ML20206N148

Submits Response to Re Lessons Learned from Review of Westinghouse Application for Certification of AP600 Std Plant Design

ML20206N148
Person / Time
Site:	05200003
Issue date:	05/12/1999
From:	Travers W NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO)
To:	Powers D Advisory Committee on Reactor Safeguards
References
NUDOCS 9905170149
Download: ML20206N148 (16)

Text

,

d t mrg

% y *, UNITED STATES

e ;5 j

NUCLEAR REGULATORYi.;OMMISSION WASHINGTON, D.C. 20066 0001

\***,,/ May 12, 1999 Dr. Dana A. Powers, Chairman Advisory Committee on Reactor Safeguards U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001

SUBJECT:

LESSONS LEARNED FROM THE ACRS REVIEW OF THE AP600 DESIGN

Dear Dr. Powers:

Thank you for your letter of March 22,1999, regarding lessons learned from the review of Westinghouse's appkcation for certification of the AP600 standard plant design. In your letter, the Advisory Committee on Reactor Safeguards (ACRS) made eight recommendations to the '

Office of Nuclear Reactor Regulation (NRR) that could affect reviews of future applications or be relevant to operating plants. The staff sent four letters to the ACRS on the AP600 design review on March 23, May 20, August 3, and August 27,1998, discussing some of the ACRS's recommendations. The discussion that follows responds to the comments and j recommendations in each section of your letter.

Quality and Timeliness of Material Submitted The staff agrees with the ACRS comments on the poor quality of the application for the AP600 design that was submitted by Westinghouse (W). In response to your recommendation for guidelines, the staff has guidance for the completeness and acceptability of applications in the Licensing Project Manager's Handbook and in NRR Office Letter No. 803, " License Amendment Review Procedures." Also, the NRC's criteria for sufficiency of documentation are

/

stated in 10 CFR 52.47(a)(2) and discussed further in the staff's letters to the ACRS dated March 23 and August 3,1998. Furthermore, the staff rejected W's initial submittalin 1992 and expressed its dissatisfaction to W during NRC's AP600 review on timeliness of W submittals

/  !

and quality of W documentation. The staff documented its concerns on W documentation in several letters to W and in SECY-97-051 and SECY-97-298. Also, major modifications were kl made to the AP600 design and changes to analytical approaches were adopted by W during the review period, which affected staff review time. However, after many requests for additional information and on the basis of the staff's design assurance review findings and W's corrective actions, the staff eventually concluded that the AP600 design documentation complied with Appendix B to 10 CFR Part E0 and was acceptable.

In response to ACRS's comment regarding sufficient technical rationale in the NRR staff's safety evaluation report (SER), the staff's SERs' are summary documents that identify the l informc. tion that was reviewed and the criteria that were used to evaluate the information submitted. The SER also documents the results of the staff's review with a conclusion and a .

summary of the basis for that conclusion. The appropriate level of detailin an SER is a subjective judgement. The staff believes that its SER on AP600 (NUREG-1512) met this purpose and offered a sufficient basis for finding the AP600 design acceptable. In addition, the NRR Licensing Action Task Force is working with the industry to develop guidelines that will improve the quality of SERs in the future.

9905170149 990512 PDR O

ADOCK CS200003 PDR m* - m r- I' P~

f

e 4~ j 2

Thermal-Hvdmulic Code Devetooment The staff agrees with ACRS's recommendation to continue the research phogram on thermal-hydraulic codes. It is evident that the existence of the four separate codes consumes NRC resources, as multiple efforts must be expended to support code improvements, assessments, analyses, and maintenance. The plan to consolidate and improve the NRC thermal-hydraulic code capabilities into a single state-of-the-art code with better physics, modular architecture, and much greater computational efficiency is continuing, although the effort is affected by budget reductions and loss of key staff. Throughout the consolidation effort, NRC is developing innovative code numeric and physical models that, when completed, will be merged with the consolidated code, either during the consolidation program or upon its conclusion. Additionally, a Graphical User Interface is being developed to aid in the use of the code and to provide a tool that will assist the user with the conversion of input decks, so that previously constructed input decks will not become obsolete.

, Code Validation Process l I

In response to ACRS's comment on the AP600 test program, the staff believes that there was  !

sufficient scaling analyses and phenomena identification and ranking tables (PIRTs) for the  !

reactor systems test programs, e.g., Oregon State University tests. However, the staff agrees  !

that there may not have been sufficient pre-test efforts, in the areas of scaling and PIRT, for tho ,

Large-Scale Test program and this made the review more difficult. Furthermore, the staff has l initiated development of guidance for a more proactive, comprehensive, and structured process i for review and approval of thermal-hydraulic codes, and we will continue to meet with the ACRS to discuss the development of our review criteria.

1 One of the key processes in assessing system code capability for transient analyses of nuclear reactors is the establishment of the code scale-up capability to plant conditions. Review of a specific code or model for plant applications involves review of code-governing equations and numerics, review of the constitutive models and correlations, assessment of integrated code i performance, and establishment of code scale-up capability. Code scale-up capability is an 1 important step and it includes the following:

1. assessment of scaling base for facility design and data

2. establishment of code applicability at facility scale I

3. verification of code scale-up capability to plant conditions  ;

I It is clear that this process is complex and requires intimate knowledge of the inner working of l the code, but is necessary to ensure code scale-up capability.

