The Cyclic Behaviour of Reinforced Concrete Chimney Sections with … · Advances in Structural Engineering Vol. 12 No. 3 2009 411 The Cyclic Behaviour of Reinforced Concrete Chimney

Advances in Structural Engineering Vol. 12 No. 3 2009 411

The Cyclic Behaviour of Reinforced Concrete Chimney

Sections with and without Openings

John L. Wilson*

Faculty of Engineering and Industrial Sciences, Swinburne University of Technology, Australia

(Received: 8 January 2008; Received revised form: 20 August 2008; Accepted: 9 December 2008)

Abstract: The most vulnerable feature of most existing tall reinforced concretechimneys when subject to severe earthquake excitation is the relative weakness of thesections around the openings near the chimney base. This paper documents anexperimental study undertaken to investigate the cyclic behaviour of typical chimneysections with openings orientated to be bending critical and shear critical and theresults are compared with previous tests with no openings. Significantly, no suchcyclic tests have been previously been completed. The experimental results presentedinclude; failure mode, over-strength factor, ultimate curvature, available ductility,hysteretic behaviour and strain distribution and the results are compared with somepredictive analytical section models and design guidelines. The research indicated thatchimney sections with openings were not brittle, but were significantly less ductilethan sections without openings. The paper is relevant for both assessing existingchimney stacks and for the design of new chimneys in the vicinity of openings.

Key words: chimney, stack, seismic, cyclic behaviour, experimental tests, reinforced concrete, thin walled sections.

*Corresponding author. Email address: [email protected]; Fax: +61-3-9214-4882.

studies, a series of code design recommendations weredeveloped and incorporated into the 2001 CICIND codefor the design of reinforced concrete chimneys (Wilson2000). The recommendations included a reduction in theelastic seismic loads by a factor of R=2 throughdetailing for ductility and providing over-strengtharound openings to prevent the formation of brittlefailure modes.

A literature review has indicated few failures of tallreinforced concrete chimneys from earthquakeexcitation. This is not unexpected given theconservative design parameters traditionally specifiedfor the seismic design of chimney structures. TheTupras chimney is one of the few examples of chimneysthat have failed under seismic loading (Huang 2004;Kilic 2003). This refinery chimney stack was 115m talland failed in a brittle and catastrophic manner in thevicinity of the openings at around one third height. Anumber of researchers have studied the failuretheoretically using non-linear finite element techniques

1. INTRODUCTION

Tall reinforced concrete chimneys have traditionallybeen designed to resist earthquake excitation in theelastic range in the belief that such structures,characterised by a large diameter/thickness (D/t) ratio,were inherently brittle with no redundancy. This hasresulted in reinforced concrete chimneys beingprohibitively expensive in regions of high seismicity. Aseries of experimental tests which examined the cyclicbehaviour of chimney sections demonstrated that welldetailed chimneys were not brittle but possessed someductility (Wilson 2002, 2003). These testscomplimented some of the previous tests carried out oncircular hollow reinforced concrete sections where theD/t ratio was much smaller than those used in typicalchimney sections (Omote 1975; Mokrin 1985; Regan1981; Schober 1984; Whittaker 1987; Zhan 1990; Yeh2001). The experimental tests were used to develop anon-linear dynamic analysis procedure for evaluatingthe inelastic response of tall chimneys. Based on these

and concluded that the opening was lightly reinforcedand the failure was precipitated by the unzipping failureof the poorly detailed reinforcement splice connectionsin a region where flexural yielding occurred. This paperdescribes an experimental study commissioned to betterunderstand the complex behaviour around significantchimney openings under extreme cyclic loading and theresults are compared with previous tests undertakenwithout openings. Analyses indicated that the sectionwas weaker with the openings orientated to be ‘bending’critical, however there was considerable uncertaintywhether the section could successfully transfer the shearforces with the large openings orientated to be ‘shear’critical. Consequently, two tests were undertaken withthe openings orientated to be (i) bending critical and (ii)shear critical to study the behaviour.

