eactor Safety Division, Bhabha Atomic Research Centre, Mumbai 400085, India
a b s t r a c t
habha Atomic Research Centre (BARC), Trombay has organized an International Round Robin Analysis program to carry out the ultimate load capacityssessment of BARC Containment (BARCOM) test model. The test model located in BARC facilities Tarapur; is a 1:4 scale representation of 540 MWeressurized Heavy Water Reactor (PHWR) pre-stressed concrete inner containment structure of Tarapur Atomic Power Station (TAPS) unit 3&4. There are
large number of sensors installed in BARCOM that include vibratory wire strain gauges of embedded and spot-welded type, surface mounted electricalesistance strain gauges, dial gauges, earth pressure cells, tilt meters and high resolution digital camera systems for structural response, crack monitoringnd fracture parameter measurement to evaluate the local and global behavior of the containment test model.The model has been tested pneumatically during the low pressure tests (LPTs) followed by proof test (PT) and integrated leakage rate test (ILRT) during
ommissioning. Further the over pressure test (OPT) has been carried out to establish the failure mode of BARCOM Test-Model. The over-pressure testill be completed shortly to reach the functional failure of the test model. Pre-test evaluation of BARCOM was carried out with the results obtained from
he registered international round robin participants in January 2009 followed by the post-test assessment in February 2011. The test results along withhe various failure modes related to the structural members – concrete, rebars and tendons identified in terms of prescribed milestones are presented inhis paper along with the comparison of the pre-test predictions submitted by the registered participants of the Round Robin Analysis for BARCOM test
odel.
. Introduction
Over more than three decades, the demonstration of robust con-ainment design has become important for the public acceptance ofuclear power plants, which needs to be verified through the func-ional and structural failure mode tests. The earlier beyond designasis accidents at Three Mile Island-1979 (USA) and Chernobyl-986 (former USSR) and more recently at the Fukushima-2011Japan) nuclear plants have led to wide attention on the ulti-
ate load capacity assessment of nuclear containments structures.he design, safety and regulatory requirements demand ruling outontainment structural failure by design measures and hence andequate factor of safety over the design pressure has been empha-ized (see for example Liersch et al. (1994), Davies et al. (1995) andeview of various Sandia model tests reported by Hessheimer andameron (2006)). The assessment of old containment structuresnd design and safety assessment of new nuclear containmentseed to follow the standard practice of the pre-stressed concreteeactor vessel design philosophy of ASME design code section IIIiv-2 (ASME Boiler and Pressure Vessel Code, 2007), where a factorf safety of 2 is assured over the design basis accident.
The ultimate load capacity assessment of all the nuclear plantsn India has been carried out over the years using in-house com-uter codes ULCA and ARCOS-3D in Bhabha Atomic Research Centre
(BARC). These codes have earlier been extensively used to predictthe ultimate load capacity of 220/540 MWe PHWR containmentdesigns and a factor of safety ∼2.3–2.5 over the design pressurehas been demonstrated (Singh et al., 1993; Gupta et al., 1995).BARC, Trombay also participated in the round robin analysis ofPre-Stressed Concrete Containment Vessel (PCCV) of a typical PWRsteel lined pre-stressed concrete containment sponsored by SandiaNational Laboratory, USA and Nuclear Power Engineering Corpora-tion, Japan during the pre-test phase (PCCV, 1997). The predictionsmade by BARC in-house finite element code ULCA, developed bySingh et al. (1993) and Basha et al. (2003a) were found to be inexcellent agreement with the observed and published test results,as reported in Basha et al. (2003a,b) and Singh (2007).
Further for the ultimate load capacity assessment of IndianPHWR nuclear containments, 1:4 size BARC Containment (BAR-COM) Test Model (Fig. 1) representing 540 M We PHWRpre-stress concrete inner containment with design pressure (Pd)of 0.1413 MPa has been constructed and commissioned at Tarapursite (BARC, 2006; Singh et al., 2009). Under this containment safetyresearch program initiated at BARC Trombay for Indian PHWRs,BARCOM performance is being evaluated with the internationalround robin analysis results obtained from 15 registered partici-pants during the pre-test and post-test phases (Singh, 2009, 2011a).
The objective of the present test program is to obtain the pres-
sure, displacement and strain data related to the various functionaland structural failure modes of BARCOM in terms of the loss ofpre-stress in the membrane and discontinuity regions of majoropenings, first appearance of concrete surface cracks followed
y first through thickness cracks, first yielding of reinforce-ent/tendons and significant loss of leak tightness, the maximum
ressure sustained by the model before significant leakage to iden-ify functional failure pressure of the test-model and finally the
aximum static pressure sustained by the model for the structuralailure assessment. All these assessments would finally lead to thestimation of the functional and structural behavior of the contain-ent model up to the ultimate pressure for beyond the design basis
evere accident studies.The experiments on BARCOM Test Model have been planned in
he following four phases:
. Phase I – Calibration of sensors during low pressure test (LPT)-limited up to 0.35 Pd.
