GF S2-HF CI-RF S2-HF CI-CF S2-HFCP-GF S2-HF CP-RF S2-HF CP-CF S2-HF

Figure 7-4: Time-temperature Profiles for the Central Stud

Hot Flanges in Non-Load Bearing Wall Test Specimens 4 to 9

Note: CI-GF/RF/CF S2-HF: Time-temperature profile followed by the hot flange of the central stud in the specimen using glass fibres/rock fibres/cellulose fibres as cavity insulation

Amongst the three types of insulations tested, rock fibre gave the maximum

separation (see Figure 7-2). This was possibly due to its superior insulating properties

over the other two. The insulation gave the best results when placed on the outside of

studs. The lower levels of conductivity which had caused extensive damage when

placed in the wall cavity were seen to offer the best protection to the studs by

maintaining their temperatures at lower levels with minimum temperature gradients

across them. This ensures minimum lateral deflections and maximum periods of fire

endurance.

The wall specimens with cellulose fibre insulation on the other hand showed

minimum separation in the time-temperature profiles of their studs (see Figure 7-3).

This was considered to be due to the rapid disintegration of the cellulose fibre in the

wall cavity reducing it to the case of a wall specimen without the cavity insulation.

This actually helped by lowering the temperature gradients in the studs as more

energy could be dissipated into the cavity. Thus the disintegration of the cellulose

fibre in the cavity actually helped the studs to survive longer. The beneficial effect of

placing the insulation on the outside is also seen in Figure 7-3.


The temperature escalations of the hot flanges in the cavity insulated and externally

insulated wall specimens are shown in Figure 7-4, which clearly brings out the

benefits of using the insulation on the outside rather than in the cavity of the wall

specimens. Among the externally insulated specimens rock fibre insulation was seen

to give the best results.

Figures 7-5 ((a) to (d)) give the hot flange temperature values of the central stud for

Specimens 3 to 9 at 60, 90, 120 and 150 minutes from the start of the test . The hot

flange temperatures of the studs in externally insulated specimens surpassed that of

the non-insulated specimen (Specimen 3) after 150 minutes. This was probably due to

the heat redirected towards the cavity by the external insulation on the ambient side.

By the end of 90 minutes the hot flange temperatures of the cavity insulated

specimens ranged from 1.5 to 3.5 times that of the externally insulated specimens.

(a) 60 minutes (b) 90 minutes

(c) 120 minutes (d) 150 minutes Figure 7-5: Hot Flange Temperatures of the Central Stud in NLB Wall Test

Specimens 3 to 9


Note:

CI-GF: Non-load bearing wall specimen using Glass Fibre (GF) as cavity insulation (CI)

CI-RF: Non-load bearing wall specimen using Rock Fibre (RF) as cavity insulation (CI)

CI-CF: Non-load bearing wall specimen using Cellulose Fibre (CF) as cavity insulation (CI)

CP-GF: Non-load bearing wall specimen using Glass Fibre (GF) as external insulation in

composite panels (CP)

CP-RF: Non-load bearing wall specimen using Rock Fibre (RF) as external insulation in


CP-CF: Non-load bearing wall specimen using Cellulose Fibre (CF) as external insulation in


Figures 7-6 ((a) to (d)) give the temperature difference across the central stud in Test

Specimens 3 to 9. For up to 120 minutes from the start of the test the temperature

difference across the externally insulated wall specimens can be seen to be much less

than the non-insulated and the cavity insulated specimens. At 90 minutes from the

start of the test the temperature difference across the cavity insulated specimens was

2.5 to 7.5 times more than the externally insulated wall specimens. The drop in the

temperature difference values of cavity insulated Test Specimen 5 (CI-RF) at about

150 minutes is probably due to the effect of the disintegration of insulation and fire

side plasterboards, thus reducing the temperature difference across the stud cross-

section.


Figure 7-6: Temperature Difference across the Central Studs in NLB Wall Test Specimens 3 to 9


(c) 120 minutes (d) 150 minutes

Figure 7-6: Temperature Difference across the Central Studs C/S in NLB Wall Test Specimens 3 to 9

Figures 7-7 ((a) to (d)) show the temperature of the ambient surface of the external

plasterboard 2 for non-load bearing (NLB) wall Test Specimens 3 to 9. In the case of

cavity insulated Test Specimens 4 to 6 this was the temperature of the Pb2,Ins

interface and in the case of non-insulated and externally insulated Test Specimens 3

and 7 to 9, respectively, it was the Pb2-Cav temperature.

At the end of 90 minutes, the temperature of the Pb2,Ins interface is approximately 2

to 5.5 times greater than the Pb2-Cav temperature of the non-insulated or externally

insulated Test Specimens. At 150 minutes, the value in the case of Test Specimen 5 is

seen to cross 10000C due to a breach in the external platerboards.






(Pb2,Ins in case of cavity insulated Test Specimens and Pb2-Cav in case of non-insulated and externally insulated Test Specimens)

Figures 7-8 ((a) to (d)) show that the temperature on the ambient face of the externally

insulated test specimens is consistently lower than the cavity insulated test specimens

during the entire test.






The fall off times of Plasterboard 2 in the case of externally insulated specimens 7 and

8 was seen to be around 200 minutes, almost an hour more than that observed in

cavity insulated specimens (see Figure 7-9). This could be due to the following three

factors.

1) Extra protection offered to them by the layer of insulation on the fire side.

2) Lower stress levels in the plasterboard due to smaller lateral and axial deformations

as compared to cavity insulated specimens.

3) Lower rate of temperature growth in the plasterboard on account of dissipation of

heat into the cavity as opposed to the accelerated temperature growth due to the

redirected heat from the insulation in the cavity insulated specimens.

Figure 7-9: Fall off Times of Plasterboard 2 in Non-Load Bearing Wall Test Specimens 4 to 9


The fall off times of Plasterboard 2 is very critical as its collapse would expose the

steel frame directly to fire and commence the failure of the wall. The thermal

insulation property of the externally insulated walls was also found to outperform the

cavity insulated specimens. The temperatures of the cavity facing surface of

Plasterboard 2 in the case of externally insulated specimens was seen to be one third

to one half of the temperatures of the corresponding face in the cavity insulated

specimens. This was the major factor responsible for delaying, the fall off time of

Plasterboard 2 in the case of externally insulated specimens. The ambient side

temperatures of the externally insulated specimens were also seen to be consistently

lower than those of cavity insulated specimens at any given time (see Figure 7-10).

0

20

40

60

80

100

120

140

160

180

0 20 40 60 80 100 120 140 160 180 200 220

Time (min)

Tem

per

atu

re (

oC

)

CI-GF CI-RF CI-CF CP-GF CP-RF CP-CF

Figure 7-10: Ambient Side Time-temperature Profiles of Test Specimens 4 to 9

Note: CI-GF/RF/CF: Test Specimen using glass fibres/rock fibres/cellulose fibres as cavity insulation

After 3 hours of test the ambient side temperatures of the cavity insulated specimens

had exceeded 1000C whereas the temperatures of the externally insulated specimens

were around 800C. The lowest temperatures were recorded by Specimen 8 using rock

fibre as external insulation. All of these observations imply that the new wall system

with external insulation is likely to provide improved performance under the three fire

rating criteria of stability, integrity and insulation.


7.1.2 Comparison between conventional load bearing wall models and proposed

externally insulated wall models.

Figures 7-11 and 7-12 show the thermal responses of the central studs in the cavity

insulated load bearing wall specimens along with the externally insulated load bearing

wall specimens using rock fibre and cellulose fibre insulation, respectively. Graphs

comparing the performance of cavity insulated and externally insulated specimens

using glass fibre as insulating material (Specimens 4 and 7 respectively) could not be

made as the furnace malfunctioned during the testing of Specimen 7. The graphs were

drawn using the average temperature values of the central studs at mid-height of the

wall.