2. RESEARCH SIGNIFICANCEThe most vulnerable feature of most existing tallreinforced concrete chimneys when subject to severeearthquake excitation is the relative weakness of thesections around the openings near the chimney base. Thispaper documents an experimental study undertaken toinvestigate the cyclic behaviour of typical chimneysections with openings orientated to be bending criticaland shear critical and the results are compared withprevious tests with no openings. Significantly, no suchcyclic tests have been previously been completed. Theexperimental results presented include; failure mode,over-strength factor, ultimate curvature, availableductility, hysteretic behaviour and strain distribution andthe results are compared with some predictive analyticalsection models and design guidelines. The research isrelevant for both assessing existing chimney stacks and forthe design of new chimneys in the vicinity of openings.

3. EXPERIMENTAL TEST SET-UPCircular hollow reinforced concrete specimens weredesigned and constructed to investigate the inelasticbehaviour of chimney sections under severe cyclicloading as shown in Figures 1 and 2. The reinforcedconcrete pipes were assembled and configured ashorizontal cantilevers and tested by applying a cyclictransverse load at the free end. The cantilever fixed endsupport consisted of a 320mm thick reinforced concreteblock (which simulated a pilecap) rigidly connected to asteel anchor block which was securely fixed to thelaboratory strong floor. The outside diameter of thepipes was D = 1194mm with a thickness of t = 35–40mmresulting in a diameter to thickness ratio of D/t = 30–35which was representative of real chimneys.

The control Test #1 had no openings (as shown inFigure 2a) whilst Test #2 and Test #3 (both t = 40mm)

The Cyclic Behaviour of Reinforced Concrete Chimney Sections with and without Openings

412 Advances in Structural Engineering Vol. 12 No. 3 2009

4565

2450

Steel tube

1

250kN Hydraulic actuator

Steel strap connection

Steel anchor block

R/C Pilecap

R/C pipe

2115

1200

Figure 1. Test arrangement

(a)

(b)

Figure 2. Test set-up, (a) Test #1; (b)Test #3

had 600mm wide openings, which is the largest openingsize is permitted in the CICIND and ACI307 codes ofpractice (subtended angle of 60°). The 600mm longopening for Test #2 was located 300mm from the fixedend and orientated to be bending critical (ie. theopenings were located in the region of maximum

bending stresses), whilst the 800mm long opening forTest #3 was rotated 90° to be shear critical as shown inFigure 2b. The 800mm long opening corresponded to 20times the chimney thickness which is the maximumallowable length under the ACI307 rules, for the designof reinforced concrete chimneys.

An axial load of 226KN was applied using two 16mmdiameter prestressing wires placed symmetrically top andbottom within the pipe void. The maximum length of pipewhich could be spun by the commercial pipe manufacturerwas 2.44m which was effectively reduced to 2.20m toallow a development length in excess of 200mm forcasting the pipe reinforcement into the foundation. Anadditional length of steel pipe was designed and fabricatedand the two pipes connected using twelve steel strapswhich were bolted and epoxied to the steel and concretesections respectively as shown in Figures 1 and 2. The4.60 metre length of hybrid pipe resulted in a shear span tomember diameter of 3.8, which was consideredrepresentative of reinforced concrete chimneys.

The longitudinal steel for all tests consisted of 5.8mmdiameter deformed bars with properties and percentagessummarised in Table 1 and characteristic of typicalchimneys. The reinforcement was embedded 200mm inthe anchor block and no laps were used. The hoop steelconsisted of 4.8mm diameter bars placed in a helix at80mm centres and corresponded to around ρh = 0.5%which was greater than the 0.2% minimum typicallyspecified for chimneys. In addition, three extra rebarswere placed within a 100mm wide strip each side of theopening in both the longitudinal and circumferentialdirections in accordance with the CICIND and ACI307recommendations. The key parameters for Tests #1, #2and #3 are summarised in Table 1 for completeness.

