NUREG/IA-0064

PWR/HTWG/P(88)629 (Rev)

InternationalAgreement Report

Analysis of Semiscale TestS-LH-1 Using RELAP5/MOD2

Prepared byP. C. Hall, D. R. Bull

National Power NuclearBarnett WayBarnwood, Gloucester GIM 7RSUnited Kingdom

Office of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555

April 1992

Prepared as part ofThe Agreement on Research Participation and Technical Exchangeunder the International Thermal-Hydraulic Code Assessmentand Application Program (lCAP)

Published byU.S. Nuclear Regulatory Commission

NOTICE

This report was prepared under an international cooperativeagreement for the exchange of technical information. Neitherthe United States Government nor any agency thereof, or any oftheir employees, makes any warranty, expressed or implied, orassumes any legal liability or responsibility for any third party'suse, or the results of such use, of any information, apparatus pro-duct or process disclosed in this report, or represents that its useby such third party would not infringe privately owned rights.

Available from

Superintendent of DocumentsU.S. Government Printing Office

P.O. Box 37082Washington, D.C. 20013-7082

and

National Technical Information ServiceSpringfield, VA 22161

NUREG/IA-0064GD/PE-N/725PWR/HTWVG/P(88)629(Rev)

International.Agreement Report

Analysis of Semiscale TestS-LH-1 Using. RELAP5/MOD2

Prepared byP. C. Hall, D. R. Bull

National Power NuclearBarnett WayBarnwood, Gloucester GL4 7RSUnited Kingdom

Office of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555

April 1992

Prepared as part ofThe Agreement on Research Participation and Technical Exchangeunder the International Thermal-Hydraulic Code Assessmentand Application Program (ICAP)

Published byU.S. Nuclear Regulatory Commission

NOTICE

This report is based on work performed under the sponsorship of the

United Kingdom Atomic Energy Authority. -The information in this

report has been provided, to the USN~RC under the terms of the

International Code Assessment and Application Program (ICAP)

between the United States and the United Kingdom (Administrative

Agreement - WH 36047 between the United States Nuclear Regulatory

Commission and the United Kingdom Atomic Energy Authority Relating

to Collaboration in the Field of Modelling of Loss of Coolant

Accidents, February 1985). The United Kingdom has consented to the

publication of this report as a USNRC document in order to allow

the widest possible circulation among the reactor saf ety community.

Neither the United States Government nor the United Kingdom 'or any

agency thereof, or any of their employees, makes any warranty,

expressed or implied, or assumes any legal liability of

responsibility for any third party's use, or the results of such

use, or any information, apparatus, product or process disclosed

in this report, or represents that its use by such third party

would not infringe privately owned rights.

GD/PE-N/725PWR/IITWG/P(88)629 (Rev)

CENTRAL ELECTRICITY GENERATING BOARD

GENERATION DEVELOPMENT AND CONSTRUCTION DIVISION

PLANT ENGINEERING DEPARTMENT

NUCLEAR PLANT BRANCH

Title: Analysis of Semiscale-Test S-LH-l Using RELAP5/MOD2

Author: P. 7 (Hall and D.R. Bull, Safety Technology Section

ARTvroved by: ! IJrf Sa ety Technology Section

Summary:

The REIAP5/MOD2 code is being used by GDCD for calculating Small BreakLoss of Coolant Accidents (SBLOCA) and pressurised transient sequencesfor the Sizewell 'B' FWR. These calculations are being carried out atthe request of the Sizewell 'B' Project Management Team.

To assist in validating REI.AP5/fl0D2 for the above application, the codeis being used by GDCD to model a number of small LOCA and pressurisedfault simulation experiments carried out in various integral testfacilities. The present report describes a RELAP5/MOD2 analysis of thesmall LOCA test S-LH-l which was performed on the Semiscale Mod-2Cfacility. S-LH-l simulated a small LOCA caused by a break in the coldleg pipework of an area equal to 5% of the cold leg flow area.

RELAP5/MOD2 gave reasonably accurate predictions of system thermalhydraulic behaviour. In particular, the hydrostatic core leveldepression resulting from the hold-up of water in the steam generatortubes and pump suction legs was well predicted. A reasonable predictionof core inventory was also obtained in the period of the test in whichthe core level fell as a result of coolant boil-off.

The code did not give an accurate prediction of the liquid distributionwithin the core during the uncovering phases. Consequently the fueltemperature excursions due to uncovery were not captured by the code.Failure to calculate the correct void fraction distribution and dryoutbehaviour is believed to be due to numerical approximations inrepresenting the core by a small number of nodes, rather than due toerrors in the physical models and correlations used in the code.

iii

Based on the present study it is suggested that, in reactor analyses inwhich the potential for core uncovery occurs, the mixture leveltrajectory and peak fuel temperatures are calculated outside-RELAP5 usinga code employing a fine axial mesh, using boundary conditions from theRELAF5 analysis.

Date: February 1989

Security: This report is security classified. It may be passed toother people by recipients on the distribution list on a'need-to-know' basis and may not be passed outside therecipient's own organisation without the consent of theSafety Technology Engineer.

i V

CONTENTS

1. INTRODUCTION

2. CODE VERSION AND INPUT MODEL

3. INITIAL~ AND BOUNDARY CONDITIONS1

4. DESCRIPTION OF TEST S-LH-l 2

5. BASE CASE CALCULATION 2

6. SENSITIVITY STUDIES 4

7. DISCUSSION AND COMPARISON WITH PREVIOUS ANALYSIS 7

8. GENERAL CODE PERFORM4ANCE AND CPU TIMES 8

9. CONCLUSION 9

10. REFERENCES 10

TABLES 11

FIGURES 13

V

1. INTRODUCTION

The REIAP5/MOD2 code [1] is being used by GDCD for calculating SmallBreak Loss of Coolant Accidents (SBLOCA) and pressurised transientsequences for the Sizewell 'B' PWR. These calculations are beingcarried out at the request of the Sizewell 'B' Project ManagementTeam.