In the case of cavity insulated specimens, the temperature plateau (second phase) of

the hot flanges was seen to last only up to 40 minutes in comparison to 65 minutes in

the externally insulated specimens. The hot flange temperatures of cavity insulated

specimens were seen to rise very rapidly with large temperature differences across the

stud cross-sections due to the presence of insulation between the flanges in the wall

cavity. The hot flange temperatures of externally insulated wall specimens on the

other hand were seen to rise gradually with a small temperature difference across the

stud cross-sections due to the faster transfer of heat by radiation across the empty

cavity.

0

100

200

300

400

500

600

700

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140Time (min)

Tem

per

ature

(oC

)

Sp 5 HF Sp 5 W Sp 5 CFSp 8 HF Sp 8 W Sp 8 CF

Figure 7-11: Average Time-temperature Profiles for the Central Studs in Load

Bearing Wall Test Specimens 5 and 8


0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100 110 120 130

Time (min)

Tem

per

atru

e (o

C)

Sp 6 HF Sp 6 W Sp 6 CFSp 9 HF Sp 9 W Sp 9 CF


Note:

Sp 5/6/8/9 HF: Average time-temperature profile followed by the hot flanges of the central studs in Test Specimens 5/6/8/9

Sp 5/6/8/9 W: Average time-temperature profile followed by the webs of the central studs in Test Specimens 5/6/8/9

Sp 5/6/8/9 CF: Average time-temperature profile followed by the cold flanges of the

central studs in Test Specimens 5/6/8/9

Figures 7-13 ((a) to (d)) give the average hot flange temperature values for the central

studs of load bearing wall Test Specimens 3 to 9 at 60 , 80, 100 and 120 minutes from

the start of the test. The hot flange temperatures of the externally insulated specimens

are noted to be substantially lower than the hot flange temperatures of non-insulated

and cavity insulated test specimens. At 80 minutes from the start of the test, the

externally insulated specimens had hot flange temperatures less than half of the

temperatures recorded by the cavity insulated specimens. As the non-insulated and the

cavity insulated test specimens had collapsed before 120 minutes, their values are not

displayed in Figure 7-13 (d).




Figure 7- 13: Average Hot Flange Temperatures of the Central Studs of Load

Bearing Wall Test Specimens 3 to 9

Figures 7-14 ((a) to (d)) give the average temperature difference across the central

studs for load bearing wall Test Specimens 3 to 9. The temperature differences across

the studs in the externally insulated test specimens were seen to be very much lower

than the values observed in the cavity insulated test specimens during the entire fire

test. The non-insulated Test Specimen 3 had values intermediate to those of cavity

insulated and externally insulated test specimens at 80 minutes from the start of the

test. The temperature difference values for the cavity insulated specimens were 3 to 5

times higher than the values of the externally insulated specimens, whereas those

values for the non-insulated specimen were about 1.5 times higher.




Figure 7-14 : Average Temperature Difference across the Central Studs for Load

Bearing Wall Test Specimens 3 to 9

The higher temperature differences across the stud cross-sections in the cavity

insulated specimens led to higher lateral deformations as compared to the externally

insulated specimens. Figures 7-15 and 7-16 show the temperature differences across

the stud cross-sections and the corresponding lateral deformations (LD) for the cavity

insulated and externally insulated specimens using rock fibre and cellulose fibre

insulations, respectively. The lateral deformations were seen to be proportional to the

temperature difference across the stud cross-sections.


0

50

100

150

200

250

300

350

400

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Time (min)

Tem

p D

iffe

ren

ce (

HF

-CF

) (

oC

)

0

5

10

15

20

25

30

35

40

Lat

eral

Def

orm

atio

n (

mm

)

(HF-CF) CI RF (HF-CF) CP RF LD-CI-RF LD-CP-RF

Figure 7-15: Average Temperature Difference across Central Studs and their Lateral Deformations (LD) versus Time for Test Specimens 5 and 8

0

50

100

150

200

250

300

350

400

450

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

p D

iffe

ren

ce (

HF

-CF

) (

oC

)

0

5

10

15

20

25

30

35

40

45

La

tera

l D

efo

rma

tio

n (

mm

)

(HF-CF) CI CF (HF-CF) CP CF LD-CI-CF LD-CP-CF


Figure 7-17 compares the average lateral deformation of the central studs at the end of

60, 80, 100 and 120 minutes from the start of the test for Test Specimens 3 to 9. At 60


minutes from the start of the test, the lateral deformations in the cavity insulated

specimens was close to 16 mm as compared to about 9 mm in the case of externally

insulated specimens and 10 mm in the non-insulated Test Specimen 3. By the end of

80 minutes, the lateral deformation in the cavity insulated specimens had reached 30

mm, whereas it was still less than 11 mm for the externally insulated specimens and

14 mm for the non-insulated test specimen The lower lateral deformations of Test

Specimens 3 and 5 at 100 minutes as compared to their values at 80 minutes is on

account of the reversal of lateral deformation leading to progressive failure of the test

specimens.



Figure 7- 17:Average Lateral Deformations of the Central Studs in Load Bearing Wall Test Specimens


The average hot flange and cold flange temperature profiles of the central studs for

Test Specimens 4 to 9 are shown in Figures 7-18 and 7-19, respectively. A distinct

separation in the time-temperature profiles of the studs in the two kinds of wall

specimens shows that the studs of the externally insulated wall specimens are well

protected and thus take longer time to reach the temperatures attained by the studs in

the cavity insulated specimens.

The hot flange and cold flange temperature profiles of the cavity insulated specimens

show that the thermal response of the studs is identical for rock fibre and cellulose

fibre insulation whereas for glass fibre insulation the stud temperatures are marginally

higher. This signifies a low influence on the stud temperatures by the material of

insulation used in the cavity. This is probably because, in the cavity insulated

specimens, the insulation is on the ambient side of the hot flange and thus incapable

of offering any protection to it.

In the case of externally insulated specimens, it is seen that the temperature profiles of

the studs are well separated implying the effect of the material of insulation on the

stud temperatures. From Figures 7-18 and 7-19 it is clear that rock fibre insulation

offers the maximum protection to the studs.

0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140

Time (min)

Tem

per

atru

e (o

C)

Sp 4 HF Sp 5 HF Sp 6 HF Sp 8 HF Sp 9 HF

Figure 7-18: Average Time-temperature Profiles of Hot Flanges for the Central Studs in Test Specimens 4 to 9


0

50

100

150

200

250

300

350

400

450

500

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140Time (min)

Tem

per

atu

re (

oC

)

Sp 4 CF Sp 5 CF Sp 6 CF Sp 8 CF Sp 9 CF

Figure 7-19: Average Time-temperature Profiles of Cold Flanges for the Central Studs in Test Specimens 4 to 9

Figure 7-20: Temperature Difference across the Central Studs in Cavity Insulated and Externally Insulated Specimens

The temperature difference between the hot and cold flanges of the central studs with

the passage of time during the fire test is shown in Figure 7-20. The graphs for all the

cavity insulated specimens using different types of insulation are seen to lie in a very


narrow band signifying the very low influence of the material of insulation on the

building up of cold flange temperatures. This is probably because any insulation by

virtue of its physical presence essentially serves the main function of eliminating the

transfer of heat across the wall cavity by radiation and convection which essentially

are the faster modes of heat transfer as compared to conduction. No cavity insulation

can reduce the transfer of heat towards the cold flange by conduction along the

metallic cross-section of the stud. Thus the cold flange picks up heat from the hot

flange by conduction along the web, which would be the fastest mode of heat transfer

in the case of cavity insulated specimens. Because of the very low conductivity of the

cavity insulating material as compared to steel, most of the heat gets directed and

channeled along and across the steel studs which act as the heat sink thus raising their

body temperatures much faster than in the case of non-cavity insulated specimens.

This makes the very presence of cavity insulation a threat to the survival of steel

during fire conditions.

Externally insulated specimens on the other hand can offer a much higher level of

protection to the studs as they are installed on the fire side of the studs thus

minimizing the transfer of heat by radiation (by virtue of their physical presence) and

conduction (on account of their low conductivity). Hence the quality of insulation

used externally would directly influence the level of fire protection offered to the

studs. Rock fibre insulation when used externally was seen to give the maximum

protection.