The cyclic lateral load was applied to the cantilevertube using a hydraulic actuator with a ±125mm travel

range. The behaviour of the reinforced concrete pipeunder pseudo static cyclic loading was monitored usingdisplacement transducers (LVDT), displacement dialgauges, Demec gauges (Tests #2, 3), photogrammetry(Test #3) and load cells. The pipe units were testedusing quasi-static cyclic load testing procedures andcyclically displaced to increasing ductility levels of:µ = ±0.75, ±1, ±2, ±3 etc, for Tests #1 and #2 andincremented at a finer resolution of ±0.25% driftincrements for Test #3. The yield displacement(displacement at µ = 1) was found by extrapolating themeasured secant stiffness at the lesser of either 0.75 ofthe theoretical ultimate load or at the onset ofreinforcement yield (Priestley and Park 1987). From thequasi-static cyclic load testing undertaken in thisresearch study on limited ductile structures, it isrecommended that the displacements be increaseddirectly using drift increments rather than ductilityincrements to ensure a sufficiently fine resolution of tipdisplacements for each cycle.

In constructing the reinforced concrete models, thelaws of similitude were followed so that the behaviourof the reinforcement, concrete and overall modelreflected the characteristics of a full scale prototypethus enabling the results to be directly scaled (Wilson2002). In particular, deformed reinforcement and atypical full scale concrete mix was used for the modeltests (with the restriction of a 10mm maximumaggregate size) to avoid the unrealistic bondcharacteristics associated with micro-concrete and toensure that the experimental results were representativeof equivalent full scale prototype tests. A fulldescription of the experimental investigation and testresults is provided in journal references Wilson (2002,2003) (Test #1) and the CICIND reports by Wilson(2003, 2006) (Tests 2 and #3).

John L. Wilson


Table 1. Summary of key parameters for Tests #1, 2 and 3

Parameter Test #1 Test #2 Test #3

Openings Nil 2 off 600 × 600mm 2 off 600 × 800mmLength 4565 mm 4600 mm 4345 mmExternal Diameter (D) 1194 mm 1194 mm 1194 mmThickness (t) 35 mm 40 mm 40 mmD/t 35 30 30Axial Force (N) 226 kN 226 kN 226 kNAxial Stress (fc) 1.75 MPa 1.55 MPa 1.55 MPaF′c 40 MPa 40 MPa 40 MPaRebar (pv) actual 0.94% 0.53% 0.53%Rebar (pv) equiv. 400MPa 0.70% 0.75% 0.75%Rebar stress fy 200 MPa 530 MPa 530 MPaRebar stress f3% 300 MPa 560 MPa 560 MPaRebar stress fu 400 MPa 590 MPa 590 MPa



−100

−80

−60

−40

−20

0

20

40

60

80

100

−100

(a)

−80 −60 −40 −20 0 20 40 60 80 100Displacement (mm)

For

ce (

kN)

0.0

−20.0

−40.0

−80.0

−60.0

20.0

40.0

60.0

80.0(b)

Displacement (mm)

For

ce (

kN)

−80.0 −60.0 −40.0 −20.0 40.020.0 60.0 80.00.0

(c)

Displacement (mm)

For

ce (

kN)

100

80

60

40

20

0

−20

−40

−60

−80

−100

−100 −80 −60 −40 −20 0 20 40 60 80 100

Figure 3. Lateral force versus displacement, (a) Test #1; (b) Test #2, (c)Test #3

4. OVERALL CYCLIC BEHAVIOUR4.1. Test #1

The chimney section with no openings (Test #1)behaved as a tough and ‘ductile’ structure as distinctfrom a ‘brittle’ structure as shown in Figure 3a whichplots the applied force versus tip displacement. Thespecimen achieved a tip displacement of ±85mm and a1.9% drift after 13 cycles. A series of cracks developedalong the length of the pipe, and opened and closed andlengthened as the longitudinal strains increased onsubsequent cycles.

