NUREG-0635, Forwards SER Re TMI Item II.K.3.30 for C-E Plants,Per Conditional Acceptance of CEN-203(P),Rev 1, Response to NRC Action Plan Item II.K.3.30 Justification of Small Break LOCA Methods

Forwards SER Re TMI Item II.K.3.30 for C-E Plants,Per Conditional Acceptance of CEN-203(P),Rev 1, Response to NRC Action Plan Item II.K.3.30 Justification of Small Break LOCA Methods

ML20127K401
Person / Time
Issue date:	06/20/1985
From:	Thomas C Office of Nuclear Reactor Regulation
To:	Wells R C-E OPERATING PLANTS OWNERS GROUP, NORTHEAST UTILITIES
References
RTR-NUREG-0390, RTR-NUREG-0635, RTR-NUREG-0737, RTR-NUREG-390, RTR-NUREG-635, RTR-NUREG-737, TASK-2.K.3.30, TASK-TM NUDOCS 8506270411
Download: ML20127K401 (16)

Text

,,

• 4kmar(%,

8 ,i UNITED STATES NUCLEAR REGULATORY COMMISSION tc q WASHINGTON, D. C. 20555

% ,,,, /a '*

f, JUN 2 0 7985 '

i Mr. R. W. Wells, Chairman CE Owners Group Northeast Utilities' .

Post Office Box 270 f Hartford, Connecticut 06141-0270

SUBJECT:

CONDITIONAL ACCEPTANCE FOR REFERENCING OF LICENSING TOPI CEN-203(P) PEV.1, RESPONSE TO NRC ACTION PLAN ITEM II.K.3.30 s JUSTIFICATION OF SMALL BREAK LOCA METHODS" g

Veone of have completed item.

confirmatory our review of the subfect topical report with the exception March 31,-1982. This report was submitted initially by letter dated letter. We find Revision 1 was submitted April 15, 1982 by CE Owner's Group the report to be acceptable for referencing in license applications to the extent specified and under the limitations delineated in the report and the associated NRC evaluation, which is enclosed. The evaluation defines follows: the basis for, acceptance of the report. The specific limitation is as Our review fir.ds the submittal acceptable pending a confirmatory benchmark analysis to demonstrate good agreement between CEFLASH-4AS and the data from Semiscale' test S-UT-08.

In a letter from R. W. Wells (CEOG Chairman) to C. O. Thomas (NRC), the CEOG tommitted to submit results of the benchmark analysis by December 31, 1985. We find this commitment acceptable. The staff will issue a supplerrent to the SER following receipt and evaluation of the benchmark submittal..

On the condition .that' we find acceptable the comparison of CEFLASH-4AS results L with the S-UT-08 exp'erimental data, licensees and applicants with CE NSSS designs will not be required to reanalyze the SBLOCA for their plants (as required by Action Item II.K.3.31) since their present model was demonstrated 4

to be in compliance with Appendix K. We do not intend to repeat our review of 7, the matters s

-described in the report and found acceptable when the report appear' las a reference in license applications, except to assure that the material presented is 'aoplicable to the specific plant involved. Our acceptance applies only to the matters described in the report.

In accordance with procedures established in NUREG-0390, it is requested that CEOG publish accepted versions of this report, proprietary and non-proprietary, within three months of receipt of acceptance of our review of the benchmark analysis to be submitted by CE on December 31, 1985. The accepted versions shall incorporate this letter and the enclosed evaluation between the title page and t h abstr3ct.

accepted) followiry the report identification symbol.The accepted versions -

Should our criteria or regulations change such that our conclusions as to the acceptability of the report are invalidated, CEOG and/or the applicants referencing the topical report will be expected to revise and resubmit their /

l 8506 270411 _850620 l

C PDR e t

Mr..P,1c W. Wells respective documentation, or submit justification for the continued effective applicability of the topical report without revision of ineir respective documentation.

Sincerely, C ' o O C:==

Cecil 0. Thomas, Chief Standardization and Special Projects Branch Division of Licensing

Enclosure:

As stated D'

i 6

k k

.u,,,-

i,.

Juna 20, 1985' Mr.'R. W. Wells -

2-

. ,9 1

respective documentation, or submit justification for the continued effective

. applicability of the topical report without revision of their respective documentation.