. Phase II – Proof test (PT) and Integrated leakage rate test (ILRT)up to 1.0 Pd.
. Phase III – Over pressure test (OPT) up to the Functional Failureof model.
. Phase IV – Ultimate load capacity test (ULCT) of BARCOM.
The Phase-I pressurization of BARCOM Text Model up to.049 MPa (0.35 Pd) was carried out for the functionality tests ofarious sensors and data logger systems consisting of 350 surfaceounted electrical resistance strain gauges (SMERs), 768 vibra-
ory wire strain gauges (VWSGs), 32 dial gauges, 9 potentiometerased dial gauges, 20 Tilt-meters, 17 RTDs, and 17 earth pressureells installed on the BARCOM. Fig. 1 gives an overview of variousmbedded sensors in BARCOM along with an exhaustive cable workp to the Control & Instrument (C&I) Building. Subsequently, BAR-OM Test-Model was pressurized up to the design pressure (Pd)f 1.44 kg/cm2 (0.1413 MPa) on August 06, 2010 at 01.00 AM andince then three more design pressure tests have been completeduring October–December, 2010 under the Phase-II experiments.he commissioning and Proof Test of BARCOM has been completeduccessfully and data from all the sensors were recorded during theressurization and depressurization cycles. In addition, consistentnd repeatable leakage-rate was obtained during all the Phase-IIxperiments. Further under the Phase-III over-pressure test (OPT)rogram, the BARCOM Test Model was pressurized up to a pressure
f 0.2207 MPa (1.56 Pd) on December 17, 2010. The milestone withegard to “first appearance of crack” was recognized with onlineonitoring of the inelastic strain developed in the discontinuity
egions of Main Air Lock (MAL) and Emergency Air Lock (EAL) with
embedded VWSGs, which was also confirmed with soap bubble testduring depressurization at 0.0981 MPa. A second OPT was also con-ducted during January 12–15, 2011 to check the repeatability of thetest data and obtain the localized strain field in the fracture processzone (FPZ) at few selected locations. The strain pattern obtainedshowed development of parallel cracks with localized strain field inthe fracture process zone (FPZ) near MAL, EAL locations and the firstthrough thickness cracks in BARCOM test-model were identifiedsuccessfully.
For the pre test predictions and post test ultimate load capacityevaluation of BARCOM test model, fifteen participants from Austria(1), Brasil (1), Czech Republic (1), Finland (1), France (2), SouthKorea (2), United Kingdom (3) and India (4) have registered. Theregistered round robin participants submitted the pre-test predic-tions at sixty-nine (69) specified sensor locations (SSL) out of total∼1100 sensors for studying BARCOM failure modes in terms of pre-scribed milestones, which were discussed in pre-test meeting andworkshop in 2009 (Singh et al., 2009; Singh, 2009). The post-testmeeting and workshop after the above mentioned over-pressuretests (OPTs) was convened in February 2011 to appraise the reg-istered round robin participants on the test results (Singh, 2011a).There was an overall agreement on the observed failure mode ofBARCOM and the participants will now carry out improved posttest computations with this interim feedback for comparison of thetest results. This will aid in improved understanding of the contain-ment performance under over-pressure and help in benchmarkingvarious numerical finite element inelastic codes.
2. BARCOM experimental program and international roundrobin analysis
After the completion of the civil works for the auxiliary shed,installation and commissioning of the equipments and piping sys-tem for pressurization system and MALB door fabrication andinstallation, the commissioning tests on BARCOM Test Model(Fig. 1) were initiated. Additional 350 surface type sensors andprocess parameter sensors were installed and cable routing worksfrom the test model to the Control and Instrumentation (C&I)building and installation, commissioning and configuration of data
loggers were completed in parallel to initiate the experimental pro-gram (Fig. 1). As per the plan, the model was tested pneumaticallywith compressed air during the low pressure tests (LPTs) fol-lowed by proof test (PT) and integrated leakage rate test (ILRT) for
BARC-Tarapur test facility with details of embedded sensors and cable panels.
eering
cdpbptttts
1
2
3
4
5
6
2
jt
•
••
••
•
••
•
•
•••
tfapmbt
R.M. Parmar et al. / Nuclear Engin
ommissioning of the test model. A few additional tests at theesign pressure (Pd = 0.1413 MPa) of the model were also com-leted under Phase-II. Presently, the over pressure test (OPT) haseen carried out to establish the failure mode of BARCOM and thishase-III experiment is planned to be completed up to the func-ional failure of the test model. After this major repair work will beaken up with installation of internal rubber liner to enable con-inued pressurization of BARCOM till the ultimate load capacityest (ULCT). The test matrix along with the schedule of activities isummarized below:
. Calibration of sensors during LPTs (limited up to 0.35 Pd) – Com-pleted May 2010.