Guidance to carry out the preceding process does not currently exist in the NRC's regulations, regulatory guides, or the standard review plan. The staff is currently preparing a regulatory I guide to describe how to validate and assess codes for particular applications. Guidance for )

acceptable scaling methods will be described in the regulatory guide. Also, code validation using integral and separate effects data will be an important element of the regulatory guide.

In particular, in preparing the regulatory guide, the staff will emphasize the lessons learned from carrying out the original Code Scaling Applicability and Uncertainty Evaluation Methodology (CSAU), and its application to demonstrate the adequacy of the RELAPS code to simulate a small-break loss-of-coolant accident (LOCA) in AP600. Guidance on how to carry out a CSAU process is documented in " Quantifying Reactor Safety Margins," NUREG/CR-5249. This report was focused on the large-break LOCA. The results of application of the CSAU methodology to

o 1 3

AP600 are published in " Top-down Scaling Analyses Methodology for AP600 Integral Tests," l INEL-96/0040, May 1997, and " System Scaling for the Westinghouse AP600 Pressurized .

Water Reactor and Related Test Facilities," NUREG/CR-5541, January 1998. I Main Control Room Staffina Levels in response to a " user need request" from NRR to develop a technical basis for staffing levels needed in conventional and passive reactors, the Office of Nuclear Regulatory Research l completed a study with the HALDEN Reactor Project to investigate issues affecting staffing requirements for advanced reactor designs. The empirical study was conducted to gain data in support of developing review guidance. Two factors were evaluated across a range of plant operating conditions: (1) control room crew staffing sizes and (2) the characteristics of the operating facility itself, whether employing conventional or advanced plant design features. The Loviisa plant in Finland was the conventional design and the HALDEN Human-Machine l Laboratory provided the advanced control room. Preliminary results indicate that control room crew staffing and plant technology have significant effects on operator performance. Plant type, crew size, or their interaction appears to affect every dimemion of performance.

The report on this study will be published in the NUREG/CR series. The study has recently been completed and the draft report is in the final stages of preparation. After the staff has reviewed the report, a decision will be made on developing review guidance or related policy.

In-Vessel Retention of Core Debris The AP600 Design Control Document, which has been approved by the NRC, describes provisions for flooding the AP600 reactor cavity. Therefore, this design provision is in fact part of the AP600 licensing basis. Furthermore, the staff believes that for a well-mixed homogenous molten ceramic pool, the amount of heat that can be removed from the external surface of the AP600 reactor pressure vessel lower head by cavity flooding exceeds the heat imposed on the vessel lower head by molten pool natural convection, which is a requisite condition for -

preventing core debris from penetrating the vessel. However, as stated in your letter, the results of the study used to support the AP600 design are sensitive to numerous assumptions.

For this reason, the staff required and relied upon the results of supporting deterministic analyses of ex vesseiloads to demonstrate that the AP600 containment structure would maintain its integrity in the event of reactor vessel failure. ,

I The staff agrees with ACRS's comment that there are unresolved questions about material interactions, such as intermetalic reactions between molten Zircaloy cladding and the reactor vessel. As a part of the RASPLAV project, a small-scale experiment is planned to address corium melt crust interaction with steel at a maximum temperature of 1670 'K. Because of the '

tungsten structural wall in the RASPLAV test facility, intermetalic reaction experiments cannot be conducted and results will not be conclusive. A follow up to the RASPLAV program, if approved by member countries, includes experiments to deal with corium chemistry and intermetalic reactions, fission products released from a molten corium at elevated temperatures

(~2900 *K), fission product distribution betwee'1 oxidic and metallic layers in the molten pool, and power distribution between liquid and solid corium.

l I

m -

4

- Catalvtic Hydroaen'Recombiners The staff intends to reduce the standards for combustible gas control syst' ems, such as hydrogen recombiners and passive autocatalytic recombiners (PARS), on the basis of their safety significance. As part of the Nuclear Energy Institute whole-plant study, the staff has received risk-informed requests for exemption from the hydrogen rule (10 CFR 50.44). The staff's review has shown that the relative importance of hydrogen combustion for large, dry containments, with respect to containment failure, is quite low. Hydrogen recombiners are incapable of mitigating the consequences from an initial bum caused by the zirconium-cladding metal-water reaction. These firdings support the argument that the hydrogen recombiners are j of little or no risk significance from a containment integrity perspective.  ;

The hydrogen recombiners are of value from a severe-accident management perspective by preventing the uncontrolled burning of hydrogen generated by radiolytic decomposition of water. The risk importance of the hydrogen recombiners does not justify their safety-related ,

classification or inclusion within the plant's technical specifications. However, the stoff still l expects plants to have the capability to remove hydrogen generated by radiolysis in the long l term. This capability would be described in the severe-accident management guidelines and l could be provided by one of several options, such as existing recombiners, dedicated I penetrations available for connection to a portable recombiner, or PARS.

The staff approved the use of PARS for combustible gas control at Indian Point 2 (IP2). This approval is based on their safety significance, the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry. Presently, IP2 uses open-flame recombiners to meet the i requirements of 10 CFR 50.44. Open-flame recombiners require supplementary hydrogen and ,

oxygen for reliable ignition and flame control. Approximately 20 cylinders containing hydrogen )

and oxygen are stored just outside the IP2 personnel airlock as part of the flame recombiner '

system.- The licensee and the staff consider removal of these cylinders a significant safety Improvement.

