Federal Building and Fire Safety Investigation of the World Trade Center Disaster

Building and Fire Research LaboratoryNational Institute of Standards and Technology

U.S. Department of Commerce

Condition of Thermal Insulation:Methodology

June 23, 2003

Nick Carino John Gross

Kuldeep PrasadFahim Sadek

Monica StarnesJiann Yang

DisclaimerDisclaimer

Certain commercial entities, equipment, products, or materials are identified in this presentation in order to describe a procedure or concept adequately or to trace the history of the procedures and practices used. Such identification is not intended to imply recommendation, endorsement, or implication that the entities, products, materials, or equipment are necessarily the best available for the purpose.

ThermalThermal--Structural ModelingStructural Modeling

• Analysis of undamaged buildings exposed to postulated fires

• Analysis of damaged buildings exposed to 9/11 fires• Condition of fireproofing needs to be characterized

OverviewOverview

• Sensitivity study• In-place conditions

History of sprayed fire resistive material (SFRM)MeasurementsStatistical analysis

• Equivalent thickness• Thermal properties• Response to impact• Summary

Sensitivity StudySensitivity Study

• Effects of thickness variability and “gaps” in SFRM• Simplified finite-element analysis

1 in. thick steel plate, 60 in. longAverage thickness: 0 to 2.0 in.Standard deviation: 0 to 1.0 in.Gap length: 0 to 30 in.

• Exposure: 1100 ºC fire• Compute temperature history of steel

ModelModel

2 in.

2 in.

60 in.

1 in. thick steel plate Fireproofing

1 2 3 4 5

ModelModel

• Use psuedo-random number generator to select SFRM thickness at cross-section

Based on average thickness and standard deviation

SFRM

High conductivity material

ExampleExample

Uniform Thickness Standard Deviation = 1 in.

0.5 in. 1.75 in.

Effect of Gap in SFRMEffect of Gap in SFRM

1 2 3 4 5

With Gap

No GapTime = 50 min

Sensitivity StudySensitivity Study

• Variability in thickness of SFRM reduces its effectiveness

• Gap in SFRM:Local heatingPath for heat flow into member

History of SFRM in WTC TowersHistory of SFRM in WTC Towers

• 1969: Decision to use ½ in. CAFCO® BLAZESHIELD ® Type D• 1970: CAFCO® BLAZESHIELD® Type D discontinued at 38th floor,

replaced with CAFCO® BLAZESHIELD® Type D/CF• 1994: Thickness measurements on floor trusses on 23rd and 24th

floors• 1995: PA initiated study of SFRM thickness during tenant alterations• 1999: PA established guidelines for SFRM replacement and repair• Late 90s: SFRM upgraded to 1 ½ in. with CAFCO® BLAZESHEILD®

Type II

Specified ThicknessSpecified Thickness

• October 30, 1969 PA CorrespondenceColumns < 14WF228: 2 3/16 in.*Columns ≥ 14WF228: 1 3/16 in.Beams, spandrels, floor trusses: ½ in.

• Alcoa Drawings (Note 11)3 h on spandrels ½ in. 1/2 in. V.A.**4 h on columns (heavy) 1 3/16 in. 7/8 in. V.A.

• 1995 Upgrade StudyFloor trusses 1 ½ in.

*CAFCO® BLAZESHIELD® Type D**V.A. = Vermiculite aggregate plaster

Thickness MeasurementsThickness Measurements

• 1994 measurements on 23rd and 24th floors of WTC 1• Analysis of photographs taken in the 1990s• Thickness measurements in PA Construction Audit

Reports from late 1990s• 1999 measurements of beams and columns in WTC 1

shaft 14/15

Floors 23 and 24 Trusses (1994)Floors 23 and 24 Trusses (1994)

• 16 randomly selected trusses per floor• 6 replicate measurements on “flanges and web”

0

0.2

0.4

0.6

0.8

1

1.2

22 23 24 25

Aver

age

Thic

knes

s, in

.

Floor

Grand Average = 0.74 in.St. Dev. of Avg. = 0.16 in.

Normal Probability Plots Normal Probability Plots

• Thickness appears better described by log-normal distribution

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

.01 .1 1 5 10 2030 50 7080 90 95 99 99.9 99.99

y = 0.74187 + 0.1607norm(x) R= 0.97209

Ave

rage

Thi

ckne

ss, i

n.