Sincerely, j Original signed by Cecil 0. Thomas, Chief Standardization and Special Projects Branch Division of Licensing-s

Enclosure:

As stated r

i

/

4 DISTRIBUTION CentraLJ11el SSPB Reading PDR DCS PNoonan DMoran HBerkow i- RDiggs ,

CThomas SSPB:DL <

y PNoona .dk <n w M 6/ 5 /85 /gp/85 M 85 L - . . . . - _ _ _ _ _ - - - _ - - - _

SAFETY EVALUATION REPORT TMI ACTION ITEM II.K.3.30 FOR COMBUSTION ENGINEERING PLANTS I. BACKGROUND i -

NUREG-0737 is a report transmitted by a letter from D. G. Eisenhut, Director of the Division of Licensing, NRR, to licensees of operating power reactors and applicants for operating reactor licenses forwarding TMI Action Plan require-ments which have been approved by the Commission for implementation.Section II.K.3.30 of Enclosure 3 to NUREG-0737 outlines the Commission requirements for the industry to demonstrate its small break loss of coolant accident (SBLOCA) methods continue to comply with the requirements of Appendix K to 10 CFR Part 50.

The technical issues to be addressed were outlined in NUREG-0635, " Generic Evaluation of Feedwater Transients and Small Break Loss-of-Coolant Accidents in Comb'ustion Engineering Designed Operating Plants." In addition to the con-cerns listed in NUREG-0635, the staff requested licensees with U-tube' steam generators to assess their computer' codes with the Semiscale S-UT-08 experi-mental' results. This request was made to validate the code's ability to calcu-late the core coolant level depression as influenced by the steam generators prior to loop seal clearing.

In response to TMI Action Item II.K.3.30, the Combustion Engineering Owners Group (CEOG) has elected to justify the continued acceptance of the CEFLASH-4AS i computer program for small break LOCA evaluation. The CEOG response to Action Item II.K.3.30 was transmitted as a generic topical report CEN-203 (including Supplements 1 and 2). -

5 l 05/23/85 1 -

CE II.K.3.30

II. SlM4ARY OF REQUIREMENTS NUREG-0635 required licensees and applicants with Combustion Engineering NSSS designs to address the following concerns:

A. Demonstrate the acceptability of the condensation heat transfer correlation used in the steam generator.

B. Address the effects of noncondensible gases on the condensation heat transfer.

C. Justify the conservatism of modeling ECC mixing as an equilibrium process.

D. Validate the computer program with LOFT integral experiments L3-1 and L3-6, and with Semiscale experiment S-UT-08.

E. Validate the' steam generator model.

F. Validate the core heat transfer and the liquid level models.

G. Confirm adequate accounting of the stored energy (heat) within the primary system metal structures.

H. Validate the acceptability of applying a 1.0 discharge coefficient on both the subcooled and saturated break flow models.

~'

Detailed responses to the above items are documented in CEN-203 (including Supplements 1 and 2) and in a letter from Rick W. Wells (Chairman for the CE Owners Group) to Ceci'l 0. Thomas, " Commitment to Perform Post-Test Analysis of Semiscale Test S-UT-08," dated March 11, 1985. The following is our evaluation of the CEOG response to the concerns outlined above.

e f

e 05/23/85 2 CE II.K.3.30

III. EVALUATION A. Condensation Heat Transfer Model NUREG-0635 requested validation of the condensation heat transfer correlations used in the Combustion Engineering SBLOCA model, CEFLASH-4AS. The correlation used in CEFLASH-4AS is the Akers, Deans and Crosser model. This correlation was demonstrated to provide low heat transfer coefficients. As a consequence the primary system depressurization rate is minimized. This provides greater coolant depletion rates and lower emergency core cooling (ECC) injection. The net result is higher primary system pressures, minimized refluxing (steam gen-erator condensate returning to the core) and delay in core recovery due to lower ECC delivery. -

The staff requested additi6nal validation of the acceptability of the low con-densation heat transfer correlation. The staff had concerns that higher con-2 densation rates could lead to a depression of the mixture level in the core as a result of added inventory accumulation withi, the steam generator tubes.