The hysteresis shape was stable with increasingdisplacements associated with strain hardening of thereinforcement and increasing bending moments. Thereduction in stiffness associated with an increase in driftis characteristic of the closure of wide cracks, softening ofthe concrete matrix and the softening of the reinforcementdue to the “Bauschinger” effect. The pinched shape of thehysteresis loops is common for members with low axialloads. The loading section of the hysteresis loop is

consistent with a member undergoing inelasticdeformations with the tensile reinforcement yielding andwide cracks developing. The member displays quite highstiffness initially on unloading which is characteristic ofthe concrete unloading elastically in the compressionzone. Once the concrete stress has reduced to zero thecracks begin to re-open leaving a section with anessentially continuously open crack around thecircumference. Consequently the stiffness is greaterreduced (characterised by the pinched hysteresis), sincethe section stiffness is essentially provided by thereinforcement only. The section stiffness increases againon reloading as the cracks close and the concretecontributes to the stiffness as the concrete compressivestresses develop. The pinching effect is less severe onmembers with a high axial load since the neutral axisdepth will be greater and a portion of concrete will remainin compression at all times to carry the axial load.

This ductile behaviour was achieved through yieldingof the reinforcement in tension, rather than non linear

John L. Wilson


Strain

For

ce (

kN)

100

80

60

40

20

0

−20−0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07

−40

−60

−80

−100

Figure 4. Lateral force versus extreme fibre strain Test #1

Figure 5. Damage in vicinity of fixed end Test #1

compressive behaviour of the concrete. For example ata drift of 1.9%, the average maximum tensile strainmeasured over a gauge length of 200mm was 5% (withcrack widths in the order of 3mm) compared with acompressive strain less than 0.3% as shown in Figure 4.

As the cyclic drift increased, the cover concretearound the hoop and longitudinal reinforcement inthe vicinity of the crack began to steadily spall, as theconcrete was cycled back and fourth from extremetension to compression. The damage developed in azone of concrete approximately 350mm wide adjacentto the opening and was characterised by circumferentialand diagonal cracking (refer Figure 5). The pipe wasdeemed to fail at a tip displacement of 99mm and driftof 2.2%) when two of the longitudinal steel bars whichhad buckled on the previous cycle in compressionfractured in tension. The buckling of the reinforcementbar occurred due to the loss of the concrete cover in thevicinity of the opening combined with the reduced Evalue from the ‘Bauschinger’ effect. The buckling of thelongitudinal steel resulted in further concrete spalling

and a sharp kink developing in the reinforcement. Thelongitudinal steel then fractured at the kink when theload direction was reversed and the reinforcement wassubjected to extreme tensile strains. Transverseconfinement steel at a spacing of 5db (where db is thelongitudinal steel diameter) would be required toprevent the longitudinal reinforcement from buckling.

4.2. Test #2

The chimney section with openings orientated to be‘bending’ critical performed in a similar manner to Test #1except the behaviour could better be described as‘limited ductile’ with an ultimate tip displacement of±67mm and a 1.5% drift, after 5 cycles (refer Fig 3b).The average maximum tensile strain measured over agauge length of 200mm was 1.2% compared with acompressive strain less than 0.3%. The lateral forceversus longitudinal strain plot was similar to that ofFigure 4 except the maximum strain was 1.2% whichwas considerably less than the 5.0% measured for Test #1.The specimen was deemed to fail at a tip displacementof 76mm and a drift of 1.6% with the buckling andsubsequent fracture of exposed reinforcement barsadjacent to the opening in a similar manner to the failurein Test #1 (refer Figure 6).

The local strains measured with Demec gauges weregenerally symmetrical around the openings withsignificant longitudinal strains adjacent to the edge ofthe opening. The maximum longitudinal tensile strainsmeasured over a gauge length of 200mm were in theorder of 1.2% with localised strains varying widely inthe range 0.2%–4.0% (measured over a gauge length of50mm) reflecting the crack pattern in the pipe. Thelongitudinal strains in the pipe near the centre at eachend of the opening were negligible and indicatedsignificant local stress redistribution. Thecircumferential strains and stresses around the openingswere minimal and generally less than the crackingstrength of the concrete resulting in minimal stressesbeing transferred to the additional circumferentialreinforcement. As expected the diagonal strains at thecorners of the openings were significant with localstrains varying in the range 0.7%–4.0%.