. Proof test (PT) and integrated leakage rate test (ILRT) up to 1.0 Pd– Completed August 06, 2010 at 01.00 AM.
. Three additional tests at design pressure up to 1.0 Pd – Com-pleted October–December, 2010.
. Over pressure test (OPT) up to 1.56 Pd to establish the failuremode of model – Completed December 15–19, 2010 and January12–14, 2011.
. Over pressure test (OPT) up to the Functional Failure of model –July–August 2011.
. Ultimate load capacity test (ULCT) – February–March 2011.
.1. Milestones and failure modes of BARCOM
The milestones and different failure modes of BARCOM areudged with plots at 69 Specified Sensor Locations (SSLs) prescribedo the round robin participants as included below:
First appearance of concrete cracking in wall and dome in hoopand longitudinal directions.First through wall thickness cracking in the wall and dome.First appearance of crack at discontinuity regions such as basewall junction and near ring beam.First cracking of dome at SG opening, MAL, EAL and FMLB.First yielding of the reinforcement bars in hoop and longitudinaldirections and subsequent strain levels at the various pressurelevels.Total loss of pre-stress in concrete in hoop and longitudinal direc-tions (zero hoop and longitudinal stresses when the pre-stressinduced compressive stresses are overcome.).First yielding in tendon.Tendon strain levels and gradual change in the pre-stress forceduring pressurization.First significant loss of leak tightness and the functional failureof the model based on tendon strain level of 1% – loss of pre-stressing function of tendons.Maximum pressure sustained by the model before massive leak-age signifying the functional failure based on tendon strain levelof 1% in cylinder and dome membrane region, the discontinuityregions and around openings.Tendon strain level of 2% & 3%.Tendon rupture at strain level of 3.5%.Ultimate collapse pressure of the containment test model signi-fying the structural failure including the best estimate, the lowerand upper bounds and the criteria used for arriving at the same.
The present round robin exercise has been planned to obtainhe response from the sensors at all the critical SSLs of BARCOM toacilitate posttest experimental data comparison with the pretestnalysis predictions made by the various registered round robin
articipants. Global and local deformations of the model have beenonitored by VWSG based displacement transducers and SMER
ased dial gauges. The local strains in concrete, steel rebars andendons are obtained with VWSG and SMERs. The participants will
and Design 269 (2014) 177– 183 179
provide posttest analysis results of the specified output quantitiesversus gage pressure at these locations. The selections of these loca-tions (SSL) are based on our containment analysis experience andthe judgment based on the preliminary inelastic analysis.
2.2. Tendon characterization in laboratory and in situ tendonresponse during pre-stressing
Fig. 2 shows the different tendon profiles used for BARCOM con-struction, which are horizontal H and C cables in wall and domerespectively and J cables from ring beam to raft. The stress–straincharacteristics of seven tendon samples have been obtained in lab-oratory tests with dial gauge and six strain gauges on differentstrands. The 0.2% Proof Stress, Ultimate Stress, and Modulus of Elas-ticity, stress (load) for 1%, 2% and 3% tendon strain levels wereestimated from the stress–strain data and all were very closelyrelated as shown in Fig. 2. Mean and standard deviations from theseven sets of tensile test data have been derived for the tendonstress–strain characteristics up to failure and the measurementwith strain gauges and dial gauges are consistent. The followingobservations can be made on the test results:
• The elastic modulus calculated using both the strain gauge anddial gauge data was found to be very close.
• In most of the cases, the elastic modulus was determined usingthe strain gauge data in preference to the dial gauge data due toits better accuracy and precision.
• It was possible to obtain tendon stress (load) at 1%, 2% and 3%strain levels in majority of the cases, which is considered suffi-cient for studying the behavior of tendons in BARCOM experimentand numerical analysis.
During pre-stressing of the tendons, strains and jack loads wererecorded. The strains were measured with the SMER type straingauges in the tendons. The behaviors of the tendons are plottedalong with the layout of the tendon as shown in Fig. 3. The followingobservations can be made:
• Tendons were stressed well within the yield limit as shown in thecomparison plots (Fig. 3).