The staff has found that the PARS proposed for use in IP2 conform to the regulations and standards in Standard Review Plan Section 6.2.5 (NUREG-0800, Revision 2, dated July 1981).

Given the safety significance of these devices, the staff's evaluation found that a persuasive cualification has been established. The depletion rate assumed by IP2 is based on testing conducted by Battelle Frankfurt and is described in two Electric Power Research Institute (EPRI) reports: " Qualification of Passive Autocatalytic Recombiners for Combustible Gas ,

Control in ALWR Containments," dated April 8,1993, and "NIS Passive Autocatalytic j Recombiner Depletion Rate Equation for Evaluation of Hydrogen Recombination During AP600 Design-Basis Accident," dated November 15,1995. Subsequent testing conducted by EPRI  ;

and Electricit6 de France supports the conclusion of the Battelle testing as documented in EPRI Report TR-107517, Volumes 1,2, and 3, " Generic Model Tests of Passive Autocatalytic Recombiners (PARS) for Combustible Gas Control in Nuclear Power Plants," dated June 1997.

This testing was not conducted by IP2 or in accordance with IP2's quality assurance program.

To address this condition, IP2 conducted confirmatory performance testing under its Appendix B quality assurance program at Wyle Laboratories. The test results are documented in Wyle Laboratory Report No. 45971-1, " Functional Testing and Radiation Exposure Test Report," which was forwarded to the staff via letter dated May 16,1997. The test report supported the depletion rate assumed by IP2.

g

.e

a. 5 Hydrogen depletion tests of a scaled PAR were performed at Sandia National Laboratories, under the sponsorship and. direction of the staff. The experiments were to confirm the hydrogen depletion rate of a PAR in the presence of steam and also to ev'aluate the effect of scale on the PAR performance for a variety of hydrogen concentrations. The results of these experiments are documented in NUREG/CR-6580,"Ferformance Testing of Passive Autocatalytic Recombiners," dated June 1998, and show that the depletion rate assumed by IP2 is acceptable.

The PARS are included in the IP2 environmental qualification (EQ) program and have been tested in accordance with Institute of Electrical and Electronic Engineers (IEEE) Standards 627 i and 344. It is the staff's position that the fission products that make up the post-accident I radiation environment must be addressed as possible catalytic poisons. Specifically, the fission products listed in Table 3.13 of NUREG-1465 should be addressed as possible poisons to the PAR catalyst. IP2's response to the staff's position, dated May 2,1997, referenced the EPRI I report, "The Effects of inhibitors and Poisons on the Performance of Passive Autocatalytic Recombiners for Combustible Gas Control in ALWRs," dated April 30,1997. The report

- combines qualitative information based on established chemical and physical principles with quantitative information from testing of catalyst systems subjected to a wide range of inhibitors and poisons. The report concludes that, "Even if the accident were to progress to beyond a DBA [ design-basis accident) to substantial in-vessel damage, PAR recombination capacity 1 would be reduced by no more than 25 percent." The response also described a PAR test  ;

program being conducted in the H2 PAR facility at the Cacarache Research Center by the l French nuclear regulatory agency. These tests attempt to subject a PAR to a severe-accident  !

atmosphere. Simulated fission product aerosols are released into the test facility by an  !

Induction furnace. Within the furnace oven, two dozen elements are being used to simulate a I reactor core inventory. Among these elements are all of the elements in Groups 2,3, and 4 of I the radionuclide groups in NUREG-1465 (Group 1, the stable noble gases xenon and krypton, is not included because the gases in that group are not chemically active). The testing I supports the preceding EPRI conclusion that recombination capacity would be reduced by no l more than 25 percent. '

As part of the IP2 EQ testing, PAR cartridges were exposed to prototypical LOCA temperature, pressure, and spray chemicals (2000 ppm boron, buffered with trisodium phosphate to a pH of 8). This exposure caused the PAR to be initially slow to respond. However, the impact was temporary and functional acceptance criteria were met. It is noted in the qualification summary report that functional acceptance criteria were not met when the spray chemicals were buffered to a pH of 10. This is a much higher loading of trisodium phosphate than would be the case at IP2.

The post accident radiation load on the containment may decompose polyvinyl chloride insulation and yield vapor phase sulfur species and halogen compounds capable of poisoning palladium surfaces. Sulfur oxides, sulfurous acid, and hydrogen sulfide are expected in the containment's post-accident environment. IP2 responded to this issue in its May 2,1997, letter.

The response stated that sulfur oxides and hydrogen sulfide were addressed in the EPRI report, but sulfurous acid was not. IP2 noted that the use of polyvinyl chloride in containment is discouraged and the only cables containing polyvinyl chloride insulation inside containment are for two radiation monitors and motor-bearing thermocouples for five fan coolers. This amount of cabling is not likely to_ be a significant source of potential poisons. A more likely source is polymer chlorosulfonated polyethylene (CSPE), which is commonly used as a jacket material

, over electrical cables. To assess such effects, the EPRI report documented a test in which the recombination rate of a PAR model was measured after being exposed to the fumes from

l l ..