Percent

(c)

-0.8

-0.6

-0.4

-0.2

0

0.2

.01 .1 1 5 10 2030 50 7080 90 95 99 99.9 99.99

y = -0.32122 + 0.21269norm(x) R= 0.98754

Ln (T

hick

ness

)

Percent

(e)

Thickness Natural Logarithm of Thickness

Lognormal DistributionLognormal Distribution

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 0.5 1 1.5 2 2.5

Freq

uenc

y

Thickness, in.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

-2 -1.5 -1 -0.5 0 0.5 1

Freq

uenc

y

Ln (Thickness)

Analysis of PhotographsAnalysis of Photographs

• Original SFRM: WTC 1 (22, 23, 27) in mid 1990s• Upgraded SFRM, WTC 1 (below 31) in 1998

Bar Radius + Fireproofing Thickness

Center Line

Reference Length

Results of Photo AnalysisResults of Photo Analysis

0

0.5

1

1.5

2

2.5

.01 .1 1 5 10 2030 50 7080 90 95 99 99.9 99.99

MainBridgingStrutMain-Upgrade

Thic

knes

s, in

.

Percent

-3

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

.01 .1 1 5 10 2030 50 7080 90 95 99 99.9 99.99

Ln MainLn BridgingLn StrutLn Main-Upgrade

Ln T

hick

ness

Percent

• Original SFRM thickness distribution appears to be lognormal

• Upgraded SFRM thickness distribution appears to be normal

Summary of Photo AnalysisSummary of Photo Analysis

0.50.2 in.0.4 in.Original

Strut

0.20.4 in.1.7 in.Upgraded Main

0.60.25 in.0.4 in.Original Bridging

0.50.3 in.0.6 in.Original

Main

Coeff. of VariationStd. Dev.Average

ThicknessFloor Truss

Construction Audit ReportsConstruction Audit Reports

• 18 data sets for WTC 1 (93, 95, 98, 99 and 100)• 14 data sets for WTC 2 (77, 78, 88, 89, 92)

1

1.5

2

2.5

3

3.5

4

4.5

5

0 1 2 3

Thic

knes

s, in

.

Tower

Combined MeasurementsCombined Measurements

• Lognormal distribution appears more appropriate• Overall average = 2.5 in.• Overall standard deviation = 0.6 in.

0.4

0.6

0.8

1

1.2

1.4

1.6

.01 .1 1 5 10 2030 50 7080 90 95 99 99.9 99.99

y = 0.90833 + 0.2266norm(x) R= 0.99261

Ln (T

hick

ness

)

Percent

1

1.5

2

2.5

3

3.5

4

4.5

5

.01 .1 1 5 10 2030 50 7080 90 95 99 99.9 99.99

y = 2.5449 + 0.57398norm(x) R= 0.98583

Thic

knes

s, in

.

Percent

Core Beams and ColumnsCore Beams and Columns

• April 1999 measurements in shaft 14/15 of WTC 1 (1st to 45th floor)

Average = 1.0 in.St. Dev. = 0.2 in.

Average = 0.8 in.St. Dev. = 0.2 in.

0

0.5

1

1.5

2

0 10 20 30 40 50

IndividualAverage

Thic

knes

s, in

.

Floor

Beams

0.75 in.

0.0

0.5

1.0

1.5

2.0

0 10 20 30 40 50

IndividualAverage

Thic

knes

s, in

.

Floor

Columns

0.5 in.

0.20.25

0.2 in.0.2 in.

1.0 in.0.8 in.

Core BeamsCore Columns

PA Measurements

0.20.6 in.2.5 inTruss UpgradedPA Measurements

0.50.60.50.2

0.3 in.0.25 in.0.2 in.0.4 in

0.6 in.0.4 in.0.4 in.1.7 in.