Figure 7-21 shows the average temperature profiles on the ambient surface of the

exposed base layer plasterboard 2 for the load bearing wall specimens. The profiles of

the cavity insulated specimens were seen to lie in a very narrow band with

temperature profiles well above the externally insulated and non-insulated wall

specimens. The cavity insulation caused the temperature profiles to rise sharply by

blocking and redirecting the heat flow back on to the cavity facing surface. The

temperature profiles of the externally insulated wall specimens were found to be the

most favorable. The externally insulated specimen using rock fibre as insulation gave

the best results.


0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Time (min)

Tem

per

atu

re (

oC

)

LBW-CI-GF LBW-CI-RF LBW-CI-CFLBW-No Ins LBW-CP-RF LBW-CP-CF

Figure 7-21: Average Time-temperature Profiles of Pb2-Cav Surface in Test Specimens 3 to 9

Note: LBW: Load bearing wall CI: Cavity Insulated

CP: Composite Panel (Externally insulated)

Figures 7-22 (a) to (d) show the temperature profiles at 60, 80, 100 and 120 minutes,

respectively. At 80 minutes the temperature profiles of non-insulated and cavity

insulated test specimens ranged from 2 to 4 times that of externally insulated test

specimens. At 100 minutes (Figure 7-22 (c)) the plasterboard temperature of cavity

insulated specimens approached 7000C, whereas, it was below 3500C in the externally

insulated specimens. As the non-insulated and cavity insulated specimens collapsed

before 120 minutes, only the values of the externally insulated wall specimens is

shown in Figure 7-22 (d).




Figure 7- 22: Average Time-Temperature Profiles on the Ambient Side of the

Exposed Base Layer Plasterboard 2 in LBW Test Specimens 3 to 9

The ambient side temperatures of all the wall specimens were observed to be under

700C and well below the insulation failure temperature (see Figures 7-23 (a) to (d)).

The failure of the specimens was always by the structural failure of the studs and

never by insulation or integrity failure. In some of the cavity insulated specimens, the

external plasterboards collapsed prior to stud failure thus hastening the collapse of the

wall by exposing the steel frame to direct furnace heat.




Figure 7- 23: Ambient Side Temperatures of LBW Test Specimens 3 to 9

Table 7-1 gives the failure times of the load bearing wall specimens tested under a

constant load of 15 kN on each stud during the fire test. Only the failure times of Test

Specimens 3 to 9 can be compared as they have two plasterboards on either side of the

steel frame. The failure times of the cavity insulated specimens was noted to be less

than the failure time of Test Specimen 3 (without cavity insulation), however, the

failure times of externally insulated specimens was found to be the maximum. The

failure time of Test Specimen 7 cannot be considered as the furnace had

malfunctioned for some time during the test. The failure times of Test Specimens 8

and 9 could have been higher if the studs were allowed free thermal expansion near


the end of the test without additional load and stresses building up in the studs due to

the closure of the expansion gaps provided in the loading mechanism of the jacks.

Table 7-1: Failure Times of Test Specimens

Specimen No.

Configuration Test Insulation Ave. HF Temp. *

Failure Time (min)

01 Ambient None -

02 Fire None 48

03 Fire None 561 111

04 Fire Glass Fibre (Cavity

Insulation)

671 101

05

Fire Rock Fibre (Cavity

Insulation)

642 107

06


(Cavity Insulation)

720 110

07 Fire Glass Fibre (External

Insulation)

522 181

08 Fire Rock Fibre (External

Insulation)

524 136

09 Fire Cellulose Fibre


610 124

*Average hot flange temperature at the centre for the middle studs at failure

For cavity insulated specimens the average hot flange temperature at the time of wall

failure was observed to be 6770C, whereas for the externally insulated specimens the

average hot flange temperature at wall failure was 5670C. Table 7-2 shows the stud

reversal times signifying the local buckling of hot flanges along with the


corresponding hot flange temperature. The average hot flange temperature for the

cavity insulated specimens initiating the local buckling of hot flange and reversal in

lateral deformation was noticed to be approximately 5700C. The higher temperature of

7060C in Stud 3 of Test Specimen 6 was not considered as it appears to be faulty. The

studs in Test Specimens 8 and 9 were seen to fail at an average hot flange temperature

of approximately 5000C slightly lower than the cavity insulated specimens, probably

because of the increased load on the individual studs after the closure of the free

thermal expansion gap in the loading mechanism.

Table 7-2: Stud Reversal Times for Cavity Insulated and Externally Insulated Specimens along with the Corresponding Temperatures

Sp.

No.

Specimen Insulation Reversal time

Stud 2

Reversal time

Stud 3

HF temp.

Stud 2

HF temp.

Stud 3

4

Glass Fibre

(Cavity Insulation)

85 95 582 566

5

Rock Fibre

(Cavity Insulation)

92 103 551 570

6

Cellulose Fibre

(Cavity Insulation)

96 107 582 706

7

Glass Fibre


174 179 - 500

8

Rock Fibre


123 127 468 457

9

Cellulose Fibre


110 120 529 523


7.2: Simplified Method for the Determination of Failure Times of Wall

Specimens

A graphical method is developed to yield approximate failure times of load bearing

wall systems. The wall systems considered are non-insulated, cavity insulated and

externally insulated wall specimens using glass fibre, rock fibre or cellulose fibre as

the insulating material.

The failure time of a load bearing wall system with the studs assumed to be

effectively restrained laterally about the minor axis depends upon the load ratio, the

type and placement of the insulation used in the wall system, the critical temperature

at which the studs undergo local buckling and lateral deflection. To help determine

the critical temperature the results from Chapter 3 are used in which the reduced

mechanical properties of cold-formed steel at elevated temperatures are presented.

Figures 7-24 shows the reduction in yield strength of 1.15 mm G500 cold-formed

steel at elevated temperatures, determined based on the 0.2% proof stress method

(Refer Eqs. 1(a) to (c) in Chapter 3). The reduction in yield strength occurs from a

temperature of about 3000C at a steady rate with a constant gradient and reaches about

12% of its original value by 6000C.

Figure 7-24: Variation of Yield Strength Reduction Factor of 1.15 mm G500

Steel with respect to Temperature


The load bearing wall specimens tested were subject to a load of 15 kN per stud

giving a load ratio of 0.2. Assuming the cross-sectional areas of the studs to remain

constant during the test, the load ratio at failure can be considered equivalent to a

strength reduction factor of 0.2 giving a corresponding critical temperature of

approximately 5650C from Figure 7-24. Load ratio at the time of failure is equal to the

ratio of yield load at elevated temperature to the yield load at ambient temperature,

which is equivalent to the strength reduction factor (ratio of the yield stress at elevated

temperature to the yield stress at ambient temperature) considering the cross-sectional

area to remain constant during the test (assuming full yielding).

Figure 7-25 shows the development of the maximum hot flange temperature at the

mid-height of the central studs of different wall specimens. The development of these

hot flange temperature profiles has been broken down into linear segments, which

closely approximate the temperature profiles observed in the experiments.

Figure 7-25: Idealized Hot Flange Temperatures of Load Bearing Test

Specimens 2 to 9

Note:

NI-1x1: Test Specimen 2 – Large scale non-insulated load bearing wall specimen

lined on both sides by a single layer of plasterboard.

NI-2x2: Test Specimen 3 – Large scale non-insulated load bearing wall specimen

lined on both sides by two layers of plasterboard.


CI-GF: Test Specimen 4 – Large scale load bearing wall specimen lined on both sides

by two layers of plasterboard with Glass Fibre used as cavity insulation.

CI-RF: Test Specimen 5 – Large scale load bearing wall specimen lined on both sides

by two layers of plasterboard with Rock Fibre used as cavity insulation.

CI-CF: Test Specimen 6 – Large scale load bearing wall specimen lined on both sides

by two layers of plasterboard with Cellulose Fibre used as cavity insulation.

CP-RF: Test Specimen 8 – Large scale load bearing wall specimen lined on both

sides by two layers of plasterboard with Glass Fibre used as external insulation.

CP-CF: Test Specimen 9 – Large scale load bearing wall specimen lined on both

sides by two layers of plasterboard with Cellulose Fibre used as external insulation.