6 1

cabling is not likely to be a significant source of potential poisons. A more likely source is l l polymer chlorosulfonated polyethylene (CSPE), which is commonly used as a jacket material j over electrical cables. To assess such effects, the EPRI report documentdd a test in which the  !

recombination rate of a PAR model was measured after being exposed to the fumes from burning a length of CSPE jacketed cable beneath the PAR model. The PAR model was soaked in water before it was exposed to the fire so that moisture needed to produce such acids as hydrochloric and sulfurous acid would be present. The test showed a decrease in recombination rate of 10 percent at most, but did not verify the amount or presence of acid. In order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it would most likely have to be washed from the cable trays by the containment sprays to the ,

sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest of l

the containment sump inventory. This solution may be steamed back into containment, but at i this point it would be much less of a poisoning threat.  ;

On the basis of the preceding reports and on testing and substantial margin provided (the 4-percent flammability limit is not calculated to be exceeded even if the effectiveness of one PAR is reduced to 10 percent and the second installed PAR is not considered at all), the staff considers the issue of potential catalytic poisoning sufficiently addressed. l l

In a large, dry containment, the buildup of hydrogen caused by radiolysis is not expected to exceed the flammability limit for many days. If the PARS were not performing satisfactorily there would be sufficient time for contingency actions such as bringing an external recombiner from off site. On the basis of their safety significance - the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry - the staff believes that an adequate basis for qualification has been established and that the current standards for combustible gas control systems, such as hydrogen recombiners and PARS, are overly burdensome and that relief is in order.

Sincerely,

1. ~

j e-William D. ravers Executive Director for Operations cc: Chairman Jackson Commissioner Dicus Commissioner Diaz Commissioner McGaffigan Commissioner Merrifield SECY CFO CIO 1

w 6

burning a length of CSPE J cketed cable bene:th ths PAR mod 1. Tha PAR mod:l was soak::d

. in water before it was exposed to the fire so that moisture needed to produce such acids as hydrochloric and sulfurous acid would be present. The test showed a decrease in recombination rate of 10 percent at most, but did not verify the amount or presence of acid. In order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it would most likely have to be washed from the cable trays by the containment sprays to the

. sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest of the containment sump inventory. This solution may be steamed back into containment, but at this point it would be much less of a poisoning threat.

On the basis of the preceding reports and on testing and substantial margin provided (the 4-percent flammability limit is not calculated to be exceeded even if the effectiveness of one PAR is reduced to 10 percent and the second installed PAR is not considered at all), the staff

~ considers the issue of potential catalytic poisoning sufficiently addressed.'

in a large, dry containment, the buildup of hydrogen caused by radiolysis is not expected to exceed the flammability limit for many days, if the PARS were not performing satisfactorily there would be sufficient time for contingency, actions such as bringing an external recombiner from off site. On the basis of their safety significance - the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry - the staff believes that an adequate basis for qualification has been established and that the current standards for. combustible gas control systems, such as hydrogen recombiners and PARS, are overly burdensome and that relief is in order.

' ncerely, Si CC *<

ygraLtTc.s William D. Travers Executive Director for Operations cc: Chairman Jackson Commissioner Dicus ^

Commissioner Diaz Commissioner McGaffigan Commissioner Merrifield '

SECY CFO l CIO DISTRIBUTION: See next page DOCUMENT NAME: G:\ WILSON \ACRS-LTR.WPD *See previous concurrence OFFICE RLSB: DRIP RLSB: DRIP SRXB SPSB D: DRIP NAME JNWilson:sg* CGrimes* JWermiel* RBarrett* DMatthews*

DATE 04/23/99 04/20/99 04/22/99 04/22/99 04/23/IP OFFICE RES lOHB ADIP ADPT y DIR:NRR , g EDh

-4 \1 '- N NAME AThadans RGallo* WKane* BSheron* hCollins WTraveys DATE 04/23/99 04/22/99 04/27/99 04/27/99 h/[/99. _, (/h99 OFFICIAL RECORD COPY g7

' DISTRIBUTION: '

Docket File (50-003)

PUBLIC RLSB R/F -

EDO GT19990155 EDO R/F WTravers,0-16 E15 '

FMiraglia,0-16 E15 MKnapp,0-16 E15 ,

l PNorry,0-16 E15 JBlaha,0-16 E15 ..

SCollins/RZimmerman,0-5 E7 WKane,0-5 E7.

BSheron,0-5 E7 DMtthews/SNewberry CGrimes -

JNWilson OGC OPA OCA NRR Mail Room, GT19990155 MManahan .

JWermiel,0-10 E4 RBarrett,0-10 H1 RGallo TKenyon

.WHuffman EThrom,0-10 D17 Alevin,0-10 B3 RCaruso,0-10 B3 MSnodderly,0-10 D17 RPalla,0-10 D17 JBongarra

. Feitawila Abehbahani SBums,0-15 B18 CPaperiello, T-8 A23 JMitchell,0-16 E15 ACRS File U..

6 burning a li:ngth of CSPE J cketed cabla bantath the PAR mod:1. Tha PAR modal was soaktd

-in witar bsfors it was exposed to the firo so that moisture necdad to producs such acids as hydrochloric and sulfurous acid would be present. The test showed a decrease in recombir,ation rate of 10 percent at most, but did not verify the amount or presence of acid. In order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it would most likely have to be_ washed from the cable trays by the containment sprays to the sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest of the containment sump inventory. This solution may be steamed back into containment, but at this point it would be much less of a poisoning threat.