To assist in validating RELAP5/MOD2 for the above application, thecode is being used by GDCD to model a number of small LOCA andpressurised fault simulation experiments carried out in variousintegral test facilities. The present report describes aRELAP5/MOD2 analysis of the small LOCA test S-LH-l which wasperformed on the Semiscale Mod-2C facility [2, 3]. S-LH-l simulateda small LOCA caused by a break in the cold leg pipework of an areaequal to 5% of the cold leg flow area.

In addition to a detailed description of the new analysis of S-LH-l,comparisons are also described with earlier calculations withRELAP5/MODi and RELAP5/MOD2, described in reference [4].

2. CODE VERSION AND INPUT MODEL

The code version used for these calculations was RELAP5/MOD2/cycle36.05 E03. This code is a version of the standard release of INELcycle 36.05 containing error corrections (primarily Cray conversionerrors) implemented by UKAEA, Winfrith. In addition the horizontalstratification model of Ardron and Bryce [5] is optionally availablein version E03. This model was used in the present calculation tomodel the junction connecting the break line to the cold leg.

The RELAP5/MOD2 model was taken from reference [4). Dischargecoefficients in the break junction were set at 0.9 for both singleand two-phase flow. The RELAP5/MOD2 model consisted of 181 volumes,172 junctions and 256 heat structures :the noding diagram is shownin figure 1.

Microfiche listings of code input and output have been filed underSafety Technology Section in the Microfiche Archive at GDCD,Barnwood. Code input is archived ;ander BBWDQAA, at Barnwood.

3. INITIAL AND BOUNDARY CONDITIONS

As stated above, initial and boundary conditions for the presentcalculations were taken from reference [4]. To establish therequired steady state conditions, a steady state calculation wasfirst run for 300s of problem time. Parameters controlled toachieve the desired steady-state were steam and feed flow and thepump speed. A dummy time-dependent volume was connected to the topof the pressuriser to maintain the desired steady primary pressure.After 300s these steady-state conditions were removed, the dummyvolumes deleted and the calculations allowed to proceed for 50sbefore initiating the transient.

1

Figures 2-4 show the hot leg pressure and pressuriser level, theflows into and out of the intact loop SG separator; and the intactloop steam generator pressure and level during the steady-staterun. These figures illustrate that a satisfactory steady-state wasachieved. The RELAP5 calculated steady-state conditions arecompared with experimental values from reference [4] in Table 1. Itcan be seen that the steady-state conditions are satisfactorilycalculated, except for the intact loop steam generator secondaryside level, which had to be set artificially low to allow RELAP5 tocalculate stable operation of the steam generator. Use of a reducedinventory is considered acceptable, since in the test S-LH-l the SGsecondary plays only a minor role in the overall primary systemenergy removal.

The High Pressure Injection System (UPIS) was modelled via a tableof flow versus primary pressure. The secondary feed water flow wasmodelled using a table of flow versus time, with the flow beingstopped by the activation of a trip signal. The steam generatorsteam bypass valves were modelled as set to open when theappropriate steam generator exceeded '7.214Pa. The Semiscale facilitymakes extensive use of external guard heaters 'to reduce heat lossfrom the facility to the surroundings. The time variation of powersupplied to these heaters was modelled in the RELAP5 calculation.

4. DESCRIPTION OF TEST

The sequence of key events is given in Table 2. The transient isbriefly described as follows. The test was initiated by opening ablock valve, allowing primary fluid to flow through the breakorifice to the break flow condensation system. Reactor trip wassimulated by reducing core power to decay heat levels 3.4s after thepressuriser pressure fell to the set point level of l2.614Pa. Lossof offsite power coincident with reactor trip was simulated in thetest. Therefore feed water was terminated and 1451V closed uponreactor trip and the primary circulating pump also tripped. Highpressure injection was initiated 25s after reactor trip to simulatedelay in starting the emergency diesel generators.

In the S-111-1 transient break flow was initially insufficient toremove decay power. Therefore primary pressure remained above thesecondary pressure until the intact loop pump suction pipeworkcleared of liquid at 171s. A core level depression and dry outoccurred prior to pump suction clearance. After clearance a slowcore boildown began; dry out of the fuel rod simulators commenced atabout 400s and terminated at about 650s, 150s after the initiationof accumulator injection.

5. BASE CASE CALCULATION

The calculated timing of key events is compared with data in Table 2.

The measured and calculated primary system pressure histories areshown in Figure 5. The period of subcooled blowdown up toapproximately 50s is accurately calculated. The small overestimateof pressure in the period 50-250s probably arises from errors in thecalculation of secondary side behaviour. Following loop seal

2

clearance, the calculated depressurisation is too rapid, suggestingoverestimation of discharge enthalpy. The accumulator injectionpressure set point is-reached at 400s in the calculation,approximately 100s earlier than in the experiment. Measured andcalculated secondary side pressures are compared in Figure 6.Pressures are systematically overpredicted at all times afterclosure of the main steam control valves (MSIV), but the errorsappear to have only a minor influence in this test. Somesensitivity calculations were performed with an arbitarily specifiedsteam leakage in the MSIV. These indicated that discrepanciesbetween measured and calculated secondary pressures later in thistest could be accounted for by assuming a plausible leakage area ofless than 0.16% of the full MSIV area.

Figure 7 shows a comparison of measured and calculated dischargemass flow rates. Overall agreement is seen to be good, except inthe low quality discharge period of 50-175s. Critical flow rate isstrongly sensitive to upstream stagnation enthalpy under theseconditions, and the agreement between measured and calculatedresults in the period 50-175s is therefore considered acceptable.The calculated increase in discharge flow at l40s occurs because ofthe arrival of lower quality fluid in the broken loop cold leg, asthe broken loop pump suction pipework is cleared of liquid. In theperiod between loop seal clearance and accumulator injectiondischarge flow rate is underestimated by about 20% (0.O2kgs').