Main TrussBrdg. TrussStrutMain Upgraded

Photos

0.2*0.2 in.*0.7 in.TrussPA Measurements

Coeff. of Variation

St. Dev.AverageElementData Source

Summary of SFRM ThicknessSummary of SFRM Thickness

*Variability of averages

Equivalent ThicknessEquivalent Thickness

• Variability of thickness reduces effectiveness of SFRM• Not practical to include variable thickness in thermal

modeling• Establish “equivalent uniform thickness” that provides

thermal protection equivalent to variable thickness• Approach

Bar model with variable thickness SFRMCompute elongation under 12,500 psi tensile stress as a function of timeCompare with elongation obtained with uniform thickness SFRM

Cases Considered (Floor Truss)Cases Considered (Floor Truss)

• Lognormal distribution• Original SFRM: tavg = 0.75 in., st. dev. = 0.3 in.• Upgraded SFRM: tavg = 2.5 in., st. dev. = 0.6 in.• 3 sets of psuedo-random numbers• 100 elements for 60 in. bar• 5-point smoothing

55--point Smoothingpoint Smoothing

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

0 10 20 30 40 50 60

Rough TextureSmooth Texture

Thic

knes

s, in

.

Bar Position, in.

Results for Original SFRMResults for Original SFRM

0.00

0.20

0.40

0.60

0.80

0 5 10 15 20 25 30

Rough 1Rough 2Rough 3Smooth 1Smooth 2Smooth 30.4" Uniform FP0.5" Uniform FP0.6" Uniform FP

Def

orm

atio

n (in

)

Time (min)

Load = 9800 lbs

Results for Upgraded SFRMResults for Upgraded SFRM

0.00

0.20

0.40

0.60

0.80

1.00

0 20 40 60 80 100

Case 1Case 2Case 32.0 in. Uniform2.2 in. Uniform

Def

orm

atio

n, in

.

Time, min

Fireproofing Thickness in Thermal ModelingFireproofing Thickness in Thermal Modeling

• Floor trusses (Original):Main: T = 0.6 in.Bridging (Two-way): T = 0.6 in.Bridging (One-way) and struts: T = 0.3 in.Saddle and damper: T = 0 in.

• Floor trusses (Upgrade):T = 2.2 in., except dampers T = 0 in.

• Other elements: Specified thicknessAverage tends to exceed specified thicknessVariability reduces effectiveness

One-way

One-way

Two-way

Core

Thermal and Physical Properties of SFRMThermal and Physical Properties of SFRM

• MaterialsCAFCO® BLAZESHIELD® Type DC/F (Original)CAFCO® BLAZESHIELD® Type II (Upgrade)Monokote® MK-5® (WTC 7)

• Properties as function of temperatureThermal conductivitySpecific heat capacityDensityCoefficient of thermal expansion

Example ResultsExample Results

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0 200 400 600 800 1000 1200 1400

Monokote MK-5BLAZE-SHIELD IIBLAZE-SHIELD DC/FTh

erm

al C

ondu

ctiv

ity (W

/mK)

Temperature, oC

600

800

1000

1200

1400

1600

0 200 400 600 800 1000 1200 1400

Monokote MK-5BLAZE-SHIELD IIBLAZE-SHIELD DC/FS

peci

fic H

eat,

J/(k

g. K)

Temperature, oC

Thermal Conductivity Specific Heat Capacity

Impact DamageImpact Damage

• SFRM was dislodgedDebris fieldLocalized accelerations and deformations

• Estimate extent of dislodged SFRMMeasure static adhesive and cohesive tensile strengthDevelop “failure criteria”Impact analysis and engineering judgment to estimate extent of dislodged SFRM

Tensile PullTensile Pull--off Testoff Test

SFRM

Aluminum Plate

Adhesive

¼ in. Steel Plate

Simple ModelsSimple Models

afb

fbρ t

a =

t

Planar Substrate

ft

fb

a

dodi

4 ft (do + (α – 1) di

(do2 – di

2) ρπa =

α = ftfb

Encased Substrate

Impact TestsImpact Tests

• Determine acceleration to dislodge SFRM

Accelerometer

Accelerometer

Plate Encased Bar

SummarySummary

• Methodology for assessing condition of SFRM has been reviewed

• Variability of SFRM thickness is taken into account by use of “equivalent thickness”

• Available data used to obtain rational thickness valuesFloor trusses: original T = 0.6 in.Floor trusses: upgraded T = 2.2 in.Others: Specified thickness

• Temperature dependence of thermal properties established

• Mechanical damage to be estimated on the basis of tests and analysis