Such a depression was observed in the Semiscale Test S-UT-08 experimental data.

The CEOG agreed to benchmark the CEFLASH-4AS computer program with the S-UT-08 experimental results. We find this commitment acceptable.

B.

Noncondensible Gas Effects on Condensation Heat Transfer l

NUREG-0635 requested an assessment of the consequences of noncondensible gases in the primary coolant system during a design basis small break LOCA event.

{

The CEOG addressed this concern by degrading the heat transfer coefficient in the steam generators.' The heat transfer degradation was calculated using a boundary layer approach. For this calculation, the noncondensible gases gen-erated within the primary coolant system were collected and deposited on the surface of the steam generator tubes. The sources of noncondensibles consi-dered were: .

(i) Air dissolved in the RWST.

(ii) Hydrogen dissolved in the primary system.

(iii) Hydrogen in the pressurizer vapor space. 03 05/23/85 3 -

CE II.K.3.30

With a degradation factor on the heat transfer coefficient, the limiting SBLOCA was reanalyzed for a typical PWR.

The CEOG, thereby, concluded that formation of noncondensible gases in quantities that may reasonably be expected during a i design basis small break LOCA presents no serious detriment on the PWR system response in terms of core uncovery or system pressure. What perturbation was observed was minor in nature.

The overall heat transfer coefficient would decrease by 3% and the reactor coolant system pressure would increase by 2%. _

The staff finds acceptable the CEOG submittal relating to influences of noncon-densible gases on design bases SBLOCA events.

Our conclusion is based on the fact that a limited amount of noncondensible gas is available during a design basis SBLOCA. In addition, similar conclusions were reached based on Semiscale experiments in which noncondensible gases'were injected in excess amount com-pared to that expected during a design basis SBLOCA.

This issue is resolved.

C. ECCS MODELING NUREG-0635 requested validation of the equilibrium assumption in CEFLASH-4AS for modeling interactions between the injected subcooled emergency core cooling (ECC) water with the saturated primary coolant.

The staff had concerns that the equilibrium model would lead to an overprediction of the primary system depressurization rate (at the point of ECC injection) and thereby create an artificial driving force for excessive makeup of reactor coolant inventory.

This ",self-feeding" process could result in premature core recovery, i

The CEOG response to the above concern demonstrated the SBLOCA modeling tech-

,- nique to be conservative. This was accomplished by modifying a version of CEFLASH-4AS to conduct parametric studies on nonequilibrium thermal-hydraulic conditions.

The licensing model used by C-E injects the ECC directly into the reactor ves-sel downcomer. This is an artificial means of minimizing the impact of the equilibrium model in CEFLASH-4AS.

By injecting the subcooled ECC into a large volume with significant liquid inventory, the incremental change of thermody-namic state is not significantly altered, as would occur if injecting into a small-highly voided node, such as the cold leg.

05/23/85 4 CE II.K.3.30

, - - - + , -, ,,,-w-,w,-w-~,,, w,- r ,,----.r,-- - - - ---,-,,-,,,,--,-,,.-w--e--,- -----,-----wr-

To assess the impact of modeling the ECC injection at the cold legs, the licen-sing model was appropriately modified. As before, the break was postulated at the bottom of the pipe. This maximizes the flow out the break (low quality fluid conditions) while minimizing the rate of depressurization (e.g., minimiz- _

ing the energy crossing the primary system boundary).

The parametric .ctudies did not spill the ECC injected at the broken loop directly to containmet, as assumed in the licensing calculation. Since the break is postulated at the bottom of the pipe and the ECC injection line is attached to the top of the cold leg, direct ' spillage to containment is not physically possible. The staff questioned if postulating the break at the in-jection line would be more limiting. Mr. Gerhard Menzel of C-E replied that this assumption would be less severe sinc'e the break flow will rapidly turn to pure steam and the ratio of depressurization rate to mass depletion rate would increase significantly. This would lead to earlier core recovery.

Results of the parametric analyses on varying the rate of condensation are illustrated in Figure 1 (as taken from CEN-203). Upon completion of this study, the CEOG arrived at the following conclusions:

(i) The limiting SBLOCA occurs for a break size which does not result in SIT actuation.