Figures 8, 9, 10 and 11 compare the measured and calculatedcollapsed liquid level in the intact and broken loop SG U-tubes andpump suction legs. Experimental data are derived from differentialpressures measurements and are therefore invalid until thetermination of forced loop flow at about 50s. It is also suspectedthat experimental errors are subsequently significant, since somemeasured values show a large offset late in the transient.

It can be seen in Figures 8 and 9 that RELAP5 predicts considerableliquid hold-up in the up-side of the SG tubes in the period 120-250s. However, the calculated hold-up is generally - lm of waterless than the measurements indicate, and is less prolonged. Figure10 shows that the collapsed liq'uid levels in the intact loop pumpsuction are calculated accurately, though with a time shift of about25s. Agreement between measured and calculated collapsed liquidlevels in the broken loop pump suction (Figure 11) is relativelypoor. RELAP5 predicts that this loop seal is the first to clear, at180s, followed by partial refilling as the other loop seal clears.In practice, this loop seal did not clear until - 260s, after thebroken loop SG U-tube upside had drained. The ability to predictwhich loop seal clears first is of little practical importance, andthe present result is therefore considered to be acceptable.

Figure 12 compares measured and calculated collapsed liquid levelsin the reactor vessel downcomer and core. The key features of theexperimental data are the deep core level depression which recoveredat about 170s and the gradual boildown which occurred between 300sand 530s. The first core level depression, which is caused byhydrostatic pressure due to the build up of liquid in the upsides of

3

the SG tubes and pump suction pipework, is seen to be accuratelycalculated by RELAPS. The second core level depression, caused byboil-off of water in the core, is less accurately calculated by thecode. In addition, the calculation of an over-rapiddepressurisation rate leads to the premature activation of theaccumulators, which results in the calculated boildown beingterminated too early. These errors are discussed further below.

Measured and calculated core void fractions are shown in Figures 13and 14 respectively. The experimental data of Figure 13 arereproduced from figures given in reference [4]. Figure 14 shows theaxial variation of core void fractions calculated by RELAP5.Inspection of Figure 14 shows that in the period of the hydrostaticcore uncovery, (150-200s) the calculated void fraction reaches amaximum value of approximately 0.94 in the highest core volume, and0.86 in the lowest. This is in marked contrast with the test, wherethe void fraction reaches unity in approximately the upper twothirds of the core, but remains at about 0.1 at the bottom of thecore. Thus, even though the calculated core collapsed liquid levelwas less than the measured value, the void fraction at the top ofthe core is evidently underestimated by RELAP5/MOD2.

To compare the measured and predicted core axial void fractions inthe core boildown phase (350-500s), we examine condition at t -400s, immediately prior to the calculated time of accumulatorinjection. At this time the measured and calculated collapsedliquid levels are approximately ~the same. It is seen by comparingthe figures that at t - 400s, a well-defined two-phase mixture levelexisted at the 250 cm. elevation in the test (see Figure 13b).However, in the REI.AP5 calculation a near homogeneous two-phasemixture is calculated to exist in the top part of the core.Therefore, although RELAP5 adequately calculated core water/inventory at this time, the detailed axial distribution in the corewas in error.

Figure 15 compares measured and calculated heater rod temperaturesat around the 250cm elevation (the discrepancy prior to trip arisesbecause the REI.AP5 value is for the heater rod surface, whereas theexperimental value is measured wit *h an embedded thermocouple.Discrepancies due to the thermocouple location become insignificantafter trip). It is seen that REIAP5 fails to calculate either ofthe experimentally observed dryouts. These discrepancies are dueprimarily to the errors in the calculated core void fractiondistribution described above, and are considered further in the nextsection.

6. SENSITIVITY STUDIES

The primary shortcomings of the calculation described above areinaccuracies in the calculated pressure history, core void fractiondistribution and the failure to calculate the core dryouts.Numerous sensitivity studies were carried out to investigate thecause of these errors. Results are described in this section.

4

6.1 6.1 Modelling of Hydrostatic Core Uncovery

Figure 12 illustrates that the hydrostatic core leveldepression occurring in the period 170s was wellcalculated by RELAP5. However, although the corecollapsed liquid level was actually underpredicted,dryout was not calculated because RELAP5 predictedsignificant liquid fractions at the top of core.

Detailed investigation of calculated void fraction andflows in the core, upper plenum and hot legs, indicatedthat during this period of the hydrostatic leveldepression, a small counterflow of water was predictedto be refluxing from the hot legs into the upper plenumand core. In the core volumes this downflow of watermaintained the void fraction below about 0.95; this wassufficient to prevent dryout according to the CHFcorrelation applied by RELAP5/MOD2 for these condition(modified Zuber 'correlation). In practice such adownflow would probably run down the wall of the corevessel and would be unlikely to prevent dryout in therod bundle centre region. Loomis and Streit [4) indeednoted that in the test, whilst some pins dried out,others at the same elevation did not, indicating thatfalling films of water were present on some rods abovethe mixture level. This is consistent with the viewthat a spatially non-uniform liquid distribution existsin the core region. To try to model such liquiddistribution effects in RELAP5, the core model wasmodified as illustrated in Figure 16. The core wassplit into two channels. One channel represented theouter subchannel, next to the vessel wall. (The vesselheat structures were connected to this channel.) Theother channel represented the remaining centralsubchannels (all heater rod heat structures wereconnected to this channel). In initial studiescrossflow junctions were included between parallel corechannels, but recirculating flow patterns arose whichinhibited the calculation of dry out. The crossflowjunctions were therefore deleted. Finally to ensurethat no refluxing water entered the central core zone,the core outlet junction (volume 146 to 161) was definedas a homogeneous ('one velocity') junction.

Figure 17 compares experimental collapsed liquid levelsin the core vessel and downcomer for the basecase andthe sensitivity calculation. It is seen that the effectof the core noding change is to produce a more prolongedcore depression, with earlier clearance of the pumpsuction. In spite of the increased core liquidinventory in the hydrostatic core uncovery phase (t -140s), the void fraction in the upper part of the coreis now calculated to reach unity (see Figure 18).Subsequently dryout is now calculated to occur asillustrated in Figure 19. The rate of rise oftemperatures is well calculated, but because thecalculated dryout is too prolonged, the maximum heaterrod temperature is over-predicted.