On the basis of the preceding reports and on testing and substapth margin provided (the 4-percent flammability limit is not calculated to be exceeded evepif the effectiveness of one PAR is reduced to.10 percent and the second installed PAR is notponsidered at all), the staff considers the issue of potential catalytic poisoning sufficienp addressed, j l

In a large, day containment, the buildup of hydrogen cause/d by radiolysis is not expected to exceed the flammability limit for many days. If the PARS were not performing satisfactorily i there would be sufficient time for contingency actions,such as bringing an external recombiner I from off site. On the basis of their safety significanc,e - the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry- the staff be'lieves that an adequate basis for qualification has been established and that the current standaids for combustible gas control systems, such as hydrogen recombiners and PARS, are overly burdensome and that relief is in order.

/

, Sincerely,

/ William D. Travers, Executive Director for Operations

/

cc: Chairman Jackson /

Commissioner Dicus

/

Commissioner Diaz /

Commissioner McGaffigan

/ ,

Commissioner Merrifield 1 SECY CFO e ClO DISTRIBUTION: See next page DOCUMENT NAME: G:\ WILSON \ACRS-LTR.WPD *See previous concurrence OFFICE RLSB:DylP RLSB: DRIP SRXB SPSB D: DRIP  ;

NAME [hhsg CGrimes* JWermiel* RBarrett* DMatthews*

DATE h /99 04/20/99 04/22/99 04/22/99 04/23/99 OFFICE RES NRR EDO IOHB ADIP ADFh d

NAME AThadani SCollins WTravers RGallo* WKane* BShN DATE 04/23/99 / /99 / /99 04/22/99 04 / 27 /99 Y/$/99 OFFICIAL RECORD COPY V

q

o 6

- burning a Inngth of CSPE jackstrd cabla benrath the PAR model. The PAR model was soaked

, . in water before it was exposed to the fire so that moisture needed to produce such acids as hydrochloric and sulfurous acid would be present. The test showed a decrease in 4

recombination rate of 10 percent at most, but did not verify the amount or presence of acid. In order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it would most likely have to be washed from the cable trays by the containment sprays to the sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest of the containment sump inventory. This solution may be steamed back into containment, but at this point it would be much less of a poisoning threat.

l On the basis of the preceding reports and on testing and substantial margin provided (the 4-percent flammability limit is not calculated to be exceeded even if the effectiveness of one PAR is reduced to 10 percent and the second installed PAR is not considered at all), the staff  ;

considers the issue of potential catalytic poisoning sufficiently addressed.

In a large, dry containment, the buildup of hydrogen caused by radiolysis is not expected to exceed the flammability limit for many days. If the PARS were not performing satisfactorily ,

there would be sufficient time for contingency actions such as bringing an external recombiner from off site. On the basis of their safety significance -the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry - the staff believes that an adequate basis for qualification has been established and that the current standards for combustible gas control systems, such as hydrogen recombiners and PARS, are overly burdensome and that relief is in order.

Sincerely,

)

William D. Travers, Executive Director for Operations cc: Chairman Jackson Commissioner C:cus Commissioner Diaz Commissioner McGaffigan Commissioner Merrifield SECY CFO l CIO DISTRIBUTION: See next page DOCUMENT NAME: G:\ WILSON \ACRS-LTR.WPD OFFICE RLJfu2 RIP, RLSB:D@P/[ SRXB SPSB D: DRIP NAME hv'6p[ CGrimeh JWermiei RBarrett DMatthews DATE //e/99 [ /do /99I / /99 / /99 / /99 OFFICE RES NRR EDO IOHB ADIP NAME AThadani SCollins WTravers RGallo WKane DATE k //h99

~

/ /99 / /99 / /99 / /99

/ OFFICIAL RECORD COPY l

6 burning a 1:ngth of CSPE jrcket:d cible benuth the PAR model. Tha PAR mod 11 was soaked in water before it was exposed to the fire so that moisture needed to produce such acids as .

. hydrochloric and sulfurous acid would be present. The test showed a decrease in recombination rate of 10 percent at most, but did not verify the amount or presence of acid, in order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it would most likely have to be washed from the cable trays by the containment sprays to the sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest of the containment sump inventory. This solution may be steamed back into containment, but at this point it would be much less of a poisoning threat.

On the basis of the preceding reports and on testing and substantial margin provided (the 4-percent flammability limit is not calculated to be exceeded even if the effectiveness of one PAR is reduced to 10 percent and the second installed PAR is not considered at all), the staff considers the issue of potential catalytic poisoning sufficiently addressed.

In a large, dry containment, the buildup of hydrogen caused by radiolysis is not expected to exceed the flammability limit for many days. If the PARS were not performing satisfactorily there would be sufficient time for contingency actions such as bringing an external recombiner from off site. On the basis of their safety significance -the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry - the staff believes that an adequate basis for qualification has been established and that the current standards for combustible gas control systems, such as hydrogen recombiners and PARS, are overly burdensome and that relief is in order.