,(ii)

For breaks which lead to SIT actuation, the present licensing

' model with ECC injected into the downcomer showed that nonequili-brium conditions could lead to a 42*F increase in the calculated PCT.

(iii) For the limiting break size (with injection into the cold legs),

ideal nonequilibrium conditions could lead to an 80*F increase in the calculated PCT, as shown in Figure 1.

1 l (iv) While the sensitivity studies showed a potential for elevated PCTs for the ideal nonequilibrium case (with ECC injected directly into the cold legs), the licensing (equilibrium) ~ evaluation model, by i

l 05/23/85 5 CE II.K.3.30 i

2000 , ,

• -M TEMP = 1940 F .

lEQUILIBRIUMDOWNCOMERINJECTION)

LICENSING MODEL WITH DOWNCOMER ECC INJECTION 1900 -

AND DIRECT SPILLAGE OF BROKEN LOOP INJECTED ECC. .

TO CONTAINMENT.

M -

1800 -

c .L*2 .

. COLD LEG ECC INJECTION WITHOUT g DIRECT SPILLAGE TO CONTAINMENT 3 1700 - E. BROKEN LOOP INJECTED ECC. -

g .

g

• g . ..

g -

m 3

k ~ .

W 1600 -

5 _

e '

d -

~

W 4 E 1500 -

• 1400 ' ' ' '

0.0 ,

0.2 _

O'. 4 0.6 0.8 1.0 NORMALIZED CONDENSATION RATE-  ?

.- FIGURE 1 EFFECT OF CONDENSATION ON PEAK CLAD TEMPERATURE 05/23/85, 6 CE II.K.3.30

p injecting the ECC directly into the downcomer and assuming 1/4 of the ECC spills directly into containment, conservatively calcu-lated a PCT on the order of 190*F higher (see Figure 1).

~

However, as shown in Figure 1, the licensing model with downcomer ECC injection andspillageofthe,brokenloop'ECCtocont[inmentresultsinthehighestpeak

~

clad temperature.

i The staff finds the present licensing model acceptable. This issue is resolved. 3 -

D. CODE VALIDATION Following the Three Mile Island event of 1979, staff analyses with NRC developed computer codes led to concerns that detailed nodalization was required to simu-late realistic s~ystems responses to postulated SBLOCAs. As a consequence, licensees and applicants with Combustion Engineering (C-E) plants were requested to validate their licensing methodology with integral experiments. Specifically, '

the NRC requested that the computer codes be validated with the LOFT L3-1 and L3-6 experimental data. In addition, the staff also requested that the code CEFLASH-4AS be benchmarked with the Semiscale S-UT-08 experimental data.

C-E performed the benchmark analyses for the LOFT L3-1 and L3-6 experiments.

- The CEOG showed good agreement between the CEFLASH-4AS calculations and the l experimental data.

The staff raised concerns regarding the acceptability of the steam generator model in CEFLASH-4AS 'to accurately calculate resistance to core flow prior to clearing of 0.3 reactor coolant loop seals. This integral system response was observed in ttm Semiscale Test S-UT-08. The data showed total core uncovery (based on collapsed core liquid level) prior to clearing of the reactor coolant pump loop seals.

1 -

In response to this issue, the CEOG presented informaticn to justify the acceptability of the steam generator model in CEFLASH-4AS. Although the infor-mation was helpful, the staff judged the submittal as inconclusive. As a 05/23/85 7 CE II.K.3.30

result, the CEOG, in a letter dated March 11, 1985, from R. Wells (Chairman of

'. the CEOG) to C. Thomas (NRC), committed to benchmark the CEFLASH-4AS computer program with the Semiscale S-UT-08 experimental data. This benchmark analysis will be used to confirm the acceptability of the CEFLASH-4AS thermal-hydraulic modeling of the steam generators and its feedback on pre-loop seal clearing core water level depression. Results of this benchmark analysis will be'trans- l mitted to the NRC by December 31, 1985.

The staff concludes the CEFLASH-4AS computer program acceptable with the 1 s

condition that the confirmatory benchmark calculation of Semiscale test S-UT-08 shows acceptable agreement with the data. The CEOG committed to submit j this analysis by December 31, 1985. The staff finds this commitment acceptable.