5

It is concluded that under conditions of core leveldepression in which reflux condensation is taking place,RELAP5/MOD2 may fail to predict a dryout. Care musttherefore be taken in using the code to predict peakfuel clad temperatures in these conditions.

6.2 Modelling of Core Boildown Phase

As noted above, RELAP5/MOD2 also failed to calculatedryout during the boildown phase of this transient (350-500s). This appears to be due to errors in thecalculated core void distribution, and also to theprediction of early accumulator injection.

Figure 17 shows the collapsed liquid level in thedowncomer obtained in the base case' and sensitivitycalculations. The sensitivity calculation is seen to bemore accurate than the base case calculation.

The core void fraction distribution in *the sensitivitycalculation is shown in Figure 18. Again these resultsare in much better agreement with the test data (Fig13b) than results for the base case calculation.However the improvement seems mainly due to the use ofthe one-velocity junction at the core outlet, whichincreases the calculated liquid carry-out from the topof the core. The agreement with test data is thereforeregarded as somewhat fortuitous.

In spite of the improved agreement with the corecollapsed liquid level and void distribution RELAP5/YOD2still fails to predict a core dry-out during theboildown phase, as can be seen from Figure 19.Modifications to the level sharpener model in the codewould probably be necessary to successfully calculatethe core dry-out, given the coarse axial mesh used torepresent the core in the present analysis.

6.3 Calculation of Depressurisation after Loop Seal Clearance

It is seen from Figure 5 that the depressurisation rateafter intact loop seal clearance is over-estimated byRELAP5/MOD2. Consequently the onset of accumulatorinjection which terminated the core boil-down phase ispredicted to be too early. Calculations were performedto determine if errors in the calculations of break flowand enthalpy were responsible for the errors in thedepressurisation rate.

To investigate the effect of changing the discharge flowmultiplier (CD2), a calculation was performed with thevalue arbitrarily set to 0.65, (as opposed to 0.9 in thebase case). The accuracy of the calculated pressurehistory was greatly improved, but the calculated corewater inventory was significantly overestimated in the

6

boildown period. It was therefore concluded that theerrors in calculated depressurisation rate cannot beascribed simply to uncertainties in the break flowmultiplier.

A check was then made of the break mass flow ratecalculated by REIAP5/MOD2 in the period 250 to 450s.Values were found to be within 10% of prediction of thehomogenous thermal equilibrium model of two-phasecritical flow, which would be expected to provide areasonable prediction for the orifice geometry andthermodynamic conditions in the test. Therefore theRELAP5/MOD2 results are considered reasonable.

Finally a check was made on the calculated enthalpy inthe broken loop cold leg in the period 250-400s. RELAPSresults indicate void fractions in the range 0.98-1.0 inthis period.

Experimental data based on gamma densitometermeasurements indicate high void fraction at the vesselend of the broken loop cold leg. However, the bottomdensitometer beam at the pump discharge end of thebroken cold leg reveals the persistence of a layer ofwater throughout the transient. This suggests that theerror in calculated depressurisation rate results fromerrors in the calculated rate of entrainment of waterfrom the broken loop pump suction pipework into the coldleg, and break nozzle.

Additional information suggesting that the error incalculated depressurisation rate is related to errors inthe calculation of broken loop cold leg conditions comesfrom detailed inspection of the primary pressurehistory. The experimental pressure history shows twodistinct knees at - 180 and 270s as the broken andintact loop pump suctions clear. In contrast, the-calculation shows only a single knee as the broken looppump suction clears.

It was concluded that relatively small errors in thecalculated clearing behaviour of the loops sealsprobably gave rise to the overprediction of thedepressurisation rate. However, the data isinsufficiently detailed to eliminate errors incalculated primary-to-secondary heat transfer, ormetalwork heat transfer as contributors to the errors incalculated pressure [4).

7. DISCUSSION AND COMPARISON WITH PREVIOUS ANALYSIS

The present test has been analysed by Loomis and Streit 141 usingREIAP5/MOD2/Cycle 36.05. In common with the present study, theseauthors noted the failure of REIAP5/MOD2 to calculate dryoutcorrectly; they also noted that calculated primary pressure felltoo quickly after loop seal clearance.

7

Loomis and Streit 14] found that by modifying the interphase dragcorrelations in RELAP5/MOD2 to give a 90% reduction of interphase.drag in the core volumes, a reasonable prediction of the liquidlevel trajectory in the core could be achieved. The presentcalculations have shown that an accurate prediction of the coreaxial void fraction prof-Ile can also be achieved using theunmodified interphase drag models, providing adjustments are made tothe modelling of slip in the core outlet junction. This suggeststhat errors in the void distributions are more likely to be due toapproximations used in numerical implementation of the interphasedrag correlation, than to errors in the correlations themselves. Itis believed that averaging procedures used to calculate interphasedrag in junctions, treatment of inverted void profile and the levelsharpener model, all contribute to errors in the core voiddistribution found in the present base case calculation.

Loomis and Streit also recommended that the critical heat flux (CEF)modelling in RELAP5 be modified to allow calculation of dryout inthe hydrostatic uncovering phase. They suggested that the factor(1 - a 9g) appearing in the modified Zuber CHF model should bereplaced by (0.94 - a ) to ensure that dryout is calculated when thecore void fraction exceeds 0.94.

The present studies suggest that failure to calculate dryout in theS-LH-l hydrostatic uncovering phase was not in fact due to CEFmodelling but was rather due to incorrect calculation of the liquiddistribution in the core under conditions of reflux condensation.