Following equations represent the portion of the temperature-time graph for

temperature values ranging from 1000C to 8000C for the seven test specimens listed

above excluding Test Specimen 8. (‘T’ is the temperature in degree Celsius of the hot

flange at mid-height of the central studs reached in time ‘t’ measured in seconds)

NI-1x1: T = 13.46 t – 75.0 Eq. 1

NI-2x2: T = 8.75 t – 320.0 Eq. 2

CI-GF: T = 12.743 t - 511.5 Eq. 3

CI-RF: T = 10.48 t - 401.1 Eq. 4

CI-CF: T = 10.48 t - 401.1 Eq. 5

CP-RF: T = 6.25 t - 306.2 Eq. 6

CP-CF: T = 8.85 t - 475.7 Eq. 7

From Figure 7-25, approximate failure times for the hot flange of each type of wall

specimen can be obtained using the critical temperature corresponding to the load

ratio. The development of hot flange failure times is shown in Figure 7-26 for a load

ratio of 0.2.

Table 7-3 compares the predicted hot flange failure times with the actual hot flange

failure times. The graph gives the times at which the hot flange of the stud begins to


buckle locally leading to reversal in lateral deformation. The reversal in lateral

deformation subsequently leads to failure of the stud cross-section.

(a) Determination of Critical Temperature for a given Load Ratio or Strength

Reduction Factor

(b) Determination of Stud Failure Times for Various Wall Specimens for a given

Critical Temperature

Figure 7-26: Development of Hot Flange Failure Times for a Given Load Ratio


A close agreement is noticed between the predicted and actual values. The failure

time of the entire stud is slightly higher (approximately 15 minutes) than the hot

flange buckling failure of the individual stud due to the post-buckling strength of the

stud and the redistribution of forces in the studs within the frame.

The failure time of Test Specimen 7 has not been considered as there were difficulties

encountered in maintaining the heating profile in the furnace due to some mechanical

problems. Also in the case of Test Specimens 2 and 3, the end conditions of the studs

at the top were different from those of Test Specimens 4 to 9, and hence the actual hot

flange failure times could not be compared. The actual local buckling of studs in Test

Specimens 7 and 8 was found to be earlier than the predicted values as the studs

towards the end of the test could not expand freely as assumed in the predicted values

due to the closure of the expansion gaps leading to increased thermal strains and

consequently increased load ratio.

Table 7-3: Comparison of Predicted Hot Flange (HF) failure Times of Load Bearing Wall Specimens with Actual Local Buckling of HF (minutes) at a Load

Ratio of 0.2

Test Specimen

No.

Wall Specimen

Predicted Hot Flange

Failure Time (minutes)

Actual Local Buckling of HF (minutes)

indicated by reversal in lateral deformation

of stud

Wall Failure Time

(minutes)

2 NI-1x1 48 - 53

3 NI-2x2 101 - 111

4 CI-GF 85 85 101

5 CI-RF 92 92 107

6 CI-CF 92 96 110

7 CP-GF - - -

8 CP-RF 141 123 136

9 CP-CF 119 110 124


Figure 7-27: Determination of Hot Flange Failure Times using Load Ratio

Figure 7-27 shows a relationship between the load ratio and the HF failure times. The

graph was developed by combining Figures 7-24 and 7-25. The failure time can be

obtained as the corresponding co-ordinate for the required load ratio. For example, a

load ratio of 0.2 yields the HF failure times for the different specimens as displayed in

column 3 of Table 7-3. The intercept on the ‘X’ axis gives the failure times for non-

load bearing walls. The change in gradient below the load ratio of 0.1 in Figure 7-27

is on account of the reduced rate of reduction in the yield strength of steel as seen in

Figure 7-24. For a load ratio of 0.03 the corresponding critical temperature as

obtained from Figure 7-24 is 8000C. The critical temperature for non-load bearing

walls is assumed to lie in the range of 800 to 8500C as the walls although treated as

non-load bearing will still be carrying their own self weight, which comprises of

plasterboards (weighing 13 kg/m2), insulation of varying density and the steel frame

yielding a load ratio between 0.003 and 0.03. Considering 8000C as the critical

temperature the failure times for non-load bearing wall specimens is as obtained by

the ‘X’ intercepts of the graphs in Figure 7-27. Similarly the gradient of the graph in

Figure 7-27 is very small from a load ratio of 1 to 0.9 as the yield strength of steel

hardly changes up to 3000C (refer Figure 7-24). The wall collapses instantly (i.e. at t =

0) at ambient temperature when the load reaches the ultimate load bearing capacity of

wall (i.e. LR = 1).


Table 7-4: Comparison of Predicted Failure Times of Non-load Bearing Wall Specimens with their Actual Failure Times.

Test Specimen

No.

Wall Specimen

Predicted stud failure

time (minutes)

Time taken for HF

temperature to reach 8000C

Wall failure Remarks

4 CI-GF 103 125 125 LB of HF

5 CI-RF 115 130 145 LB of HF

6 CI-CF 115 143 145 LB of HF

7 CP-GF - 185 198 Pb2 fall off

8 CP-RF 177 195 200 Pb2 fall off

9 CP-CF 144 158 163 Pb2 fall off

Table 7-4 gives a comparison of the predicted failure times of non-load bearing wall

specimens with the time taken for the hot flange of the central stud to reach a

temperature of 8000C and the actual failure time of the wall specimen. Test

Specimens 2 and 3 have not been included as conclusive stud failure times could not

be established in the experimental work. For Test Specimens 4, 5 and 6 the local

buckling of the central stud was observed when the temperature of the hot flange was

in the vicinity of 8000C leading to the failure of the wall. The local buckling of the hot

flange is characterized by the reversal in lateral deformation of the central stud. In the

case of Test Specimens 7, 8 and 9, Plasterboard 2 was observed to fall off when the

temperature of the hot flange was in the vicinity of 8000C. This was probably because,

even though there was no observed reversal in lateral deformation of the studs (as the

temperature difference across the stud cross-section was almost uniform and there was

very little lateral buckling) the local deformation of the cross section would easily


lead to the tearing of the weakened plasterboard. Hence in the case of both cavity

insulated specimens and externally insulated specimens the failure of non-load

bearing wall specimens occurred with the temperature of the hot flange reaching a

temperature of approximately 8000C.

The effects of lateral deformation are considered to be very small in the case of

externally insulated test specimens as their lateral deformations are minimal due to a

more or less uniform temperature across the stud cross-section. However, in the case

of cavity insulated test specimens the lateral deformations would result in additional

stresses due to the moments generated by the developing eccentricity along with the

secondary moments due to P-delta effect. Effect of lateral deformation has not been

considered in the development of hot flange failure times.

To account for all these effects ABAQUS finite element program was run using as

input: The temperatures of the hot flange, web and cold flange at mid-height of the

central studs along with their lateral deformations with respect to time when subjected

to the cellulosic fire curve. The change in the modulus of elasticity of steel along with

its yield strength across the depth of the cross-section due to temperature variation

was also accounted. The stud was assumed to be laterally restrained by plasterboards

on either side with a screw spacing of 300 mm along the length of the stud. The end

conditions were assumed to be pinned.

To study the stud failure, the temperature distribution across the mid-height of the

central studs was used as input at intervals of 30 minutes along with the

corresponding lateral deformation. For each input the program was run to yield a

failure load for that particular temperature variation, lateral deformation and time.

Figure 7-28 shows the variation of load ratio with respect to hot flange temperature

obtained by several runs of the program at specified intervals of time (Gunalan 2010).


Figure 7-28: Load Ratio vs Critical Hot Flange Temperatures at Stud Failure

Figures 7-29 to 7-35 show the comparison of the graph as obtained from the program

with the one plotted using the material strength factor alone ignoring the effect of

eccentricities induced by lateral deformation and varying material properties across

the depth of the cross-section. The intersection of zero eccentricity graphs with the

time axis is considered to correspond with the failure of the non-load bearing wall

specimens.

