ML20206A937

Summary of 990413-14 Meeting with Industry Representatives in Burr Ridge,Il Re Industry Initiative to Address GL 96-06 Waterhammer Issue.List of Attendees & Meeting Handouts Encl. Work Being Sponsored Through NEI & Coordianted by EPRI

ML20206A937
Person / Time
Issue date:	04/23/1999
From:	Tatum J NRC (Affiliation Not Assigned)
To:	Hannon J NRC (Affiliation Not Assigned)
References
PROJECT-689 GL-96-06, GL-96-6, NUDOCS 9904290092
Download: ML20206A937 (59)

Text

tro Y 81 UNITED STATES j

't j NUCLEAR REGULATORY COMMISSION o WASHINGTON. D C. 2055M001

% - ,$ April 23, 1999 i

MEMORANDUM TO: John N. Hannon, Chief Plant Systems Branch Division of Systems Safety and Analysis THRU: George T. Hubbard, Jr., Chief G/.

2 Balance of Plant Section f , 4, Plant Systems Branch Division of Systems Safety and Analysis  ;

FROM: James E. Tatum /

Balance Plant SystemsofBranch Plant Section /j) #

Division of Systems Safety and Analysis

SUMMARY

OF THE APRIL 13-14,1999. MEETING BETWEEN THE NRC STAFF AND INDUSTRY REPRESENTATIVES AT THE BEST I WESTERN INN OF BURR RIDGE, ILLINOIS, TO DISCUSS THE INDUSTRY INITIATIVE TO ADDRESS THE GL 96-06 WATERHAMMER ISSUE

' George Hubbard and Jim Tatum of DSSA/SPLB, Gary Hammer of DE/EMEB, Doctor Hossein Nourbakhsh (NRC contractor), and Doctor Madan Prabhakara (NRC contractor) attended a meeting at the Best Western inn of Burr Ridge in Illinois, on April 13 and April 14,1999, to discuss the current status of work being done to address the GL 96-06 waterhammer issue.

This meeting was a continuation of the dialogue that was initiated during a previous meeting on this subject that was held in Boston, Massachusetts on Februa- 2 and 23,1999. This industry initiative is being sponsored by 12 utilities through the .'. lear Energy Institute (NEI),

and the technical work is being coordinated by the Electric Pom Research Institute (EPRI).

Waterhammer testing in support of this effort is being performed by Altran Corporation and by l Fauske and Associates, Inc. (FAI). The meeting was attended by representatives of the licensees who are sponsoring this effort, Attran Corporation, FAl, and EPRI. Doctor Fred Moody from General Electric, Doctor Benjamin Wylie from the University of Michigan, and Doctor Peter Griffith from the Massachusetts Institute of Technology, who collectively represent the industry's expert panel for this initiative, were also present at the meeting. A list of the attendees is included as Attachment 1, the meeting agenda is included as Attachment 2, and

/Of meeting handouts are included as Attachment 3. y The purpose of this industry initiative is to develop a waterhamme Technical Basis Report (TBR) that can be used by licensees to evaluate the effects of waterhammer in low-pressure fluid systems. It is thought that existing methods (i.e., those provided 5 NUREG/CR-5220) are too conservative for low-pressure applications, and a less conservative methodology could result in substantial savings to the industry. During this most recent meeting, additional data from industry waterhammer testing was presented and attendees observed tests that were being conducted by Fauske and Associates, Inc. at their test facility. Plans for performing 9904290092

" 990423~ ~ " ~

~'fl*B (l/

paggy 1

"" " ER m C OPV l J

1 2

additional testing were discussed, and the various sections of the waterhammer TBR were outlined. The industry working group plans to finish the waterhammer testing within a couple of months, and provide the staff a rough draft of the waterhammer TBR by the end of June,1999, with the interim (final) waterhammer TBR being submitted for staff review and approval by l July 15,1999. Once approved by the staff, EPRI will publish the TBR in final form and l licensees will be able to use the TBR to address the waterhammer concerns discussed in GL 96-06. It is the staff's goal is to complete a timely review of the TBR (i.e., within a couple of months) after it has been submitted to the NRC.

l Following the presentations, the staff offered the following comments and suggestions:

l l

The industry working group appears to be making good progress in identifying the l significant waterhammer issues and phenomena that need to be addressed in the TBR.