It is concluded that considerable care must be taken in modellingthe liquid level trajectory and core heat up behaviour inRELAP5/MOD2 simulations in which the core is represented by a smallnumber of nodes. It is recommended that in reactor fault analysesin which the potential for core uncovering occurs, the calculationof the level trajectory and the peak fuel clad temperatures beperformed with a separate code using a fine axial mesh, takingthermal-hydraulic boundary conditions (core liquid inventory,pressure, inlet enthalpy, core power) from RELAP5/MOD2.

8. GENERAL CODE PERFORMANCE AND CPU TIMES

The present calculations were performed on the Cray-2 computer atAERE, Harwell. 2270s of CPU time was used, giving a CPU:real timeratio of 2.84:1. The CPU time per timestep per mesh cell wasl.32ms. The main limitation to timestep size was the courant limitin the broken loop pump suction pipework. The code was found to bestable and easy to use. However, it was found necessary to reducethe maximum timestep manually during the first 25s to avoidcalculation of premature dryout and subsequent code failure.

Using the split core input model, the CPU: real time ratio reducedto 1.75:1, and the CPU time per mesh cell per timestep was l.2Oms.

8

9. CONCLUSIONS

1. RELAP5/MOD2 cycle 36.05 Version E03 has been used toanalyse test S-LH-l (5% cold leg break loss of coolantaccident simulation) carried out in the Semiscale PWRtest facility.

2. The code gave reasonably accurate predictions of systemthermal hydraulic behaviour. In particular, thehydrostatic core level depression resulting from thehold-up of water in the steam generator tubes and pumpsuction legs was well predicted. A reasonableprediction of core inventory was also obtained in theperiod of the test in which the core level fell as aresult of coolant boil-off.

3. RELAP5/MOD2 did not give an accurate prediction of theliquid distribution within-the core -during theuncovering phases. Consequently the fuel temperatureexcursions due to uncovery were not captured by thecode. Failure to calculate the correct void fractiondistribution and dryout behaviour is believed to be dueto numerical approximations in the implementation ofinterphase drag modelling and in representing the coreby a small number of nodes, rather than due to errors inthe physical models and correlations used in the code.

4. Based on the present study it is suggested that inreactor analyses in which the potential for coreuncovering occurs, the mixture level trajectory and peakfuel temperatures are calculated outside RELAP5 using acode employing a fine axial mesh, using boundaryconditions from the RELAP5 analysis.

9

10. REFERENCES

1.

2.

NUREC/CR-43l2

EGG-SEMI-6813

RELAP5/MOD2 Code ManualVolumes 1 and 2V.H. Ransom et al

Experimental OperatingSpecification forSemiscale Mod-2C 5%Small Break LOCAExperiment S-LH-lG.G. Loomis

Quick Lo ok Report forSemiscale Mod-2CExperiments S-LH-1and S-LH-2G.G. Loomis, J.E. Streit

Result of SemiscaleMod-2.C Small Break(5%) LOCA ExperimentsS-LH-l, S-LH-2G.G. Loomis, J.E. Streit

Dec. 1985

Feb. 1985

3. EGG-SEMI-6884

May 1985

4. NUREG/CR-4438

Nov. 1985

5. AEEW-R2345 Assessment of HorizontalStratificationEntrainment Model inRELAP5/MOD2K.H. Ardron andW.M. Bryce 1988

6. ECG-SEMI-5622 Post Test RELAPSSimulation of theSemiscale S-UT SeriesExperimentsM.T. Leonard Oct. 1981

10

TABLE 1

INITIAL CONDITIONS FOR S-ILl-i

EXPERIMENT

Pressuriser pressure

Core power

Core AT

Pressuriser liquid level(collapsed above bottom)

Cold leg fluid temperaturesIntact loopBroken loop

Primary flow rateIntact loopBroken loop

Initial Bypass flow(% of total core flow)

MPa

kW

K

cm

15.47 ± 0.14

2014.75 + 0.15

37.65 +1.5

-0.6

395 ± 14

RELAP5

15.47

2014.75

37.30

396.1

560.9565.68

K562.12 ± 2564.05 + 2

kgs-1

7.132.35 2.35

0.920.9

Leak rate kgs-1 0.0002 0.0

SG secondary

SG secondary

pressureIntact loopBroken loop

side massIntact loopBroken loop

MPa5.72 ± 0.076.08 + 0.07

5.86.09

kg191 ± 1343 + 4.3

150. 0*43.0

* Approaching limit of stable operation of steam generator by RELAPS

11

TABLE 2

TIMING OF EVENTS FOR S-LHi-l

EVENT

Small break valve opened

Pressuriser pressurereaches trip level (l2.6MPa)

Pump coast-down initiatedIntact loopBroken loop

TIME AFTER BREAK OPENS (s)EXPERIMENT RELAP5

(BASE CALCULATION)

0.0 0.0

HPIS initiatedIntact loopBroken loop

Pressurizer empty

Minimum core collapsedliquid level

Pump suction clearingIntact loopBroken loop

Second core dryout

Accumulator inj ectionIntact loopBroken loop

14.67

21.3520.76

41.60.40.98

33.9

172.6

171.4262.3

412.8

503.8501.4

16.95

21.721.35

43.6843.28

55

182

203180

406.0407.0

12

Semniscale S-L1-1-1-

W*OK(EN LOODP

IC7. DAIW¶ OR? IAM70$AT ýLAOS

ICAIEPS

Figure I1: RELAP5 nodinq diagram tar Semniscale S-1-1--1 Calculation ( Base case)

VISION REFERENCE FILE STDY-STATE2S-Lt11l :1Semiscale S-LU-i Steady slate 2C

16.8-I I

15.2-

15.6-

r

L

LC*

1S.4-

..............I............. ............................................................. .................. .............-..

----- - - - - -L-u

.. .... ................... .... .......... ...... ..........-. ........... ... ............

.. .. ... .............................. ............ ................ .... ...

.. .. ............... ................ .... .... ................ ...... ........I. ..

..... .................................... .................. ............................Press ....ure..

... .. .... ... .. ... .. ... .... .. ... .. ... .. ... .... .. ... .. ... .. ... .. ........ .. ... .. ... .. ... .. ... .. ... .. ... .. ... .. ... .. ... .. ... ..