Sincerely, William D. Travers, Executive Director for Operations j l

cc: Chairman Jackson Commissioner Dicus Commissioner Diaz

)

Commissioner McGaffigan Commissioner Merrifield SECY CFO CIO DISTRIBUTION: See next page DOCUMENT NAME: G:\ WILSON \ACRS-LTR.WPD OFFICE RLjBARIP n RLSB:DFJjp/L SRXB SPSB D: DRIP NAME CGrimeh JWermiel RBarrett DMatheh DATE [/ e/99 [ /dv /99I / /99 / /99 Y /l.'>/99 OFFICE - RES" NRR EDO IOHB ADIP NAME AThadani SCollins WTravers RGallo WKane DATE / /99 / 19 9 / /99 /. /99 / /99 OFFICIAL RECORD COPY

m 6

r.

buming a Irngth of CSPE jacksted c bla benstth tha PAR model. The PAR model was soaked

' in water before it was' exposed to the fire so that moisture needed to produce such acids as hydrochloric and sulfurous acid would be present.. The test showed a decrease in ,

. recombination rate of 10 percent at most, but did not verify the amount or presence of acid. In order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it would most likely have to be washed from the cable trays by the containment sprays to the sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest of the containment sump inventory. This solution may be steamed back into containment, but at this point it would be much less of a poisoning threat.

On the basis of the preceding reports and on testing and substantial margin provided (the 4-percent flammability limit is not calculated to be exceeded even if the effectiveness of one PAR is reduced to 10 percent and the second installed PAR is not considered at all), the staff considers the issue of potential catalytic poisoning sufficiently addressed.

'i n a large, dry containment, the buildup of hydrogen caused by radiolysis is not expected to exceed the flammability limit for many days. If the PARS were not performing satisfactorily there would be sufficient time for contingency actions such as bringing an extemal recombiner from off site. On the basis of their safety significance - the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry _- the staff believes that an adequate basis for qualification has been established and that the current standards for combustible gas control systems, such as hydrogen recombiners and PARS, are overly burdensome and that reliefis in order.

Sincerely, i William D. Travers, Executive Director for Operations cc: Chairman Jackson Commissioner Dicus Commissioner Diaz Commissioner McGaffigan -

Commissioner Merrifield SECY CFO CIO DISTRIBUTION: See next page

. DOCUMENT NAME: G:iWILSON\ACRS-LTR.WPD OFFICE RLJBI! RIP n RLSB:D$P/[ SRg , ,. SPSB D: DRIP NAME hoiv'unp[ CGrimeh JWhel RBarrett DMatthews DATE- [/ */99 [ lh /99i +/M/99 / /99 / /99 OFFICE RES NRR- EDO IOHB ADIP NAME AThadani SCollins WTravers RGallo WKane DATE / /99 - / /99 / /99 / /99 / /99 OFFICIAL RECORD COPY

n

~

6 burning's I::ngth of CSPE JIckst::d cable bens th ths PAR modal. Ths PAR modsl was soaktd in water before it was exposed to the fire so that moisture needed to produce such acids as hydrochloric and sulfurous acid would be present. The test showed a decrease in . .

recombination rate of 10 percent at most, but did not verify the amount or presence of acid. In j

. order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it.

would most likely have to be washed from the cable trays by the containment sprays to the 1 sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest.

i of the containment sump inventory. This solution may be steamed back into containment, but  !

at this point it would be much less of a poisoning threat. I On the basis of the preceding reports and on testing and substantial margin provided (the 4-percent flammability limit is not calculated to be exceeded even if the effectiveness of one PAR is reduced to 10 percent and the second installed PAR is not considered at all), the staff considers the issue of potential catalytic poisoning sufficiently addressed.

In a large, dry containment, the buildup of hydrogen caused by radiolysis is not expected to exceed the flammability limit for many days. If the PARS were not performing satisfactorily.

there would be sufficient time for contingency actions such as bringing an extemal recombiner from off site. On the basis of their safety significance - the safety benefit they would provide over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry -~ the staff believes that an adequate basis for qualification

{

has been established and that the current standards for combustible gas control systems, such l as hydrogen recombiners and PARS, are overly burdensome and that relief is in order.

Sincerely, William D. Travers,  !

Executive Director for Operations cc: Chairman Jackson Commissioner Dicus Commissioner Diaz ,

Commissioner McGaffigan Commissioner Merrifield SECY CFO.

ClO DISTRIBUTION: See next page DOCUMENT NAME: G:\ WILSON \ACRS-LTR.WPD ,//

OFFICE RLJEQRIP, RLSB:D@y/[ SRXB SPSpj[) D: DRIP NAME. CGrimeh JWermiel MBark DMatthews DATE- [/ &/99 [ /dv /99Y / /99 Y /b/99 / /99 OFFICE RES NRR EDO IOHB ADIP NAME AThadani SCollins WTravers RGallo WKane DATE / /99 / /99 / /99 /. /99 / /99 OFFICIAL RECORD COPY

6 l

burning a 1:ngth of CSPE J ckstad cabla benteth ths PAR modsl. The PAR mods! was sotked in water before it was exposed to the fire so that moisture needed to produce such acids as hydrochloric and sulfurous acid would be present. The test showed a decrease in recombination rate of 10 percent at most, but did not verify the amount or presence of acid. In  !

order for the PAR to be exposed to acid created by the decomposition of cable jacketing, it l would most likely have to be washed from the cable trays by the containment sprays to the {

sump. Once in the sump it would be diluted and buffered by trisodium phosphate with the rest '

of the containment sump inventory. This solution may be steamed back into containment, but at this point it would be much less of a poisoning threat.