The remaining challenge is how to pull all the information that is being developed together in the TBR for plant-specific application.

l Careful review and scrutiny is needed in applying the information that is being developed to the system configurations that actually exist in the plants (closed loop vs. open loop, fan cooler bypass configurations, piping transition areas such as channel heads, etc.). The limitations of laboratory test results should be realized and accounted for in the TBR (if appropriate) based on the importance of the limitation.

• lt is important to recognize the difference between variables (which may have some uncertainty associated with them) and the uncertainties (or unknowns) that exist in general I when attempting to model and analyze waterhammer events.

• Risk should be taken into consideration when evaluating data and determining if additional information should be developed. Risk arguments may also be used to help justify plant-specific relief requests for situations where piping and support loads do not fall within the applicable design criteria for the plant.

• Waterhammer tests are being completed by both Altran and FAl, and the data from both of these efforts, as well as other information that has been collected through document j review, should be combined and considered collectively in developing the TBR (i.e., test results should not be used or evaluated in a segregated or fragmented way).

• Drain down tests that have been completed by Altran indicate that several pressure peaks i can occur, but additional testing is needed to demonstrate that the first peak will be the i I

maximum, even for the various system configurations, UD ratios, and drain down scenarios that actually exist in the plants.

• If components (such as check valves) are credited in establishing the initial system j conditions at the onset of a waterhammer event, programs must be in place to assure that i the component will perform in the manner that is assumed. For example, the hinge pins in l

- service water system check valves can become rusted or clogged with silt, causing the valves to stick in the open position. If a check valve is credited for preventing system 1 l

drainage, there should be a program in place to assure that the valve is maintained in good i l working order.

1

)

3 i

As part of the validation process, it would be a good idea to apply the TBR to a couple of l plants as a way to test the methodology and to determine if any problems with

. implementation or areas of confusion exist.

Although the current schedule does not call for another meeting with the NRC staff to discuss the results of the additional testing that is being completed, closure with the staff is needed to facilitate a complete understand,ng of the TBR methodology that is being developed. I The interaction between the NRC staff and industry representatives went very well during the two-day meeting, and the staff continues to be encouraged by this initiative.

Project No. 689 DISTRIBUTION (Hard Coov):

SPLB R/F tiCential File w, PDR OGC ACRS ASerkiz BWetzel See next page E-Mail (w/o Attachments 2 and 3)

SCollins/RZimmerman BSheron GHolahan/TCollins JStrosnider/RWessman KManoly JFair CHammer l

DOCUMENT NAME: MTGMIN2.WPD Ta receive e copy of this document, indcate in the box: *C" = Copy without attachment / enclosure "E" = Copy with attachment / enclosure *N" - No copy OFFICE SPLB:DSSA:NRR l6 PGEB:DRP,M l SC:SPLB:DS,SA f l NAME JTatum:bwgd SMagruder 4M GHubbard 86V

! DATE 4/4:V99 [)' 4/,$/99 4/d99 / /99 / /99 l OFFICIAL RECORD COPY z n e (.-

M$

?

t

1. h

~. 'i 3

As part of the validation process, it would be a good idea to apply the TBR to a couple of plants as a way to test the methodology and to determine if any problems with implementation or areas of confusion exist.

Although the current schedule does not call for another meeting with the NRC staff to discuss the results of the additional testing that is being completed, closure with the staff is needed to facilitate a complete understanding of the TBR methodology that is being developed.