40

rI',C

311

3G

4Ž.

34

1s.O-

32

1 4.2-

30so5 108 ISO ?130 2S0

P28101 [I] - CI11RI.VHIRG (2] -

Figure 2 : Hot, leg pressure and pressurizer level

30u

UJSJTrn PEFEPEiCE FILE SMhY-STflTE2:S-L If-1: 1

Semiscale S-Lit-i Steady staLe

4-

-. 4 - - .- - -- -

.. ............

I ...........................

1%

my

:30

meme

................ ............ ........................................................... ........

4~ ~~~~~Rst ----- low ----

......... ..... ..... ..... ... .... ... . . ... ... .. .. .... .. ... ....... .. .... .

S~o~Fd e..

.. . ... .. . .. . . .. .... ... .. ... . .... ....... . .. . ... . .. .. .... .

-4

-3

U,

11- -1

B-

- I--1

e 513 189 159 0 5

MrOU.16018 - MLOU.G16102 *-- t-fLOWGBI183 --

Figure 3 :Intact. loop steam generator separator mass Hlow rates

313

VISION PRENTPEtC FILE STDY -STA~TE2: S-L H- I: ISemi5cale S-LH-1 :Steady State12

(I)

6. 6-1....

... ... ... .....

........-..........................................................

L Eig -

I P.. .... ... .... ...

K--1 ............... ................. . ......................... ...... .................... j...- . .............. .......

LIL

..............

5.6r-]1.... ................... .................................

......... ............... ..........................

........................ ...........................

.......................... ..........................

.......................... ..........................

.......................... ..........................

.......................... ..........................

.......................... ..........................

..........................

-0.10

r0CS

:3

-0.06

-0.04

S. 9-1

ADO IOU C COUt 1DOj 00

PI1101 III C"iTRLI.%W43G (2)

Figure 4I : Intact loop steam generutor pressure and collapsed liquid leuel

I

I

00~C

SL

aftaSI.A.

Semiscale S-LU1-i

6 . .3..................................

1 4 ... ... ... .. .... . ... ....... . .. ... ... ... ... ... . .. ... .... . .. ... .. ..... .... . .... .. .. ... ... ..

.. . . . . .. . . . . . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .1

I a . . . . . . . . .. . . . . . .. . . . . .. . . . .. . . . .. . . . . .. . . . . .. . . . . ..I. . . . .. . . . . .. . . . . .. . . . . .. . . . . .

...... ..... ........... ..... ...................................................... ...... 1

..... . . . . . . . . . . . . . . . . . . .. . . . . ..* . . . . . . . . . ..1. . . . .. . . . . . . . . . ... . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .

......8..... ........... ................. ....................... ....................t

................................. ...... ................... ............

............. ...... ......

.. ........ ........ ...... .. ... .... . ..... ............ ... . .....>. ........ ........ . .....

.................. . ...... ...... .................. .....................

N ... ... .

-1 -2....... ... ... ...... .. .... .... .. ...... ..... .,.... .. .. . .. .. ... .

..................................... .. . . . .. . . .

a4 0

I, too, 206 300 400 SOO NO 780 800

lifne ( S)

P28181 .- P1:1

Figure 5 : Measured and calculated primary pressure ( Base case)

Semiscale S-LU-i--

a.E*

Co

7 ............ ................. . ......... .................................

...................

*4. ....

3rknlo

2 -e ....... ..... ...................

1- ...... ..... .. ... .. .. .. .... ... .. ... ..... ... .. ... . .. ....

1 1 1 r T I I/

0 180 208...8........CO. 78..8.

litre ( S)

P70SOI.......... PBG41lSG - P6I119 Pl~IS'11 __

Figure G : Measured and calculated secondary pressures ( Base case )

Semiscale S-LH-I

......I .......... ..... .............. ........ ...... . .......... .. ... ... ....... ........ ... .... ..

...... ...... ...... .......

.. . . . . . . .. . .. ..*.. . ... . . ... ... .. .. .. .. .... .. . .. . .. . . . . . . . . . . . . . .. .. . . . . . .. ... .. . . . . .. . . . . .. . . .

8.8 ... ... . ...... .. .... ...... ............ ....... ........ .0 .9

I ,,.Calctaihtec

... . .... ............ ...... ............... f .... .... ...... ........

~? 8 . 4 -. j...

.... . ............. .. . ... .. .. .. ..... . ... .... ... .... .. ... ... .... ... .

8.4- ...... ........... ..... ...... ......... .. ........... ......... ... ...........

8 100 200 300 400 Sao GOO 700 800

T in tf s

ITLOUJ37590 . - - IIOTI8REAK __

F igure 7 : Mleasured and calIculIaLed breakf Claw ( Base case)

Semiscale S-1.11-1

........ ...... ....... . ' ......

...... ...... . ........ .... .. . ...... .......... ...

...... . ... ..... .... ..... ..... .. ... ..... ..... ..... ..... .. - e

N))

..... ...... ........... ......... e

S:MP 6OW* .E............. ........ ............ ....

eee * C>?V(ERItitIT uWsL~ * *4.

*%.- EXPER1ftNT Dc*IS IDE

~~.. .....-......L2.......-...-. ....

a lee 280 308 488 see GOO 708 Goo

CMMrR?.~ 11P49?1 57E CriWLVARS .. LIP#97t 57X

Figure 8. Collapsed Liquid Levels In Intact Loop U-tubes

SEMlISCALE S-Lit-i

am

E

0b

6ws

4@e-

....... . . ..... .... .......

..... ..

~-RRLAPS UP41DE

RELR OUS

.... ... .... u.. ... ..

.... 4........ ...... ...................... ,..... .......

...... . .......... .... .......... .......... .... .

............. ................... ...t.......... ... ..

................. . . ...........I. . ...... .....

............... ......

N)

-40e

Ta 188 200

CIITRLVAII - LOPSURS3 -SU

F Igure 9.