Simple linear equations (see Table 7.5) have been proposed to predict the stud failure

times based upon the experimental results and the graphs drawn using the Abacus

program, and are graphed as dotted lines in Figures 7-29 to 7-35. Figure 7-36 shows

the comparison between the graphical representations of the proposed linear equations

for all the tested wall models, to predict the stud failure times.


Figure 7-29: Load Ratio Vs Stud Failure Times for Test Specimen 2










The Abacus graphs in Figures 7-29 to 7-35 are seen to intersect the zero eccentricity

graphs between a load ratio of 0.5 and 0.4, implying the reducing influence of

eccentricity with the lowering of failure load. Beyond the intersection the gap in the

graph probably denotes the interval of time between the time at reversal in lateral

deformation of the critical stud and the complete failure of the stud. The hot flange

buckling times and the complete stud failure times are compared with experimental

results in Table 7.6

Figure 7-36: Load Ratio Vs Stud Failure Times for all Test Specimens using

Predictive Equations


The failure times for the Test Specimens can be represented by the following

equations:

Table 7-5: Predictive Equations for obtaining Stud Failure Times for Different Wall Models

Test Specimen Predictive Equation Range Eq. No.

NI-1x1: T = -150(LR) + 150

T = -55.5(LR) + 65

0.9 ≤ LR ≤ 1.0

0 ≤ LR ≤ 0.9

Eq. 8a

Eq. 8b

NI-2x2: T = -869.5(LR) + 869.5 0.954 ≤ LR ≤ 1.0 Eq. 9a

T = -92.24(LR) + 128 0 ≤ LR ≤ 0.954 Eq. 9b

CI-GF: T = -403.2(LR) + 403.2 0.876 ≤ LR ≤ 1.0 Eq. 10a

T = -60.5(LR) + 103 0 ≤ LR ≤ 0.876 Eq. 10b

CI-RF: T = -454.54(LR) + 454.54 0.89 ≤ LR ≤ 1.0 Eq. 11a

T = -73(LR) + 115 0 ≤ LR ≤ 0.89 Eq. 11b

CI-CF: T = -449.4(LR) + 449.4 0.9 11≤ LR ≤ 1.0 Eq. 12a

T = -82.32(LR) + 115 0 ≤ LR ≤ 0.911 Eq. 12b

CP-RF: T = -1111(LR) + 1111 0.946 ≤ LR ≤ 1.0 Eq.13a

T = -123.67(LR) + 177 0 ≤ LR ≤ 0.946 Eq.13b

CP-CF: T = -576.9(LR) + 576.9 0.896 ≤ LR ≤ 1 Eq. 14a

T = -93.75(LR) + 144 0 ≤ LR ≤ 0.896 Eq. 14b


Table 7.6 : Table Comparing the Actual Stud Reversal Times and Wall Failure Times with the Predicted Values

Test Specimen

No.

Wall Specimen

Predicted Hot Flange

Failure Time

(minutes)

Reversal in Lateral

Deformation of Stud (minutes)

Predicted Stud

Failure Time

(minutes)

Wall Failure Time

(minutes)

2 NI-1x1 48 - 52 53

3 NI-2x2 101 - 107 111

4 CI-GF 85 85 91 101

5 CI-RF 92 92 101 107

6 CI-CF 92 96 106 110

7 CP-GF - - -

8 CP-RF 141 123 157 136

9 CP-CF 119 110 125 124

The predicted stud failure times correlate well with the actual wall failure times for all

the tested wall models. The reversal in lateral deformations is also predicted very

accurately by the predicted hot flange failure times.

7.3 Essential Points to Consider for Thermal Modeling

The numerous fire tests carried out on plasterboards, composite panels, non-load

bearing walls and load bearing wall specimens has helped in formulating certain

important assumptions or essential factors to be considered in the thermal modeling of

the stud wall systems.

1) The time of exposure to the cellulosic fire curve determines the approximate

depth up to which the free and chemically bound water present in the gypsum

plasterboard gets expelled. On average, 1 minute of fire exposure is required to


expel water from 1 mm thickness of plasterboard. Hence in the case of 13 mm thick

plasterboard exposed to standard time-temperature curve from one side, the

temperature on the ambient surface would be maintained at about 1000C up to 13

minutes and in the case of 16 mm plasterboard it would be maintained for up to 16

minutes.

2) The paper on the exposed face of the plasterboard lasts only for 3 to 4 minutes.

3) After the calcination of the plasterboard, the temperature can be assumed to drop

linearly across the thickness of the plasterboard from the exposed face to the

unexposed face.

4) Interfaces between plasterboards do not influence the linearity of the temperature

variation across the layers of the plasterboards, however, when 16 mm plasterboards

are used, the duration of the second phase in the temperature profile of the ambient

side is seen to extend by approximately 30 minutes per interface.

5) A temperature gradient of 40 degrees per mm thickness can be assumed for the

linear variation across a single layer of plasterboard. However, if two layers are used

the gradient can be assumed to be 26 (degrees/mm) and for three layers a

temperature gradient of 19 (degrees/mm) can be assumed after the complete

calcination of the plasterboards.

6) When three layers of plasterboard are used, the exposed layer (Pb1) should be

assumed as ineffective from 150 minutes onwards in a thermal model.

7) The thermal performance of glass fibre insulated composite panels can be

assumed to remain unchanged regardless of the thickness and density of insulation

used.

8) Regardless of insulation thickness and density, Glass fibre insulation used in

composite panels can be dropped from the thermal model at approximately 90

minutes from the start of the test as at around this time its temperature reaches 700oC

and starts to disintegrate rapidly.

9) Thermal performance of Rock fibre insulations of varying thickness in the

thermal modeling of composite panels can be assumed to be practically same.


10) Plasterboard 1 in the composite panels using rock fibre as insulation should be

removed from the thermal model when the temperature of the interface (Pb1-Ins)

reaches 900oC.

11) In the case of non-load bearing wall specimens with a single layer of plasterboard

on either side of the steel frame the influence of plasterboard joints on the insulation

failure of the wall specimens can be ignored in the thermal modeling as the joints on

the exposed plasterboards have little or no effect on the temperature profiles of the

ambient side plasterboards (see Figure 5-21).

12) The failure of 1x1 NLB wall assemblies is entirely due to the inadequacy in

insulation and not due to loss of integrity or structural stability, hence collapse or

failure of exposed plasterboards need not be considered in the thermal modeling for

such wall assemblies (see Figures 5-16 & 5-17).

13) A difference of less than 250C in the hot flange temperatures of corresponding

end studs in Specimens 1 and 2 (see chapter 5) at the time of failure implies that the

extra heating of the central stud in Specimen 2 due to the joint does not much

influence the heating of the end studs (see Figures 5-18 &5-20), hence the lateral

transmission of heat in the plane of the cavity from central stud to the end studs can

be neglected in thermal modeling.

14) Central studs should be considered for modeling as they show higher

temperatures at any time than the end studs.

15) Effect of plasterboard joints on the lateral deflection of the studs can be ignored

in the thermal models (see Figure 5-23).

16) As joints in plasterboard do not influence the failure of 1x1 non-load bearing wall

specimens, it is reasonable to ignore the effect of such joints in the thermal modeling

of 2x2 wall specimens especially when the joints are staggered.

17) In the case of 2x2 non-load bearing wall specimens without cavity insulation, the

exposed Plasterboards 1 & 2 can be assumed to remain intact and effective until the

end of the test.


18) In the case of 2x2 non-load bearing wall specimens without cavity insulation, the

studs can be assumed to have uniform temperature across the cross-section.

19) The exposed Plasterboards 1 & 2 in the case of 2x2 cavity insulated non-load

bearing specimens need to be removed in the thermal model at a certain stage of the

test. Plasterboard 1 can be assumed to be ineffective when the temperature on the

ambient side of Plasterboard 1 (Pb1-Pb2) crosses 9000C.

20) The formation of the second plateau in the time-temperature profile of the

plasterboards need not be assumed in the thermal modeling, as it is not evident in the

temperature profiles of the plasterboards.

21) In the case of 2x2 non-load bearing wall specimens using Glass Fibre or

Cellulose Fibre as cavity insulation, the insulation can be removed from the model

when the temperature of the Pb2-Ins interface crosses 7000C.