Significant distortion of the section in the vicinity ofthe openings was observed, with the edge of the openingbuckling on the compression face at drift levels ofaround 1.1%. Figure 6(c) shows the formation of an out-of-plane buckle, providing clear evidence that thechimney section in the vicinity of the opening wasdistorting locally and violating the assumption thatplane sections remain plane. As a result of the observeddistortions, the 3D photogrammetry technique wasdeveloped and introduced in Test #3 to measure the

in-plane ovalling effects which could not be capturedwith the linear transducers used to measure the surfacestrains. (The photogrammetry technique utilises a highresolution digital SLR camera, scalebar, a network ofretro-reflective targets and sophisticated software tointerpret the data. Photographs of the targets are takenbefore and after an event and the relative changes tothe 3D position of the targets is calculated using theprinciples of triangulation from which average surfacestrains can be estimated).

4.3. Test #3

The chimney section with openings orientated to be‘shear’ critical performed in a similar manner to Test #1except the behaviour could better be described as‘limited ductile’ with an ultimate tip displacement of±78mm and a 1.8% drift after 8 cycles (refer Figure 3c).The average maximum tensile strain measured over agauge length of 200mm was 1.2 % compared with acompressive strain less than 0.3%. The lateral forceversus longitudinal strain plot was similar to that of



(a) (b)

(c)

Figure 6. (a) Overview of damage Test #2; (b) Compression failure damage Test #2; (c) Elastic buckling along free edge of opening,

Test #2

John L. Wilson


(a)

v1

h1

h2

v2

Original shape

Deformed shape

Figure 4 except the maximum strain of 1.2% wasconsiderably less than the 5.0% measured for Test #1.

The specimen was deemed to fail at a tipdisplacement of 87mm and a drift of 2.0% when theconcrete shell on the compression side (between the 2openings) buckled as a wide column panel adjacent tothe fixed end in a sudden, brittle and explosive mannerafter significant inelastic deformations (refer Figure 7a).The shell buckling was precipitated by buckling of thelongitudinal steel bars which had become exposeddue to the deterioration of the concrete and bond in theanchor block adjacent to the fixed end. Surprisingly, thesection did not fail in shear, although significant sheardistortion was observed around the openings resultingfrom the 3D shell action. The shear distortion is alsodemonstrated in Figure 7(b) which shows the elasticovalling deformation of the cross-section at theopenings from photogrammetry measurements at a drift

of around 1.8%. Further, the longitudinal strainsmeasured in the vicinity of the openings did not satisfythe assumption of ‘plane sections remain plane’providing further evidence of the local shear distortion.

5. EXPERIMENTAL AND THEORETICALPERFORMANCE OF CHIMNEY

5.1. First Cracking

The cracking strength was calculated based on a tensilestrength of concrete given by ft = α. where α =0.5 - 0.75 and using simple elastic section analysistheory. These values of α are considered conservativelyhigh values for concrete in direct tension and directflexure respectively. In this application the tensilestresses were considered uniform across the wallthickness of the pipe and consequently a value of α = 0.5has been assumed, resulting in ft = 3.2MPa. Table 2shows a comparison of the calculated and actual bendingmoments at first cracking and show overall goodagreement, with some variation associated with theuncertainty of the tensile cracking strength and somelocal thickness variations associated with the fabricationprocess of the pipe sections.

5.2. First Yield

The bending moment strength at first yield wascalculated based on the reinforcement bar yield stressand using simple elastic section analysis theory ignoringthe tensile strength of the concrete. Table 3 shows acomparison of the calculated and actual bendingmoments at first yield and show overall good agreementbetween the experimental and theoretical results.

For completeness, the measured yield curvatures andyield tip displacements are presented in Table 4. Themeasurements for Tests #2 and #3 indicate that the pipewith openings was considerably more flexible and alsoindicated that the yield stress of the reinforcement barwas considerably higher than that used for Test #1.

5.3. Ultimate Bending Moment Strength

The ultimate moment capacity was calculated usingstandard ultimate section analysis theory (assumingplane sections remain plane and using the stress-strainproperties for concrete and steel reinforcement) andusing the design charts from the 2001 CICIND modelconcrete code (similar method to that recommended inACI 307). The calculated moment values from thedesign charts have been interpolated by using theequivalent percentage of 400MPa reinforcement and bymultiplying the design values by 1.15 to account for thepartial safety factor incorporated in the charts.