II 1 1306 400 sea GOO 7

-. CHTRAIRRl .. ... LOPSCR53-SUx

Collapsed Liquid Levels In Broken Loop U-Aube

86

Time ( s

SOD-

-300-

200-

i*

100-

Semiscale S-UL-i

DOak" flusd

...... ...... ..... .... ....

.. . .... ......... ..................... ....... ................ ...........

..... * . ... .... .... ... .. ... .. ... ...... . .... .

" Calculated lijcosure'd

a cmt..s bottc~ aoI suqt. Ion

-4110

.308

-108

8 108 200 308 400 SOO 6coo 780

CIIIRLIklR9 - DPI-S7Xtl4B ____ IIRLUA~l9 DPIr14RtIG

F igure 1O: Mleasured and calIculoaLed IL puImp suct ion col lapsed l iqu id l~evel

BOB

tf#-Ie

Semiscale S-LH-1-

U

A ..........

.. .. . ... .. ... .. ..... ... .

.. .. . .. ... ...... .....

.... ...... ..

.1ll f. .. . .. . . . . . .. . . .. . . . .

...... .... ... .. . 4

8 cmtIs bo~ttf a I sutw ion

.. . . .... ... ... . . . .. .. . .. . .. . . . ... .. . .

.g ~ ................ 4...... ...........

.. .. . . .. . . . . .......... .... . .. . . .. .I.. . .. . . . . . .

d-

-200

-1200

-108

Cs0I I II -______ -- I

ChT1RLM~R13 .- DPB-S7XGSA - CNTRLI.R1R4 . D P~tGSt73___

Figure H: Measured and calculaLed BL puIMp sucLion collapsed liquid level

"03

Semiscale S-LH-1

...... ........... .... . .......

.- ...... .......... I ....... ****a

* -10oukcouer Itqu~cd let-ell:

.... ..... . ..... ..

....a- . . .. . 4 .......... ...... . .............. .... I .... . ... . . . .. ... ..... .... .... ... ... . . . .

-300-- .. ........... .4 3...80.. ......... ....

Pess 1: 11 400d; ee1

.......... ..... ...... . . . . . .. . ... .. . .. . . ... . . .

.................. ......... ............ .... ........... .......... ... .... .......... .......... ... I........ ...................... ...

-49G6mi bdttom Whete Core : : : :....5-280 jm I bdttom 01pu.-p suCtion: :: :

* 138cw....... o~ . ... ...... ........... ... . I........ ...... ....... ......

108 200 300 48 00 So oe 788 808

(tITRLI%1R3 - LI.0#29-5?9 -__ CtITRLVARI . -LV-13M-S72

Figure 12 : Vessel and dow~ncomer collapsed liquid levels (Base case)

Semiscale S-LH-1

C0

0

*1.1

1.0

0.9

0.8

0.7

0.8

0.5

0.4

0.3

0.2

0.1

0.0100 125 150 175 200 2.25

Time after rupture (s)250

W~SMIB-~

Figure 13 a :Measured vessel axial void fraction distribution during the manometric depression

C

1.0

0.9

0.6

0.6

0.4 TAAll 6ele

0.3 bottom

0.2

0.1

0.0300 350

Accumulator Injection

400 450 500 550lime after rupture (a)

(500 6150 700

w8aeine-4

Figure 13b : Measured vessel axial void traction distribution during core boildown

25

Semiscale S-L14-1VISION PEFEREIICE FILE STD-ICCCK3:S-LHI1-:2

1.0

0.8

B.G

-)0.4

o.2~

0.0

.s*.......... ....................................... ........... .......... ...................... ....................... ......................

...... I ............... ..... ................4........... ......................... I.......................... ................ ~. ................................... ........... .......... ...,. . . . . . .. .....

I~lE Cleval ao bove-.4 hottomisbi Care~ .heated ilength m~r

..... . ....... . ..... ......... . .

4 .......... ..... ..... .................. .. .............. .............

. . . ... .. .. .. . .. .... . .....

'1. .... . ............ .... ...

..I ... ... ...... ...... .....

I,-.............

V0ED01 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ....... ...IOSO .. .WDl84 OD10~....... Dl~0

FiueI aclld oeui rcin

000

B0

a

Semiscale S-LU-i

760-

GSB-

o 00-

.. . . . . .. . . . . . . . . . . . . . . . . ... . .. . . . . .. . .. . ... ... ..

..... .~. ................ .......

..... ...... .

.. .. .. ... ... ..... ..

flenur eA

:Catculcted ..2S cmi Iewt'nIor

~......................... ................. ........................~~ ~ ~ ~ ~ .. . .. . .. ... . . .. . . . . . . . . . . . . . . .I . . .. . . . . . . . .. . . . . . . . . . . .

....... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...

8100 206 300 400 Soo Soo 70080

-700

-GSO

-600

-s500

liho ( S

wrvtVSeoo9ts .- Df'JsAS.S1-

Figure [5:Measured and calculaLed heaLer rod LemperaLtires (Bose case)

ONE VELOCITYJUNCTION

CORE INNERZN,CONTAINING

HEATER RODS

CORE OUTER ZONE,CONTAINING NOHEATER RODS

FIG.16 RENODALISATION OFIHE CORE

28

Semiscale S-LU-i

8A..

EU

CL

U

S~( ItIvI

.. . -......

4Studg:

... .. ... .. .. ... . ..

....... ........ ......

.. ................. .... ..........

... .. . . ... .....

.... .... . .. .... .. . ... ... . . . . . . . . .

.. .. ... .. .... ..... .. . F ... . . .. . . .. .. . . .. .

.......

..... ........

.............

..........

.............

..........

...........

-I 82

-o

--100

--300

-Sw-i.. --Soo

hI~1fl" .

e 109 200 300 400 Soo GOO 700 8600 -Tinae ( S)

CI1TRttAR3 .-- CIITRLMXR3 ......... LLO29-572 -__ Cf1TIZI-VR191 . -- CflTRLLARI ......... LV-13fl-578 -

Figure 17 :Compaorison of vessel and douncomer collapsed liquid levels

Serniscale S-LH-1VISIOM REFERENCE FILE COREV4.2:S-Ul-l:?