22) In the case of 2x2 cavity insulated non-load bearing wall specimens the specimen

can be considered to have failed structurally when the temperature of the Pb2-Ins

interface reaches 7000C. At this temperature the hot flange becomes sufficiently soft

to initiate a reversal in lateral deformation of the studs. Also the screws holding the

plasterboards to the steel frame begin to rotate downwards under the plasterboards

weight as the soft hot flange is unable to offer any degree of fixity. This leads to the

collapse of the plasterboard exposing the entire frame to direct fire.

Thus when the temperature of the interface Pb2-Ins is in the range of 7000C to

7500C, four things happen very quickly;

a) Cavity insulation (Glass Fibre or Cellulose Fibre) starts to disintegrate rapidly.

b) Studs undergo reversal in lateral deformation.

c) The screws connecting the Plasterboards to the steel frame start rotating

downwards as they lose their fixity.

d) Plasterboard 2 falls off as the sudden jerk introduced by the reversal in lateral

deformation of the studs coupled with the rotation of the screws used for fixing


the plasterboard to the frame are sufficient to induce a partial or complete

collapse in the already weakened exposed plasterboard.

Only Rock Fibre when used as cavity insulation resists burnout well beyond 7000C.

This delays the reversal in lateral deformation of the studs and allows Plasterboard 2

to remain in position till the Pb2-Ins interface temperature reaches 9000C leading to

the collapse of Pb2 and reversal in lateral deformation of the studs resulting in the

failure of the wall.

23) 2x2 non-load bearing wall specimens (both cavity insulated and externally

insulated) fail by stud buckling before insulation failure can occur.

24) Plasterboard 1, Insulation between the exposed plasterboards and Plasterboard 2

from the externally insulated wall models need to be removed after certain time in

the thermal modeling. The critical temperatures identifying the removal for these

elements are identical with the cavity insulated specimens. The insulation if it is

Glass Fibre or Cellulose Fibre can be removed from the model when the Pb1-Ins

interface temperature reaches 7000C and Plasterboard 1 can be removed from the

model regardless of the type of insulation used when the interface temperature

reaches 9000C. When Rock Fibre is used as external insulation it can be assumed to

become ineffective when the interface temperature reaches 9000C.

Similar to the cavity insulated specimens, Plasterboard 2 can be removed from the

model when the temperature of the cavity facing surface of the plasterboard (Pb2-

Cav) is in the range of 7000C – 7500C. The wall can be assumed to have failed by

structural inadequacy on the removal of Plasterboard 2.

25) Vertical plasterboard joints along the stud length affect the thermal performance

of 1x1 LBW and leads to a structural failure.

26) 2x2 load-bearing walls fail by structural inadequacy and not by insulation or

integrity failure.

27) The type of insulation used in the cavity of load-bearing walls has a low influence

on the stud temperatures.



28) Hot flange temperatures of externally insulated load-bearing walls is influenced

by the type of insulation used.

7.4: Conclusion:

In the comparative study of different wall models, the fire performance of externally

insulated wall specimens both load bearing and non-load bearing was found to be

considerably better than the non-insulated and cavity insulated wall specimens. The

thermal insulation property of the externally insulated wall specimens was seen to be

much superior when compared with the insulation characteristics of cavity insulated

specimens. The ambient side temperatures and the lateral deformations of the

externally insulated specimens were also seen to be consistently lower than what was

observed in the case of cavity insulated specimens.

In the case of externally insulated wall specimens, the quality of insulation used was

observed to directly influence the fire performance of the specimen with the Rock

fibre giving the best results, whereas, in the case of cavity insulated specimens, the

type of insulation used did not much affect the fire performance of the wall models.

The failure of all the wall specimens was noted to occur primarily due to the structural

failure of the studs and never by insulation or integrity. Cavity insulated specimens

were seen to fail earlier than similarly built non-insulated specimens, whereas, the

failure times of the externally insulated Test Specimens were seen to be maximum.

Simple linear equations and graphs based upon experimental work have been

proposed to predict the growth in hot flange temperatures of load bearing non-

insulated, cavity insulated and externally insulated wall specimens. This temperature

growth model is used to develop predictive equations to estimate the failure times of

all the tested wall models. A close agreement has been observed between the

predicted failure times and the actual failure times of load bearing and non-load

bearing wall models with and without insulation.

Chapter 8: Conclusions and Recommendations

This thesis has described a detailed investigation into the structural and thermal

performance of cold-formed LSF stud wall systems lined with gypsum plasterboards

under fire conditions. It included both the conventional steel stud wall systems with

and without cavity insulation as well as a new steel stud wall system based on a

composite panel in which a layer of external insulation was used between the two

plasterboards. Both non-load bearing walls and load bearing walls were tested in a

detailed experimental study. This research has thus developed comprehensive

experimental thermal and structural performance data for both the conventional and

the new non-load bearing and load bearing cold-formed steel stud wall systems under

fire conditions including simple and accurate methods to predict their fire resistance

rating. It has improved the knowledge and understanding of the fire performance of

cold-formed LSF stud wall systems under fire conditions, and has led to the

development of safer design methods for fire conditions and new LSF stud wall

systems with increased fire rating.

A detailed literature review of the current body of knowledge in this field was

undertaken first (Chapter 2). High grade cold-formed steels are increasingly used not

only in LSF stud wall systems, but also in other building systems. The lack of reliable

mechanical property data of these high grade cold-formed steels at elevated

temperatures was addressed in Chapter 3, which describes the steady state tensile

coupon tests undertaken in this research at ambient and elevated temperatures. The

use of non-contact Laser Speckle Extensometer (LSE) was found to be highly

successful in providing the required strain measurements at elevated temperatures

instead of the resistance type strain gauges. This experimental study led to the

development of predictive equations for the determination of yield strength and elastic

modulus of high strength steels at elevated temperatures. The developed equations

compared well with the test results, and are considered to eliminate the conservative

predictions given by the current Australian and European standards.

In Chapter 4 the thermal performance of the gypsum plasterboards was investigated

using 15 small scale fire tests of Type X gypsum plasterboards supplied by Boral

Plasterboards under the product name FireSTOP. Thermal performance of single,


double and triple layers of plasterboards was investigated in detail. It was found that

the discontinuity or interface between the layers of plasterboard increased the fire

performance of the wall system. Thermal performance of composite panels using

different types of insulating material of varying thickness and density was also

investigated, which allowed the assessment of the fire performance of insulations such

as glass fibre, rock fibre and cellulose fibre and also the determination of the

temperature at which the fall off of external plasterboards occurred.

Chapter 5 presents the details of nine small scale wall models built and fire tested to

investigate the thermal performance of conventional steel stud wall systems with and

without the use of cavity insulation and the innovative steel stud wall systems using

composite panels. The composite panels were seen to offer greater thermal protection

to the studs as compared to the conventionally built non-load bearing wall models.

The use of cavity insulation regardless of the type and density of insulation has been

shown to lower the fire rating of the walls. Rock fibre was identified to have the

maximum detrimental effect on the fire performance of non-load bearing walls when

used as cavity insulation. This chapter identifies and discusses the deficiencies in the

conventional stud wall systems. Time-temperature measurements from the tests

clearly demonstrated the superior thermal and fire performance achieved by the use of

composite panels. The benefits of adopting this new system over the conventional

stud wall systems are discussed in this chapter.

Chapter 6 presents the details of nine full scale load bearing wall models built and fire

tested to study the thermal and structural performance of the load bearing wall

assemblies lined with single or dual layers of plasterboards with and without cavity

insulation and compares the results with the thermal and structural performance of

load bearing wall models built using composite panels on either side of the steel

frame. Details of the results, including the temperature and deflection profiles

measured during the tests are presented along with the stud failure modes. The

analysis showed that the proposed cold-formed steel stud wall systems with external

insulation provided considerably increased fire resistance rating with smaller lateral

deformations than the conventional cavity insulated stud wall systems.