A comparison of the experimental and theoreticalvalues is listed in Table 5 and indicates that the

′F c

Figure 7. (a) Explosive compressive failure Test #3; (b) Ovalling

distortion of Test #3 pipe opening (Max distortion at 1.8% drift:

v2-v1 = 18mm and h2-h1 = 13mm)

experimental strengths are significantly greater than thetheoretical values. The CICIND chart values for Tests #2and #3 have been assumed to be the same and equal toMu = 250kNm, since the charts do not specify thecritical orientation of openings. The over-strengthvalues presented in Table 5 are based on the nominaldesign values from the CICIND charts and have beencalculated using Test #2 as an example: Ω = 1.15 ×292/250 = 1.15 × 1.17 = 1.35. The result is particularlysignificant since it indicates that the section withopenings orientated to be ‘shear’ critical (Test #3) isaround 1.3 times stronger than the ‘bending’ criticalcase (Test #2), and that the section is capable oftransmitting the shear forces around the wide openingthrough 3D shell action. In addition, the experimentalresults demonstrated that the usual ‘plane sectionsremain plane’ assumption is not valid for calculating theultimate strength due to the significant stressredistribution and shell action (this is the subject of aseparate investigation using non-linear finite elementanalyses). The fact that the specimen with an opening(Test #3) is stronger than the specimen with no opening(Test #1) is a reflection of the different reinforcementproperties used in Test #1, and the fact that additionalreinforcement was placed around the openings.

5.4. Ultimate Curvature

The maximum ultimate curvatures and maximumlongitudinal tensile strains averaged over a 200mmgauge length are listed in Table 6. The measurementsindicate that sections with openings develop maximumstrains and curvatures in the order of only one quarterthat of sections with no openings, showing the limitedstrain capacity. Interestingly, the maximum drifts werecomparable for the 3 tests, indicating that theconsiderable distortion around the openings contributedsignificantly to the tip displacements for Tests #2 and 3.The Test #1 curvature of 0.051 m–1 compares very wellwith a theoretical value of 0.053 m–1calculated usingstandard ultimate section analysis theory assumingplane sections remain plane.

5.5. Ultimate Displacement

The observed damage and curvature ductilitydistribution indicated that most of the inelastic curvatureoccurred over a 350mm section (0.30 times thediameter) near the base and in the vicinity of the openingsfor all 3 tests. This concentration of curvature over adefined length can be modelled as a plastic hinge with anominal plastic hinge length. Using the nominal plastichinge length and average curvature values the cantilever



Table 2. Cracking bending moment values for Tests #1, 2 and 3

Cracking moment Test #1 Test #2 Test #3

Experiment 237 kNm 108 kNm 161 kNmTheory 210 kNm 110 kNm 186 kNm

Table 3. Yield bending moment values for Tests #1, 2 & 3

Yield moment Test #1 Test #2 Test #3

Experiment 224 kNm 194 kNm 239 kNmTheory 220 kNm 215 kNm 310 kNm

Table 4. Experimental yield curvature and displacement values for Tests #1, 2 & 3


Yield curvature 0.0012 m−1 0.0025 m−1 0.0022 m−1

Yield displacement 4.4mm 13mm 15mm

Table 5. Ultimate bending moment values for Tests #1, 2 and 3

Ultimate Moment Test #1 Test #2 Test #3

Experiment 369 kNm 292 kNm 391 kNmTheory 310 kNm 300 kNm 400 kNmCICIND Charts 326 kNm 250 kNm 250 kNmOver-strength Factor 1.30 1.35 1.80

tip displacement, ∆u, can be estimated assuming thedisplacement consists of an elastic component equal tothe displacement at yield and a plastic componentrepresenting deformations associated with the plastichinge as shown in Eqn 1:

∆u = F/Fy × ∆y + (φ – F/Fy φy) lp (l-0.5 lp) (1)

F = Lateral force applied to the pipe systemFy = Lateral force at yield∆y = Yield deflection at Fy