0.Cci

ii

CA.)0

-~

. ................ ....................... ...... ...... .......................... ............

* !~ *ii I ** ~ .J330 cma.. .. .. . .. . . . . . . . . . . . . ... . ... . . .

iIi-

278270

I 210 C"............ ~ ~ ~ . ...... . ..................................

I.Ej ki'I ~)soIs Ifj'*** "*1~ . hit~

05' ~.~~~~ ~~~~ ~ ~ ~ .. .... ... T j. ....... ........ ........ .. ..... ......

... .~ ..... 0 A

90 C1

.. .... ..... ... .... ..... ..... .... ..... .............. ... ... ............ ..

o 08 ~ 0 30 8050 GB 00M0

VOIDGIIOI .. WIDGI4Oi W1014301 AIIDGII~l 30I(40 c-4GIG

Fiqure ~~~ ~ ~ ~ ...... ..C.cl.e.or.oi....o (Snstviystd

Semiscale S-LH-1

I Cn~tult~ed I..

P 1~ cm Ievatinp *

Al .

...... ..... ............ ..................................... ...... ............

I a ..a

............ .. ............. ............. ..... ...................................

IxI

.... ... ..... ....... ......

.. . . ...

....... I ............ ....... ...... ......

-~ ITrB NO 4*. OBco.7..O

I~~~Tr s 21c~e.~~

Figure. 58Maue n acl~d e~rrdLmeaue

Distribution (S - Summary only')

S5

P.M.B. V.I. L.P. D.K.H.P. C.F.M.

BillamGeorgeHirstJenkinsArdronHall (6)Evans

GDCDGDCDGDCDGDCDGDCDGDCDGDCD

PMTPMT

HSDHSDHSD

S R. GarnseyN.E. Buttery

S J. R.P. R.A. C.

HarrisonFarmerWilletts

S E.W. Carpenter BNLA.J. Clare BNL

M.W.E. Coney CERLA.H. Scriven CERLJ. Putney CERL

L.F. Wilson CISD Park Street

S D.A. Ward NNCK.T. Routledge NNCP.A.W. Bratby NNC

S D.A. HowlK.W. Hesketh

S M.R. HaynsI. BrittainI.H. GibsonC.G. RichardsW.M. Bryce

Library G)Library BLibrary CLibraryLibrary S

BNFL SprinfieldsBNFL Springfields

UKAEAUKA.EAUKAEAUKAEAUKAEA

WinfrithWinfrithWinfrithWinfrithWinfrith

DCD

ERLERL

udbury House

32

NRC FORM 335 U.S. NUCLEAR REGULATORY COMMISSION 1. REPORT NUMBER(2-89) lAssiomed by NRIC. Add Val., Supp., Reov..NRCM 1102. DATA SHEET Nmer, ton3201.3202 BIBLIOGRAPHIC D T HE

(See instructions on the reverse) NUREG/IA-00642. TITLE AND SUBTITLE DP-/2

Analysis of Semiscale Test S-LH-1 Using RELAP5/MOD2 PWR/HTWG/P(88)629(Rev)

3. DATE REPORT PUBLISHEDMONTH j YEAR

AprilI 19924. FIN OR GRANT NUMBER

____________________________________________________ A46825. AUTHOR (S) 6. TYPE OF REPORT

P.C. Hall, D.R. Bull

7. PERIOD COVER ED linciusis'e Dates)

B. PERFORMING ORGAN IZATION - NAME AND ADD RESS (IiNRC.proride Division, Ofisce or Region, U.S. Nuclear Regulatory Com~mission, and mailing address,, if contractor, procidenamne end mailing addtessi

National Power NuclearBarnett WayBarnwood, Gloucester GL4 7RSUnited Kingdom

9. SPO NSO R ING OR GAN IZATI ON - NAM E AN D ADD RESS (if NRC. type "Sm as abW' if contractor. provide NRC Division. offisce or Region, U.S. Nuclear Regulatory Commission,and mailing address.)

Office of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555

10. SUPPLEMENTARY NOTES

11. ABSTRACT 1200 words or les)

The RELAP5/MOD2 code is being used by GOCO for calculating Small Break Loss of Coolant Accidents (SBLOCA) andpressurized transient sequences for the Sizewell 'B' PWR. These calculations are being carried out at the request ofSizewell 'B' Project Management Team. To assist in validating RELAP5/MOD2 for the above application. the code is beingused by GDCD to model a number of small LOCA and pressurized fault simulation experiments carried out in various integraltest facilities. The present report describes a RELAP5/MOD2 analysis of the small LOCA test S-LH-1 which was performedon the Semiscale Mod-2C facility. S-LH-1 simulated a small LOCA caused by a break in the cold leg pipework of an areaequal to 5% of the cold leg flow area.

12. KEY WORDS/DESCR!PTORS (List words or phroses that wilt assist rmsearcher: in locating the report.) 13. AVAILABILITY STATEMENT

semiscale test iinl imitpsdTest S-LH-1 14. SECURITY CLASSIFICATIONRELAP5/MOD2 ( This Page)

cold leg pipework iinCl ac~ifiprfSmall Break Loss-of-Coolant Accidents (SBLOCA) (This Report)

15. NUMBER OF PAGES

16. PRICE

NRIC FORM 335 f2-89)

THIS DOCUMENT WAS PRINTED USING RECYCLED PAPER

NUREG/IA-OM6 ANALYSIS OF SEMISCALE TEST S-LH-1 USING RELAPSIMOD2 APRIL 1992

UNITED STATESNUCLEAR REGULATORY COMMISSION

WASHINGTON, D.C. 20555

OFFICIAL BUSINESS-PENALTY FOR PRIVATE USE, $300

FIRST CLASS MAILPOSTAGE AND FEES PAID

USNRCPERMIT NO. G-67

RES4 RZAICt

Dc 05