Chapter 7 presents the outcomes of the tests performed on non-load bearing and load

bearing conventional steel stud wall systems with and without cavity insulation and


the innovative steel stud wall systems using composite panels. Idealized hot flange

temperature profiles of non-insulated, cavity insulated and externally insulated load

bearing wall models using composite panels have been developed and presented in

this chapter along with suitable equations for predicting their failure times (fire

resistance rating). The chapter also presents the development of a simple graphical

method to predict the failure times of non-load bearing and load bearing wall models

under different load ratios.

8.1. Main Research Outcomes

The most valuable outcomes from this research are as follows:

Significantly improved the knowledge and understanding of the structural and

thermal performance of cold-formed LSF stud wall systems under fire

conditions. Both non-load bearing and load bearing walls with varying

arrangements of plasterboard and insulation were included.

Developed an innovative cold-formed LSF wall system with increased fire

resistance rating through the use of a composite panel system in which a layer

of insulation is placed externally between the two plasterboards. This thesis is

the first one to propose and investigate the use of such an innovative

composite panel system, and demonstrate its superiority and benefits over

conventional panels.

Developed comprehensive structural and thermal performance data for both

the conventional and the new LSF stud wall systems, which can be used for

accurate numerical modelling and design of LSF stud walls by fire researchers

and designers.

Developed simple predictive models for the mechanical properties of high

grade steels at elevated temperatures for use by researchers and engineers as

the values given in the current steel design standards are either too

conservative or unsafe.

Developed idealised time-temperature profiles of studs in LSF walls under fire

conditions and their fire resistance rating as a function of varying

arrangements of plasterboards and insulation, and load ratios. Engineers,


designers and researchers in this field can use them without the need for

further expensive and time consuming full scale fire tests.

This research has produced many useful outcomes for the designers of LSF

stud wall systems; for example, this research has shown that the common

industry belief that the use of cavity insulation improves the fire rating of

walls is not true. The use of cavity insulation was found to reduce the fire

rating of load bearing walls regardless of the type and density of insulation.

This research has paved the way for Australian building industries to develop

new LSF stud wall systems based on the new composite panels proposed in

this research with increased fire rating for commercial applications worldwide.

Developed an excellent fire testing facility at the Queensland University of

Technology that has the capacity to simulate both standard and real fire curves

on LSF stud walls and to obtain high quality temperature and deformation data

using the latest technologies. This is currently being used by other researchers.

8.2 Recommendations to the Construction Industry

Based on the research reported in this thesis, the following recommendations are

made to the building construction industry.

Implement the use of the new composite panel proposed in this research in the

standard wall panel systems. For this purpose develop improved cost-efficient

methods of building and installing composite panels.

Develop thinner sheets of insulation with higher fire performance and thermal

insulation characteristics to facilitate easy construction of composite panels.

Develop appropriately sized composite panel units with interlocking joints to

promote speedy construction of the new wall systems with minimum labour.

Develop improved methods for fixing multiple plasterboards and composite

panels to the steel studs of the walls in order to improve the fire performance

of wall systems with reduced thermal bridging problems.



The technique of spraying cellulose fibre insulation onto the plasterboard

needs to be standardized to achieve a uniform distribution and density of the

sprayed insulation layer. This will avoid the formation of weak areas within

the composite panel that leads to a weaker fire performance of the wall

systems.

8.3 Future Research

This research has addressed the structural and thermal performance of cold-formed

LSF stud wall systems under fire conditions using an extensive experimental study

program. However, further research is needed in some areas as shown next.

Effect of varying stud sizes on the fire performance of load bearing wall

systems built using composite panels.

Extension of external insulation concept to ceiling elements in order to

improve their fire performance.

Numerical modeling of both conventional and proposed LSF stud wall

systems to simulate both their thermal and structural behaviour. Such

numerical thermal and structural models can be validated using the vast

amount of experimental results presented in this thesis.

Additional experimental investigations on the fire performance of both

conventional and externally insulated load bearing wall specimens at different

load ratios to verify the predictive equations presented in the thesis.

Experimental investigations on the fire performance of both conventional and

externally insulated load bearing wall specimens using different types of

popularly used plasterboards used in the construction industry.

Effect of screw length and screw spacing on the fire performance of stud walls

using multiple plasterboards. This is because the screws used in the test

specimens with multiple plasterboards on the fire side or composite panels

were found to be severely bent at elevated temperatures, promoting the

collapse of the outermost external plasterboards

References

Ala-Outinen and Myllymaki, J. (1995) The Local Buckling of RHS Members at Elevated

Temperatures, Technical Research Centre of Finland, Espoo.

Alfawakhiri, F. (2001), Behaviour of Cold-formed-Steel-framed Walls and Floors in

Standard Fire Resistance Tests, PhD, Thesis, Department of Civil and Environmental

Engineering, Carleton University, Ottawa, Ontario, Canada.

American Society for Testing and Materials (ASTM E119, 1995), Standard Test Methods

for Fire Tests of Building Construction and Materials, West Conshohocken, PA, USA.

Andersson, L. and Jansson, B. (1987), Analytical Fire Design With Gypsum – A

Theoretical and Experimental Study, Institute of Fire Safety Design, Lund, Sweden.

Baux, C. Melinge, Y. Lanos, C. And Jauberthie, R. (2008) Enhanced Gypsum Panels for

Fire Protection, Journal of Materials in Civil Engineering, Vol. 20, pp. 71-77.

Benichou, N. and Sultan, M.A. (2005), Thermal Properties of Light Weight Framed

Construction Components at Elevated Temperatures, Fire and Materials, Vol. 29, pp.

165-179.

British Standards Institute-(BSI) (1990), British Standard, BS 5950, Structural Use of

Steelwork in Building, Part 8, Code of Practice for Fire Resistance Design, London, UK.

Buchanan, A. H. (2001), Structural Design for Fire Safety, John Wiley and Sons, HK.

The Steel Construction Institute (1993), Building Design Using Cold Formed Steel

Sections: Fire Protection, SCI Publication P129, Berkshire, UK.

CAN/ULC-S101-M89, Standard Methods of Fire Endurance Tests of Building

Construction and Materials, Underwriters Laboratories of Canada, Scarborough, Ontario,

1989.


Chen, J. and Young, B. (2004), Mechanical Properties of Cold-formed Steel at Elevated

Temperatures, Proceedings of 17th International Speciality Conference on Cold-formed

Steel Structures, University of Missouri-Rolla, USA, pp. 437-466.

Chen, J. and Young, B. (2007), Experimental Investigation of Cold-formed Steel Material

at Elevated Temperatures, Thin-walled Structures, Vol. 45, pp. 96-110.

Cooke, G. M. E. (1987),Thermal Bowing and How it Affects the Design of Fire

Separating Construction, Fire Research Station, Building Research Establishment, Herts,

UK.

Davies, J. M. (2000), Recent Research Advances in Cold-formed Steel Structures,

Journal of Constructional Steel Research, Vol. 55, No. 1-3, pp. 267-288.

EN 520 (2004) Gypsum Plasterboards, Definitions, Requirements and Test Methods,

CEN, Brussels.

EN 15283-2 (2008) Gypsum Boards with Fibrous Reinforcement – Definations,

Requirements and Test Methods - part 2: Gypsum fibre boards, CEN, Brussels.

European Prestandard ENV 1993-1-2, Eurocode 3 – Design of Steel Structures – Part 1 –

2: General Rules – Structural Fire Design, European Committee for Standardization (CE

N), Brussels, Belgium, September 1995.

Feng, M. Wang, Y. C. and Davies, J. M., (2003), Thermal Performance of Cold-formed

Thin-walled Steel Panel Systems in Fire, Fire Safety Journal, Vol. 38, pp. 365-394.

Feng, M. Wang, Y. C. (2005), An Experimental Study of Loaded Full-Scale Cold-

Formed Thin-Walled Steel Structural Panels Under Fire Conditions, Fire Safety Journal,

Vol. 40, pp. 43-63.

Franssen, J. M. (1999), Thermal Properties of Gypsum Board Walls Submitted to the

Fire, Literature Survey, UNIVERSITE DE LIEGE, Belgium.