• The in situ stress–strain behaviors were found to be comparablewith the laboratory test data.
• However in most of the tendons in situ stress–strain behaviorswere found to be stiffer than the tendon laboratory test data. Thiscould be due to pre-stressing losses.
• The plot indicates the stress (jack force) at the ends while thestrains at the different locations can be converted to stress at aparticular gauge location using elastic modulus from the labora-tory test data.
• The present pre-stressing data can be used to estimate losses forrefined numerical computations.
2.3. Pre-stressing, phase-II design pressure test and phase-IIIover-pressure test results
The strain data obtained during the construction period includ-ing pre-stressing (October 2007–December 2008) and beyond upto a period of one more year (up to December 2009) were obtainedfrom embedded VWSG sensors. Pre-stressing load induced strainsat these locations were identified after separating the creep andshrinkage strain components from the overall strain data with helpof pre-stressing schedule. As an example, in Fig. 4 for a typical set of
VWSGs in hoop (SSL-10 & 12) and longitudinal directions (SSL-11& 13) located near the inner and outer surfaces of BARCOM wall,the hoop strain in the range of 272–353 �� and longitudinal strainin the range of 97–109 �� were obtained from the experimental
180 R.M. Parmar et al. / Nuclear Engineering and Design 269 (2014) 177– 183
files a
dastipdmv
Fig. 2. Pre-stress tendon pro
ata analysis as compared to average computed value of 292 ��nd 99 ��, respectively based on the jack load, which shows con-istent response of embedded VWSGs. Fig. 5 describes the phase-IIest data at SSL-10 up to the design pressure (Pd) obtained dur-ng various tests and all the tests show repeatability. The pre-test
redictions from the round robin participants were obtained andiscussed in January 2009 meeting/workshop, a typical displace-ent plot for SSL-2 is shown in Fig. 6. The pre-test results from
arious participants at SSL-23 are compared with tests data up to
Fig. 3. Pre-stress tendon response during pre
nd tendons characterization.
Phase-II as shown in Fig. 7. The inelastic strain responses have beenidentified near openings during the two Over-Pressure Tests (OPTs)sup to 1.56 Pd during December 2010 and January 2011 as is shownfor SSL-27 near MAL in Fig. 8. These were confirmed during thesoap bubble tests (Fig. 9) with first appearance of through thick-
ness cracks during the de-pressurization at 0.0981 MPa. Anotherconfirmation was obtained from Optical Crack Profiling to identifythe fracture process zone (Fig. 9), more details on this are presentedin another paper by Singh (2011b). With the availability of all these
-stressing & comparison with lab test.
R.M. Parmar et al. / Nuclear Engineering and Design 269 (2014) 177– 183 181
Long t
Hoop
Axial
-10
-8
-6
-4
-2
Str
ain
(m
icro
str
ain
)
ter m data colle
direction SS L
direc tion SS L
19-Jun-07
19-Dec
000
800
600
400
200
0
ection for creep
L-10 (272 με),
L-11 (97 me), S
c-07
19-Jun-08
19-De
Time (Day
SSL-10 /VC
p, shrink age an
SSL- 12 (353 μSSL-1 3 (109 m
Dec-08
19-Jun-09
19-D
ys)
C-2113
nd during pre- s
με),Compu ted
me) ,Co mput ed
Dec-09
1
-10 00
-800
-600
-400
-200
0
Str
ain
(m
icro
str
ain
)
tress ing of ten d
d 292 μεd 99 με
19-Jun-07
19-Dec-07
S
dons
19-Jun-08
19-Dec-0
Time (Days)
SSL -12 /VC-2
0
19-Jun-09
19-Dec
211 1
ec-09
Fig. 4. Creep-shrinkage and pre-stress data on tendons.
Fig. 5. Phase-II test-typical response at SSL-10.
Fig. 6. Pre-test predictions by round robin participants at SSL-2.
182 R.M. Parmar et al. / Nuclear Engineering and Design 269 (2014) 177– 183
Fig. 7. Pre-test comparison up to 1.0 Pd with test data at SSL-23.
Fig. 8. Response during over-pressure test in BARCOM near MAL at SSL-27.
Fig. 9. First appearance of crack in BARCOM test model during over-pressu
re test phase-III experiment (December 17, 2010 and January 2011).
eering
itoswa
3
Rrasdri7rrgiitdtof
R
A
R.M. Parmar et al. / Nuclear Engin
nputs that were presented in the post-test meeting and interna-ional workshop during February 04–07, 2011, the failure modef BARCOM has been identified. Participants have been asked toubmit their post-test analysis results in the prescribed format,hich will be compared with the final Phase-III test results that
re underway (Fig. 8).