The interaction between the NRC staff and industry representatives went very well during the two-day meeting,' and the staff continues to be encouraged by this initiative.

Project No. 689 l

i-( .j

ATTACHMENT 1 o 9 MEETING PARTICIPANTS 1

Name Affiliation A. Singh Electric Power Research Institute H. Chang New York Power Authority i L. Rochino Rochester Gas and Electric Corporation J. Wade Southern Nuclear Operating Company M. Aulik Wisconsin Public Service Corporation B. Hammersley Fauske and Associates, incorporated T. Brown Duke Energy G. Hubbard Nuclear Regulatory Commission l

G. Hammer Nuclear Regulatory Commission J. Tatum Nuclear Regulatory Commission 1- B. Wylie University of Michigan P. Griffith Massachusetts institute of Technology l F. Moody General Electric Corporation B. Henry Fauske and Associates, incorporated A. Ginsberg Consolidated Edison Company R. Randels Commonwealth Edison Company G. Zyst Altran Corporation l M. Zweigle Altran Corporaaon T. Esselman Altran Corporation l V. Wagoner Carolina Power and Light Company A. Arastu Bechtel Corporation l K. Ramsden Commonwealth Edison Company H. Nourbakhsh NRC Contractor M. Prabhakara NRC Contractor

! S. Thomas Northern States Power Company A. Setlur Contractor for Northem States Power Company l

l l

L  ;

Nuclear Energy Institute Project No. 689 cc:(w/o Attachment 3)

Mr. Ralph Beedle Ms. Lynnette Hendricks, Director i

Senior Vice President Plant Support and Chief Nuclear Officer Nuclear Energy Institute Nuclear Energy institute Suite 400 Suite 400 1776 l Street, NW 1776 i Street, NW Washington, DC 20006-3708 Washington, DC 20006-3708 l

Mr. Alex Marion, Diiector Mr. Charles B. Brinkman, Director Programs Washington Operations Nuclear Energy Institute ABB-Combustion Engineering, Inc.

Suite 400 12300 Twinbrook Parkway, Suite 330 1776 i Street, NW Rockville, Maryland 20852 Washington, DC 20006-3708 Mr. David Modeen, Director Engineering Nuclear Energy Institute Suite 400 17761 Street, NW Washington, DC 20006-3708 Mr. Anthony Pietrangelo, Director Licensing Nuclear Energy Institute Suite 400 1776 l Street, NW Washington, DC 20006-3708 Mr. Nicholas J. Liparulo, Manager Nuclear Safety and Regulatory Activities Nuclear and Advanced Technology Division Westinghouse Electric Corporation P.O. Box 355 Pittsburgh, Pennsylvania 15230 Mr. Jim Davis, Director Operations Nuclear Energy Institute Suite 400 1776 I Street, NW Washington, DC 20006-3708

i Attachment 2 Preliminary Agenda for the Third Meeting of the Waterhammer Expert Panel First Day,4/13/99 Time Topic Presenter 8:00 ANI Introduction and Objectives of the Project and V. Wagoner Expert Panel- Update 8:15 A51 Thermal Layer & Steam Air Content Testing R. IIenry Solution Approach

. Testing Results

• Status 10:00 Ah! Ilreak 10:15 A31 Thermal Layer & Steam Air Content Testing R. Henry (continued) 11:00 Ah!

Travel to FAI for Demonstration and Discussion of FAI Data 12:00 Noon Lunch {

1:00 PAI TIIR Outline / Contents / Status {

T. Esselman "What will it contain?"

+

"liow will I use it?"

1:30 PSI {

TilR Introductory Sections Altran Describe the following TBR Sections: Corporation e introduction '

System Descriptions Postulated Event Description Waterhammer Occurrence e Plant Experience with Waterhammers 2:00 PNI TIIR Section on Plan Development /Prioritization T. Esselman e

include PIRT and Uncertainty Discussion 3:00 PAI Break 3:15 PSI TIIR Information on Column Closure Waterhammer Altran Testing Result Update Corporation Comparison to Analysis Update Methods of Characteristics Update

. Scaling How is it applied?