Frangi, A. Schleifer, V. Fontana, M. and Hugi, E. (2010), Experimental and Numerical

Analysis of Gypsum Plasterboards in Fire, Fire Technology, Vol. 46, pp. 149-167.

Gerlich, J.T. (1995), Design of Load-bearing Light Steel Frame Walls for Fire

Resistance, Fire Engineering Research Report 95/3. School of Engineering, University of

Canterbury, New Zealand.

Gerlich, J.T. Collier, P. C. R. and Buchanan, A. H. (1996), Design of Steel-framed Walls

for Fire Resistance, Fire and Materials, Vol. 20, No. 2, pp. 79-96.

Ghazi, W. and Hugi, E. (2009), Four Types of Gypsum Plaster Boards and their

Thermophysical Properties under Fire Condition, Journal of Fire Sciences, Vol. 27, No.1,

pp. 27-43.

Gonclaves, T. Clancy, P. and Poynter, W. (1996), Mechanical Properties of Fire Rated

Gypsum Board, Victoria University of Technology, Australia.

Harmathy, T. Z. (1998), Properties of Building Material, The SFPE Handbook of Fire

Protection Engineering, Society of Fire Protection Engineering/National Fire Protection

Association, Bostan, United States.

ISO 834 -1, (1999), Fire Resistance Tests – Elements of Building Construction – Part 1:

General Requirements, International Organization for Standardization, Geneve,

Switzerland.

Klippstein, K.H. (1978), Strength of Cold-formed Steel Studs Exposed to Fire,

Proceedings of the 4th International Speciality Conference, Missouri-Rolla, St.Louis,

Rolla, Mo, USA, pp. 513-555.

Kodur, V. K. R. Sultan, M. A. Latour, J. C. Leroux, P. and Monette, R. C. (1999),

Experimental Studies on the Fire Resistance of Load-Bearing Steel Stud Walls, Internal

Report, IRC, NRC, Ottawa, Canada.


Kwon, I. K. Choi, K. and Jee, N. Y. (1998), Fire Resistance of Bearing Wall Using Steel

and Gypsum, Proceedings of the Fourteenth International Speciality Conference on Cold

Formed Steel Structures, St. Louis, MO, pp. 379-391.

Lee, J. Mahendran, M. and Mäkeläinen, P. (2003), Prediction of Mechanical Properties of

Light Gauge Steels at Elevated Temperatures, Journal of Construction Steel Research, pp.

1517-1532.

Lie, T.T. (Ed.), (1992), Structural Fire Protection, American Society of Civil Engineers,

New York, NY.

Mäkeläinen, P. and Miller, K. (1983), Mechanical Properties of Cold-formed Galvanized

Sheet Steel Z32 at Elevated Temperatures, Helsinki University of Technology, Finland.

Manzello, S.L. Gann, R.G. Kukuch, S.R. Prasad, K. And Jones, W.W. (2005), Real Fire

Performance of Partition Assemblies, National Institute of Standards and Technology

(NIST), USA.

Manzello, S.L. Gann, R.G. Kukuch, S.R. Lenhert, D.B. (2006), Influence of Gypsum

Board Type (X or C) on Real Fire Performance of Partition Assemblies, National

Institute of Standards and Technology (NIST), USA.

Mehaffey, J.R. Cuerrier, P. and Carisse, G. (1994), A Model for Predicting Heat Transfer

through Gypsum-Board/Wood-Stud Walls Exposed to Fire, Fire and Materials, Vol. 18,

No. 5, pp 297-305.

Milke, J. A. (1999), Analytical Methods to Evaluate Fire Resistance of Structural

Members, ASCE Journal of Structural Engineering, Vol. 125, No. 10, pp. 1179-1187.

Noureddine, B. and Sultan, M. A. (2005), Thermal Properties of Lightweight-framed

Construction Components at Elevated Temperatures, Fire and Materials, Vol. 29, pp.

165-179.


Outinen, J. Kaitila, O. Makelainen, P. (2000), A Study for the Development of the

Design of Steel Structures in Fire Conditions, Proceedings of 1st International Workshop

of Structures in Fire, Copenhagen, Denmark.

Purkis, J. A. (1996), Fire Safety Engineering Design of Structures, Reed Educational and

Professional Publishing Ltd. Oxford, UK.

Ranby, A. (1999), Structural Fire Design of Thin Walled Steel Sections, Licentiate

Thesis, Department of Civil and Mining Engineering, Lulea University of Technology,

Stockholm, Sweden.

Ranawaka, T. and Mahendran, M. (2009), Experimental Study of the Mechanical

Properties of Light Guage Cold-formed Steels at Elevated Temperatures, Fire Safety

Journal 44, 219-229.

SCI Publication P 129 (1993), Building Design using Cold-formed Steel Sections: Fire

Protection.

Sidey, M. P. and Teague, D. B. (1988), Elevated Temperature Data for Structural Grades

of Galvanized Steels, Technical Note No. WL/PM/TN/251/88/D, Welsh Laboratory,

British Steel Corporation, Port Talbot, UK.

Son, B. C. and Shoub, H. (1973), Fire Endurance Tests of Double Module Walls of

Gypsum Board and Steel Studs, NBSIR 73-173, Centre for Building Technology,

National Bureau of Standards, Washington, DC.

Standards Australia (SA) (1998), AS 4100, Steel structures, Sydney, Australia.

Standards Australia (SA) (2005), AS 4600, Cold-formed steel structures, Sydney,

Australia.

Standards Australia (SA) (2005), AS 1530.4, Methods for fire tests on building materials,

components and structures, Part 4: Fire-resistance tests of elements of building

construction, Sydney, Australia.


Standards Australia (SA) (2005), AS 1391, Metallic materials-Tensile testing at ambient

temperature, Sydney, Australia.

Standards Australia (SA) (1979), AS 2291, Methods For the Tensile Testing of Metals at

Elevated Temperatures, Sydney, Australia.

Sterner, E. and Wickstrom, U. (1990), TASEF- Temperature Analysis of Structures

Exposed to Fire, Fire Technology SP Report, Swedish National Testing Institute, Boras,

Sweden.

Sultan, M. A. (1995), Effect of Insulation in the Wall Cavity on the Fire Resistance

Rating of Full-Scale Asymmetrical (1 x 2) Gypsum Board Protected Wall Assemblies,

Proceedings of the International Conference on Fire Research and Engineering, Orlando,

FL, Lund D. P. (Ed.), Society of Fire Protection Engineers, Bostan, MA, pp. 545-550.

Sultan, M.A. (1996), A Model for Predicting Heat through Non-insulated Unloaded

Steel-Stud Gypsum Board Wall Assemblies Exposed to Fire, Fire Technology, Vol. 32,

No. 3, pp. 239-259.

Sultan, M.A. (2010), Comparison of Gypsum Board Fall-off in Wall and Floor

Assemblies, 12th International Conference on Fire Science and Engineering Conference,

Nottingham, UK, pp. 1-6.

Sultan, M.A. Alfawakhiri, F. Benichou, N. (2001), A Model for Predicting Heat Transfer

through Insulated Steel-Stud Wall Assemblies Exposed to Fire, Fire and Materials,

International Conference, San Francisco, pp. 495-506.

Sultan, M.A. and Kodur, V. R. (2000), Light-Weight Frame Wall Assemblies: Parameters

for Consideration in Fire Resistance Performance-Based Design, Fire Technology,

Vol.36, No. 2, pp. 75-88.

Sultan, M.A. and Lougheed, G. D. (1994), The Effect of Insulation on the Fire Resistance

of Small-Scale Gypsum Board Wall Assemblies, Proceedings of the Fire and Materials

Third International Conference and Exhibition, Washington, DC, Interscience

Communications Limited, London, UK, pp. 11-20.



Thomas, G. (2002), Thermal Properties of Gypsum Plasterboard at High Temperatures,

Fire and Materials, Vol. 26, pp. 37-45.

Documents

Structural and Thermal Performance of Cold-formed Steel ...eprints.qut.edu.au/46654/1/Prakash_Kolarkar_Thesis.pdf · Structural and Thermal Performance of Cold ... Commonly Used Cold-formed