φ = Average curvature in plastic hinge at Fφy = Yield curvature at Fy

lp = Nominal plastic hinge lengthl = Length of cantilever to openingThis equation was applied to the test pipe data, and the

plastic hinge length back calculated so that the predictedand actual tip displacements matched. The calculatedplastic hinge lengths are summarised in Table 7and indicate an excellent correlation with the observeddamage region of Test #1 but totally unrealisticcorrelations for Tests #2 and 3. Clearly this approachbased on ‘plane sections remaining plane’ and assumingthat significant inelastic behaviour is associated withflexure only, does not provided a good match for chimneysections with large openings where 3D distortions of the section, shear effects and second order effectscontribute significantly to the overall ultimate displacement.

The approximate contribution of elastic, shear andplastic deformations to the overall tip displacement forthe 3 tests is summarised in Table 8. The elastic andplastic contributions were calculated from Eqn 1 using a

plastic hinge length of 350mm, whilst the shearcontribution was calculated as the additionaldisplacement needed to equal the total tip displacement.

6. CONCLUSIONS AND SUMMARY1. Specimen #1 with no opening displayed ‘ductile’

behaviour, whilst specimens #2 and #3 withopenings displayed ‘limited ductile’ behaviour,but importantly not ‘brittle’ behaviour’. Allspecimens achieved a drift in the range of1.5%–1.9%. The failure mode of the specimenwith openings orientated to be ‘shear’ critical(Test #3), was a sudden and explosive bucklingof the compression panel adjacent to the openingafter considerable inelastic excursions, andsurprisingly not due to direct shear failure.

2. The limited ductile behaviour exhibited by allthe pipe specimens is a result of smallcompressive strains in the unconfined concreteand larger tensile strains in the ductilelongitudinal reinforcement. All specimensdeveloped a damage zone adjacent to the fixedend or opening, consisting of a 350mm zone(0.3D) of significantly cracked concrete whichopened and closed and slowly degraded as thedisplacements increased. The ultimatelongitudinal strains and curvatures in the specimens with openings were only 1/4 ofthe values of the specimens with no openings.

3. The ultimate strength of the chimney sectionwith openings was around 1.35 and 1.8 times


John L. Wilson

Table 6. Ultimate curvature and maximum tensile strains for Tests #1, 2 and 3


Ultimate curvature 0.051 m–1 0.012 m–1 0.012 m–1

Maximum strains 5.0% 1.2% 1.2%Maximum drift 1.9% 1.5% 1.8%

Table 7. Plastic hinge lengths for Tests #1, 2 and 3

Plastic hinge length Test #1 Test #2 Test #3

Observed 350mm 350mm 350mmCalculated 350mm 1700mm 2200mm

Table 8. Displacement contributions for Tests #1, 2 and 3

Displacement Test #1 Test #2 Test #3

Elastic 10mm 20mm 25mmShear ~ 0 30mm 42mmPlastic 76mm 17mm 11mmTotal 86mm 67mm 78mm

stronger than the nominal strength predictedfrom the CICIND (or ACI 307) design charts forthe openings orientated to be ‘bending’ and‘shear’ critical respectively. The result highlightsthat the section with openings orientated to be‘shear’ critical has considerable overstrengthavailable and is around one third strongercompared with the ‘bending’ critical orientation.

4. Significant distortion was experienced in thesection around the openings demonstrating thatthe assumption of plane sections remainingplane was violated. Consequently the tipdisplacements could not be predicted fromsimple models using the concept of a plastichinge due to the extensive deformation fromshear and 3D ovalling distortions. Further, clearevidence of elastic buckling of the free edgearound the opening was observed at drift levelsin the order of ±1.1% for Test #2.

5. The experimental research has improved theunderstanding of the inelastic cyclic behaviourand actual strength of chimney sections withlarge openings which is relevant for assessingexisting chimney stacks and for the design ofnew chimneys. A further study is currently inprogress to compare the experimental resultswith a detailed non-linear and inelastic finiteelement analysis to model the complex 3Dbehaviour in the vicinity of the openings wherethe standard assumption of ‘plane sectionsremain plane’ is violated.