On the basis of the preceding reports and on testing and substantial margin provided (the 4-percent flammability limit is not calculated to be exceeded even if the effectiveness of one PAR l

is reduced to 10 percent and the second installed PAR is not considered at all), the staff j considers the issue of potential catalytic poisoning sufficiently addressed. )

In a large, dry containment, the buildup of hydrogen caused by radiolysis is not expected to exceed the flammability limit for many days. If the PARS were not performing satisfactorily '

there would be sufficient time for contingency actions such as bringing an extemal recombiner from off site. On the basis of their safety significance - the safety benefit they would provide

over the currently installed system, the excessive margin provided, and the qualification testing performed by the staff and industry - the staff believes that an adequate basis for qualification has been established and that the current standards for combustible gas control systems, such as hydrogen recombiners and PARS, are overly burdensome and that relief is in order.

Sincerely, William D. Travers, Executive Director for Operations cc: Chairman Jackson -

Commissioner Dicus Commissioner Diaz Commissioner McGaffigan )

Commissioner Merrifield j SECY i CFO  !

CIO I DISTRIBUTION: See next page DOCUMENT NAME: G:\ WILSON \ACRS-LTR.WPD OFFICE RLSfLQRIP, RLSB:DFyff/, SRXB SPSB D;r) RIP NAME hv^vnp[ CGrimeh JWermiel RBarrett DMatthews DATE [/ e/99 [ lh 199I / /99 / /99 / /99 OFFICE RES NRR EDO lOHBhM\ ADIP NAME AThadani SCollins WTravers RC WKane DATE  ! /99 /- /99 / /99 h b9 ' / /99 OFFICIAL RECORD COPY (

h

15

~

CFO CIO-DOCUMENT NAME: G:\WILSO'N%CRS-LTR.WPD !

l i

i TECH EDITOR

/Af4*dce

/// /f /99 I

-. . {

. l ( >

lCY EDO Principal Correspondence Control j

FROMt DUE: 04/22/99 EDO CONTROL: G19990155 >

DOC DT: 03/22/99 - gh F:n2 A. Powers, ACRS FINAL REPLY:

8 { l l

TO

($F

\/

J I

Tr:: vers, EDO h0RSIGNATUREOF  : ** GRN ** CRC NO:

Tr:: vers, EDO lESC ROUTING:

LESSONS LEARNED FROM THE ACRS REVIEW OF THE AP600 Travers DESIGN Knapp i Miraglia j Norry l Blaha  !

Burns DATE: 03/25/99 Thadani, RES l Paperiello,NMSS ASSIGNED TO: CONTACT: Mitchell, OEDO ACRS File NRR Collins l

SPECIAL INSTRUCTIONS OR REMARKS:

R$f. Gl9980459.

Pr: pore response to ACRS for EDO signature. Add Commissioners and SECY as cc's.

USE SUBJECT LINE IN RESPONSE. ,

'\CTION t 4](( (ld$nG ' bfdhLl}$dlwa)

~ ~ ~

M M DXCi 21

$(dd Lb&kud Lj')fagd5'/999 BY _ (E Tf 4](( L/bahq/ Q%/j papxa/O - - - -

il&

YkbiC]U ,

YkT V7klwnd

kOS, C314%o(55 OFFICE OF THE SECRETARY CORRESPONDENCE CONTROL TICKET ( -

PAPER NUMBER: CRC-99-0269 LOGGING DATE: Mar 25 99 ,

ACTION OFFICE: EDO AUTHOR: DANA POWERS ,

AFFILIATION: ADVISORY COMMITTEE ON REACTOR SAFEGUARDS ADDRESSEE: TRAVERS LETTER DATE: Mar.22 99 FILE' CODE: OM-7 ACRS

SUBJECT:

LESSONS LEARNED FROM THE ACRS REVIEW OF THE AP600 DESIGN ACTION: Appropriate DISTRIBUTION: CHAIRMAN, RF SPECIAL HANDLING: NONE CONSTITUENT:

NOTES:

DATE DUE:

SIGNATURE: . DATE SIGNED:

AFFILIATION:

1 J

l l

I j

g UNITED STATES 8 NUCLEAR REGULATORY COMMISSION

& ADVISORY COMMITTEE ON REACTOR SAFEGUARDS >

b. WASHINGTON,0. c. 30055 l March 22,1999 Dr. William D. Travers Executive Director for Operations U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001

SUBJECT:

LESSONS LEARNED FROM THE ACRS REVIEW OF THE AP600 DESIGN During the 460th meeting of the Advisory Committee on Reactor Safeguards, March 10-13, 1999, we completed deliberations regarding lessons leamed from our review of the AP600 passive plant design. As noted in our July 23,1998 report, issues on the safety aspects of the -

AP600 application were resolved to our satisfaction. In the course of our review, however, we identified some lessons leamed that could affect reviews of future applications or that could be rel.evant to operating plants.

Recommendations

1. Guidelines on the acceptable quality of documentation submitted by the applicant and on l

the lead times necessary for staff reviews should be established and enforced.

)

2. Safety evaluation reports should include more of the technical rationale leading to the regulatory decision.

3. The NRC research program to improve and consolidate thermal-hydraulic codes should be continued.

j i

4. Guidance for acceptable scaling methods, such as the Code Scaling Applicability, and Uncertainty (CSAU) evaluation methodology, and for acceptable utilization of integral l test data for the validation of computer codes should be developed.