. Plan for Completion 5:00 PSI Adjourn

Preliminary Agenda for the Third Meeting of the Waterhammer Expert Panel Second Day,4/14/99 Time Topic l 8:00 A51 Utility Caucus Utilities 9:00 ANI Discussion Session Expert Panel. All 9:30 ANI TBR Section on Condensation Induced Altran Waterhammer Corporation Test Result Update Plan for Comparison to Analysis and Status

• Scaling Pian for Close-Out i

10:30 ANI Break 10:45 A31 TBR Section on Wave Propagation Altran

. Sonic Velocity Corporation Fluid Structure Amplification Fluid Structure Attenuation Wave Reflection 12:00 Lunch i

1:00 PSI Support / Component Reactions Altran Comparison of Test and Analysis Corporation 1:45 P.\1 Support / Component Qualification T. Esselman e Tying it all together 2:30 PAI Break 2:45 PAI Caucus All 3:45 PAI Daussion and Comments All 4:15 P31 Summary / Upcoming Work Plan Review /Aleeting A. Singh/V.

Schedule Wagoner 4:30 PAI Adjourn l

t.

1 Attachment 3 i

MEETING HANDOUTS i

l

Technical Basis Report Overall Objective Provide overall description of systems and events Provide technical information that describes the waterhammer events

• Respond to RAI questions

• Provide each utility with information that will allow for demonstration of the conservatism of a prior analysis

~

• Or, provide a roadmap for the calculation of the waterhammers and the reaction loads 1

Outline of the TBR 1.0 Introduction '

l.1. Purpose 2 1.2. Scope 2 2.0 System Description 4 2.1. 13rief Description of Typical Fan Cooler System 4 2.2. System Data 5 3.0 Postulated Event Description 6 3.1. Event Time Line 6 3.2. Single Active Failure 8 4.0 Waterhammer Occurrence 9 4.1. Condensation Induced Waterhammer 9 4.2. Column Closure Waterhammer 11 5.0 Plan Development 1 13 6.0 Plant Waterhammer Experience 14 6.1. Description of Tests 14  !

6.2. Description of Other Plant Events l 14 7.0 Condensation Induced Waterhammer 14 7.1. Analysis 14 7.2. Testing 15 8.0 Column Closure Waterhammer 15 8.1. Analysis 15 8.2. Testing 15 9.0 Pressure Pulse Propagation 15 10.0 Sysam Loads / Qualification 16 11.0 Acceptance Criteria ,

, 16 1 12.0 Application to Other Hydraulic Events 16 13.0 Conclusion 16 14.0 References 17 Appendices Appendix A Survey Responses Appendix 13 Phenomenon Identification and Ihnking Table (PIRT)

Appendix C Column Closure Waterhammer Test Description and Data Appendix D Temperature 130undary/ Column Closure Waterhammer Test Appendix E Condensation Induced Waterhammer Test Description and Data 2

1

\

Introduction

• The objective of this TBR is to provide information sufficient to allow a utility to close out these issues for their plants.

l

. The information is also intended to be useful in evaluating other hydraulic transients. )

. Describe the Generic Letter 96-06 and the subsequent RAIs.

i I

l 3

I l

i

System Descriptions

• Typical system descriptions will be provided that apply, as a minimum, to all participating plants.

• Variations will be described for :

- Open Loop Plants

- Closed Loop Plants e Summary of the surveys will be provided.

4

Postulated Event Description

• Provide insight not only to the occurrence of the LOCA and the LOOP but also the effect of single active failure.

• Describe acceptable methods for the evaluation of single active failure. Describe the failures could make the postulated event more severe.