. Conclusions
The present paper has identified the objectives of the Roundobin analysis program for containment safety research. The testesults from Phase-II proof test up to the design pressure (Pd)nd Phase-III over-pressure test up to 1.56 Pd have been pre-ented. Based on the discussions with the registered participantsuring the post-test meeting and the comments and feedbackeceived from the panel expert members and other invitees dur-ng their presentations in the international workshop on February, 2011, a summary document will be issued to the registeredound robin participants for submission of their post-test analysisesults. Our endeavors to demonstrate the available safety mar-in against over-pressurization of BARCOM will address importantssues with regard to containment safety under extreme events. It ismportant to note that under such extreme events the responses ofhe lined/unlined containments need to be established with confi-ence and after the functional failure such as liner tearing or majorhrough thickness cracking in the double containment system, theverall stability of the containment structure needs to be ensuredor public acceptance for nuclear power plants.
eferences
SME Boiler & Pressure Vessel Code, 2007. Rules for Construction of Nuclear PowerPlant Components, Section III, Division-1, Division-2 and Appendices.
and Design 269 (2014) 177– 183 183
BARC, 2006. Features of 1:4 Size 540 MWe PHWR Pre-Stressed Concrete Contain-ment Test Model at Tarapur and the List of Proposed Experiments – A ReportPrepared by the Committee Constituted by Director. BARC for X-PLAN Project5.06-A1 of Reactor Safety Division, BARC, Trombay, Mumbai 400085, October.
Basha, S.M., Singh, R.K., Patnaik, R., Ramanujam, S., Kushwaha, H.S., Venkat Raj, V.,2003a. Predictions of ultimate load capacity for pre-stressed concrete contain-ment vessel model with BARC finite element code ULCA. Ann. Nucl. Energy 30(March (4)), 437–471.
Basha, S.M., Singh, R.K., Patnaik, R., Ramanujam, S., Kushwaha, H.S., Venkat Raj, V.,2003b. Studies on influence of steel liner and pre-stressing tendons on ultimateload capacity of pre-stressed concrete containment test model. In: SMiRT-17,Prague, August 17–22 (Division H).
Davies, D.R., Roberts, A.C., Ness, D., 1995. A review of the design and construction ofthe Sizewell B pressurized water reactor building containment structure. NED156, 259–268.
Gupta, A., Singh, R.K., Kushwaha, H.S., Mahajan, S.C., Kakodkar, A., 1995. Assessmentof ultimate load capacity of inner containment for Indian PHWR, Division H. In:SMiRT 13, Porto Alegre, Brazil (Invited Paper).
Hessheimer, M.F., Dameron, R.A., July 2006. Containment Integrity Research at San-dia National Laboratories: An Overview, NUREG/CR-6906, SAND2006-2274P.
Liersch, G., et al.,1994. Concept study for a combined reinforced concrete contain-ment. In: Proceedings of the Third International Conference on ContainmentDesign and Operation. CNS.
1997. PCCV Round Robin Analysis Design Package, SO-97-O47. Sandia National Lab-oratories, USA.
Singh, R.K., 2007. Nonlinear analysis of RC/PC structures using phenomenologicalstress–strain plasticity, fracture and damage mechanics constitutive models. In:Panel Workshop, WH1, SMiRT-19, Toronto, (Invited Paper).
Singh, R.K., 2009. Pre-Test Report on International Round Robin Analysis of BARCContainment (BARCOM) Test Model, ISBN 97.8-81-8372-046-5 (January 28).
Singh, R.K., 2011a. Post-Test Preliminary Report on International Round Robin Anal-ysis for the Ultimate Load Capacity Assessment of BARC Containment (BARCOM)Test Model (February 07).
Singh, R.K., 2011b. Size effect and scalability issues of model tests addressed throughfracture mechanics studies. In: Division II, SMiRT-21, New Delhi, November 6–11(Paper-757).
Singh, R.K., Gupta, A., Kushwaha, H.S., Mahajan, S.C., Kakodkar, A., 1993. Ultimate
load capacity assessment of Indian PHWRS – some pre test results. In: U02/5,SMiRT 12, Stuttgart, Germany, August 15–20.
Singh, R.K., Ghosh, A.K., Kushwaha, H.S., 2009. Indian PHWR pre-stressed concretecontainment performance evaluation with BARCOM and round robin analysisprogram. In: Division V, SMiRT-20, Espoo, Finland, August 9–14 (Paper-1977).