5.

The development of tec,hnical and policy guidelines for approving requests for reducing the main control room staffing levels below present regulatory limits should be considered.

6. More experiments and analyses will be required before in-vessel core debris retention can be credited as part of the licensing basis.

gA M EDO --G19990155 0L LG VXl f (1' .-

l' 2

7. Better standards for qualification of catalytic hydrogen recombiners should be required before approving these recombiners for use as safety-related equipment in nuclear power plants.

Quality and Timeliness of Material Submitted Our review was made particularly difficult because the associated documentation was submitted  ;

piecemeal, was sometimes of poor quality, and contained technical errors. For future applications, the staff should establish and enforce guidelines on the acceptable quality of documentation and on the lead times necessary for staff reviews.

The section of the safety evaluation report (SER) related to the AP600 test and analysis program lacked sufficient technical rationale for us to judge the quality of the staffs review. Our Thermal-Hydraulic Phenomena Subcommittee had to perform a much more exhaustive review  ;

than should have been necessary in order to become convinced of the adequacy of the staff l reviewc Future SERs should include more of the technical rationale used to make regulatory I decisions. .

Ibg.rmgi-Hydraulic Code Develooment Our review identified deficiencies in the existing suite of NRC thermal-hydraulic coden er databases. In order to ensure that the staff has an acceptable thermal-hydraulic analysis capability for confirmatory review of license applications and amendments, the NRC research program to improve and consolidate thermal-hydraulic codes should be continued.

Code Validation Proc.ess The scope of the Westinghouse test and analysis program in support of the AP600 certification was extensive. However, the test program was completed prior to both the scaling analyses and the phenomena identification and ranking process. Because of this, we had considerable difficulty in evaluating both the quality of the data used to validats the computer codes and the scaling of the test results to AP600 conditions. The staff should develop guidance for acceptable methods for scaling and uncertalnty evaluation, such as the CSAU evaluation methodology, and for acceptable utilization of integral test data for the validation of computer codes. This is especially crucial as we make more use of best-estimate models for emergency core cooling system requirements.

Main Control Room Staffino Levels The AP600 is designed to allow the reactor safety systems to remove decay heat without any required operator actions for up to 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> after the onset of a severe accident. In addition, the instrumentation and control systems and the human factors design of the main control room provide improved access to information on plant operating parameters. This facilitates and speeds the operator's ability to diagnose problems. Based on these developments and the results of current h:.: man factors research, the staff should consider developing technical and l

l l

?, .

4 3 policy guidelines for reviewing and approving licensee and applicant requests for reducing the main control room staffing levels below present regulatory limits.

In-Vessel Retention of Core Debris The AP600 design contains provisions to flood the reactor cavity to cover a significant portion of the reactor vessel. It was argued that this design provision could result in the removal of sufficient heat to prevent core debris from penetrating the vessel. Although this strategy was not part of the AP600 licensing basis, such a strategy might be included in future license I amendment requests.

The staff identified weaknesses in the in-vessel core debris retention study used to support the l AP600 application. The staff found that the results were quite sensitive to assumptions l

conceming the mass of metallic core debris in the vessel plenum and the magnitude of upward l heat flux induced by vaporization of volatile constituents of core debris. In addition, analyses by the staff questioned assumptions made in the study concoming material properties. There are also unresolved questions about materials interactions, such as intermetallic reactions between molten Zircaloy cladding and the reactor vessel.

More experiments and analyses are needed before in-vessel core debris retention can be credited as part of the licensing basis. At this time, we believe in-vessel core debris retention should only be considered as a severe accident management strategy.

Catalvtic Hydroaen Recombiners The design of the AP600 utilizes hydrogen recombiners to control the accumulation of hydrogen in the reactor containment following a design-basis accident. The AP600 design also containers hydrogen ignitors to prevent hydrogen accumulation in the event of more serious beyond-design-basis accidents. The possible use of catalytic processes to control hydrogen concentrations in reactor containments is gaining popularity throughout the world.

The catalytic recombiners that are proposed for use in the AP600 are based on palladium or platinum dispersed on alumina. There is lacking, however, a good understanding of the vulnerabilities of these devices to the environment expected to exist following either design basis or severe accidents. There is not yet a good understanding of what would constitute persuasive qualification of a catalytic recombiner. We believe that the staff should establish better standards for the qualification of these devices.

Dr. Thomas S. Kress did not participate in the Committee's deliberation regarding extemal reactor vessel cooling.

I Dr. Dana A. Powers did not participate in the Committee's deliberation regarding the results of Sandia National 1.aboratories' tests on qualification of passive autocatalytic recombiners.

j i

l I

.. . . . _ _ . _ . . _ . _ _ . _ . _ . . . _ . . , . . . l

l

~

. 4

'k.

Dr. George Apostolalds did not participate in the Committee's deliberation regarding the analyses perfoiTned by the Idaho National Engineering and Environmental Laboratory concerning extemal reactor vessel cooling.

Sincerely,

^

J_ a .cT= =e Dana A. Powers Chairman

Reference:

Report dated July 23,19g8, from R. L Seale, Chairman, ACRS, to Shirley Ann Jackson, Chairman, NRC,

Subject:

Report on the Safety Aspects of the Westinghouse Elect l Application for Certification of the AP600 Passive Plant Design.

.