E. FLOW REGIME INFLUENCES ON PRESSURE DROP t

~

Based on experimental data obtained from the Semiscale Test S-UT-08, the staff requested the CEOG to validate the adequacy of the steam generator model in l CEFLASH-4AS to calculate pressure loss in various flow regimes. The concern arose from the S-UT-08 experimental data which showed total core uncovery (based on collapsed liquid level measurements) prior to the clearing of the reactor coolant pump loop seals.

The CEOG responded to this concern by updating their model and performing the following parametric studies:

~

(i) Homogeneous versus heterogeneous steam generator modeling.

(ii) Counter current flow limit (flooding) modeling.

(iii) Bubble rise modeling.

(iv) Slug flow regime modeling.

(v) Annular flow regime modeling.

The above parametrics were performed individually. The CE0G, thereby, concluded that no significant pressure loss (and consequential core level depression) would occur if any of the above mentioned phenomenological model was incorpor-ated into the code. It was thus decided that these models were not needed to 05/23/85 8 CE II.K.3.30

I i

be incorporated in the licensing model. The licensing model appeared to do an

{

adequate job calculating core water level depression prior to cleaning of the loop seals.

Basedupontheargumentspresented,thestaffcouldnotconcludethatb' FLASH- E 4AS can adequately calculate the core leve1' depression observed in the S-UT-08 experimental data.

The staff believes the pre-loop seal clearing core level depression, as observed in the test, is due to complex interactions between the steam generator heat transfer and various flow regimes. The parametric analyses, while informative, did not provide the assurances required to conclude that the code adequately predicts this integral behavior. The CEOG agreed to benchmark the CEFLASH-4AS computer code with the S-UT-08 experimental data.

This benchmark comparison will be submitted to the staff by Decemb'er 31, 1985. The staff finds this commitment acceptable.

F. CORE HEAT TRANSFER MODEL ,

NUREG-0635 requested validation of the SBLOCA model with core uncovery data obtained from tests conducted at the Oak Ridge National Laboratory (ORNL). The data showed the Dougall-Rohsenow heat transfer correlation (used in the licen-sing model) not to be conservative.

In response to NRC's concerns, the CEOG benchmarked the CEFLASH-4AS code with the ORNL core boiloff tests and demonstrated conservative results. The bubble

~ rise model in CEFLASH-4AS was shown to underpredict the two phase coolant level in the core and the core heat transfer model was shown to calculate conserva-tive peak clad temperatures. CE attributed the conservative temperature calcu-lation to the neglect of a radiation heat transfer model. The ACRS subcommit-tee on ECCS analysis suggested that a more probable cause for the conservatism of the CE heat transfer model may be attributed to the lack of a droplet entrainment model with associated heat transfer characteristics, By neglecting radiation and droplet entrainment, and by benchmarking the co_ with the ORNL core uncovery data, CE demonstrated their licensing model to be conservative.

05/23/85 9 -

CE II.K.3.30

The staff finds the CEOG response to this issue acceptable. This item is

, resolved.

G. METAL HEAT TRANSFER NUREG-0635 requested validation that the stored energy within the metal struc-tures of the reactor coolant system is properly accounted for in the SBLOCA

! calculation. The CEDG responded by confirming that CEFLASH-4AS models the stored energy within'the primary system metal structures. The stored energy is modeled as a lumped parameter with a multiplier on the heat transfer coeffi-cient to assure correct integrated energy transfer to the primary coolant.

The staff concludes that CEFLASH-4AS adequately accounts for the stored energy of the primary system metal structures. This issue is resolved.

H. BREAK FLOW MULTIPLIER ,

i The C-E SBLOCA model applies a break flow multiplier of 1.0 during subcooled and saturated fluid conditions. NUREG-0635 requested validation of the con-servatism of this model.

Of concern was that realistic break flow character-istics require a different break flow multiplier for subcooled and saturated fluid conditions.

In response to this issue, the CEOG performed parametric studies on varying the break flow discharge coefficient. The study concluded that it is conservative to maintain a constant discharge coefficient of 1.0 on both the subcooled and saturated fluid conditions. The duration of core uncovery was determined to be relatively insensitive to the break flow multiplier (see Figure 2). As a consequence, the earlier the time to core uncovery, the higher the calculated PCT.