ACKNOWLEDGEMENTSAcknowledgements and appreciation are extended tothe CICIND organisation and University of Melbournefor the financial assistance and particularly Grant Rivettfor the technical support with the experimental program.Additional thanks to PhD students Kittipoon Rodsin andDavid Heath for assistance with the experimental dataanalysis and photogrammetry study for Test #3.

REFERENCESACI Committee 307. (1998). Standard Practices for the Design and

Construction of Cast-in-Place Reinforced Concrete Chimneys,

American Concrete Institute, Farmington Hills, Michigan, USA.

CICIND. (2001). Model Code for Concrete Chimneys - Part A: The

Shell (commentary), International Committee on Industrial

Chimneys, Switzerland.

Huang, W., Gould, P.L., Martinez, R. and Johnson, G.S. (2004).

“Non-linear analysis of a collapsed reinforced concrete

chimney”, Earthquake Engineering and Structural Dynamics,

Vol. 33, pp. 485–498.

Kilic, S.A. and Sozen, M.A. (2003). “Evaluation of effects of August

17, 1999, Marmara earthquake on two tall reinforced concrete

chimneys”, ACI Structural Journal, Vol. 100, pp. 357–364.

Mokrin, Z.A.R. and Rumman, W.S. (1985). “Ultimate capacity of

reinforced concrete members of hollow circular sections

subjected on monotonic and cyclic bending”, ACI Journal,

Vol. 82, No. 5, pp. 653–656.

Omote, Y. and Takeda, T. (1975). Experimental and Analytical

Study on Reinforced Concrete Chimneys, Japan Earthquake

Engineering Promotion Society, Tokyo.

Priestley, M.J.N. and Park, R. (1987). “Strength and ductility of

concrete bridge columns under seismic loading”, ACI Structural

Journal, Vol. 84 No. 2, pp. 61–76.

Regan, P.E. and Hamadi, Y.D. (1981). “Behaviour of concrete

caisson and tower members”, in Concrete in the Oceans,

Technical Report No. 4, CIRIA/UEG, Cement and Concrete

Association, Department of Energy, United Kingdom.

Schober, H. and Schlaich, J. (1984). “Ultimate strength of reinforced

concrete chimneys”, Proceedings of the CICIND 5th

International Conference, Essex UK, pp 37–42.

Whittaker, D. (1988). Seismic Performance of Offshore Concrete

Platforms, Report No. 88-1, University of Canterbury,

Department of Civil Engineering.

Wilson, J.L. (2000). “Code recommendations for the aseismic design

of tall reinforced concrete chimneys”, CICIND Report, Vol. 16,

No. 2, pp. 8–12.

Wilson, J.L. (2002). “Aseismic design of tall reinforced concrete

chimneys”, ACI Structural Journal, Vol. 99 No. 5,

pp. 622–630.

Wilson, J.L. (2003). “Earthquake response of tall reinforced concrete

chimneys”, Engineering Structures, Vol. 25, No. 1, pp. 11–24.

Wilson, J.L. (2003). “Investigation of the cyclic behaviour of

reinforced concrete chimney sections with openings: Part 1 –

Experimental study”, CICIND Report, Vol. 19 No. 2, pp. 9–15.

Wilson, J.L. (2006). “Investigation of the cyclic behaviour of

reinforced concrete chimney sections with openings orientated to

be shear critical”, CICIND Report, Vol. 21 No. 2, pp. 43–49

Yeh ,Y.K., Mo, Y.L. and Yang, C.Y. (2001). “Seismic performance

of hollow circular bridge piers”, ACI Journal, Vol. 98, No. 6,

pp. 862–871

Zhan, F.A., Park, R. and Priestley, M.J.N. (1990). “Flexural strength

and ductility of circular hollow reinforced columns without

confinement on inside face”, ACI Structural Journal, Vol. 87,

No. 2, pp. 156–166.



Documents

The Cyclic Behaviour of Reinforced Concrete Chimney Sections with … · Advances in Structural Engineering Vol. 12 No. 3 2009 411 The Cyclic Behaviour of Reinforced Concrete Chimney