This is attributed to higher levels of decay heat generated in the core.

The parametric results are illustrated in Figures 2 and 3 (obtained from CEN-203). These figures show that a discharge coefficient of 1.0 on both the sub-cooled and saturated break flow conditions leads to an early core uncovery. The 05/23/85 10 CE II.K.3.30

1

- "9' . , , ,

42. 0DSubcooled CD 29 -

1 5.- .

y -

1.0 1. 3 l N. ---

1.2 1.0 1 .I 0.6

~*-

35. . 1.0

~

i, l. - . . - -

1.0 0.6 A

28. .

j\..

\ .

\. -

\ N. . TOP 0F CORE 7 21- -

\

t s. .

\ \ , s.

. . p .. #.

a  % %- ....d Sf

-i 14. . .

W BOTTOM OF CORE E

x , -

i E I =

O. ' i e i ,

0. 500. 1000. 1500. 2000. 2500. 3000.

TRANSIENT TIME IN SECONDS l

l l

FIGURE 2 BREAK FLOW DISCHARGE COEFFICIENT -

PARAMETRIC STUDY FOR A 0.1 FT2 BREAK 05/23/85 11 .

CE II.K 3.30

- - - ~ - ,., n ,,swq,g, ,

~ SUBC00 LED 1.0 SATURATED .l .0 SUBC00 LED 1.0 ~

SATURATED 0.6 -

SUBC00 LED 0.6 N2a SATURATED 1.0 l a a a q{i SUBC00 LED 1.2 1

.i SATURATED 1.0 ---

' \ 's 1 .,

35.

- \s i

\ \ ',

t I, -

28. -

s \ *\ s s

• --BREAK\ '\ ' .

ELEV 25.79'

\,

\*

\ _

7 N' '

~

h-TOP OF CORE '\ .

% /

' ~

~

. \. -

g ...

g .

< 14,. - '

x BOTTOM OF CORE

7. -

O'. ' ' ' '

O 1000 2000 3000 I4000 5000 TRANSIERT TIME III SECO*1DS

l BREAK FLOW DISCHARGE COEFFICIENT PARAMETRIC STUDY FOR A -

0.05 FT2 BREAK FIGURE 3 05/23/85 12 CE II.K.3.30

. .s

. application of a 1.0 discharge coefficient on the subcooled model and a 0.6 on the saturated model is considered a more realistic approximation of expected flow behavior. While the 1.2 multiplier on the subcooled break flow led to earlier time to core uncovery, this break flow is greater than the expected flow and resulted in only a slight increase ,in the calculated PCT (the PCT was determined to be insensitive to' varying the discharge coefficient during' sub-cooled flow conditions).

The staff finds acceptable the use of a constant break flow multiplies of 1.0 for both saturated and subcooled conditions.: This issue is resolved.

IV. CONCLUSION The Combustion Engineering Dwners Group (CEOG) submitted CEN-203 in response to issues raised in TMI Action Item II.K.3.30 of NUREG-0737. The staff reviewed this submittal, with its accompanying supplements, and find it accept-able pending one condition. This condition requires confirmation that the CEFLASH-4AS computer program can acceptably calculate core level depression, prior to clearing of the reactor coolant pump loop seals, as observed in the data from Semiscale Test S-UT-08.

By letter dated March 11, 1985, from R. Wells (Chairman of the CEOG) to C. Thomas (NRC), the CEDG committed to submit a confirmatory benchmark analysis j of CEFLASH-4AS with Semiscale experiment S-UT-08. This analysis will confirm

~~

the acceptability of the steam generator thermal-hydraulic and integral models of CEFLASH-4AS. We find this commitment acceptable.

l Pending an acceptable benchmark comparison of CEFLASH-4AS with Semiscale exper-iment S-UT-08, the staff finds Action Item II.K.3.30 resolved for CE plants.

With an acceptable calculation / data comparison, the requirements to perform l plant specific analysis, per Action Item II.K.3.31, will no longer be required.

The staff will issue a supplement to this SER following receipt and evaluation of the benchmark submittal.

i l

05/23/85 13 -

CE II.K.3.30