1 h

i 5

Waterhammer Occurrence

. Describe the events:

- pump and fan coast-down

- draining

- boiling

- pressurization of the piping

- draining of the piping (open loop systems)

- eventual restart of the pumps.

e.For each " event", the fluid and thermal behavior of the system will be described.

• The potential for waterhammer occurrence and the specific plant; conditions required will be ..

described.

• The potential for occurrence of other waterhammer events will be described and, when appropriate, justification for discounting these events will be provided.

6 l

l

R Plant Experience with Waterhammers

• The experience of plants with column closure waterhammers during normal Loss-of Offsite-Power tests (open loop plants without LOCA) will be described.

• The quantitative data recorded will be provided, but it will be emphasized that no plant has experienced any loss of pressure boundary during a LOOP and the subsequent column closure waterhammer.

l l

1 7

Mechanism Prioritization and Plan .

Development

. The identification and prioritization of phenomena associated with a combined LOOP and LOCA will be described.

• The phenomenon identification and ranking table (PIRT) and the Uncertainty Analysis (UA) used in this development will be included.

• The Plan developed to address the 96-06 issues will be described. -

• The logic used to develop the scope of the TBR will be presented.

8

{

Calculation of the Waterhammer Magnitude Prior to Pump Start

. The waterhammer that may occur prior to pump start is the condensation induced j waterhammer.  !

l e Existing laboratory test data, calculations,

{

flow regime maps, and new testing will be described.

l

= Methods to calculate the waterhammer l

magnitude and the bases for the calculation of l the pressure magnitude will be provided. ..

I l

9 a

Calculation of the Waterhammer Magnitude Following Pump Start

• The waterhammer that may occur following pump start is the column closure waterhammer.

• Existing laboratory test data, calculations, and new testing will be described.

• Methods to calculate the magnitude of this  ;

waterhammer and the bases for the calculation l will be provided. l

• The information that will be supplied will -

l include all the information necessary to l

determine that a prior calculation is conservative or to calculate the magnitude of the waterhammer.  !

10

I Pressure Pulse Propagation

• The propagation of the pressure pulse in the piping system will be described including sonic velocity, pressure attenuation, and fluid-structure interaction.

Methods and the bases for the methods will be provided.

11

i Reaction Loads on Supports and System Components

. The reaction of the pressure pulse into supports and components will be described.

• The information supplied will include that necessary to determine that a prior calculation is conservative or, if necessary, to calculate the reaction loads.

i i

i 12

Acceptance Criteria e Typicalinformation contained in individual plant FSARs will be described, but a specific plant criteria will be dependent on the FSAR.

• Guidance into the application of these stress criteria to the results of the waterhammer analysis will be provided.

13

Risk Informed Implementation

. The potential for use of risk assessment for the assessment of the combined LOOP and LOCA ,

i will be described.  !

• Guidance into the performance of a specific plant assessment will be provided. ,

l I

14

l TBR Introductory Sections i

. There are several open items from the sections entitled: l

- Introduction

- System Descriptions

- Postulated Event Description

- Waterhammer Occurrence

- Plant Experience with Waterhammers W

4 1

What Waterhammers Do We Have?

e Condensation Induced in horizontal lines during draining.

. Column Closure during refill.

I

)

2

What Waterhammers Don't We Have (or Have Insignificant Effects)?

. Fluid impact at changes in direction or changes in resistance - not significant

• Condensation Induced waterhammer in the fan cooler - Froude number greater than 1.0 and tubes hot

. Velocity change when hot water passes orifice or flow control valve and flashes -loads not I significant

. Condensation induced waterhammer in horizontal lines during refill - Froude number greater than 1.0 1

. Condensation induced waterhammer in 1 vertical lines during refilling - Froude -

numbers greater than 0.67 in vertically down piping

. Check Valve Slam in pump discharge -

coastdown slow

. Valve Closure (if MOVs or AOVs stroke once flow reestablished)- closure times slow

= Others?

3

Plant Experience

. Plant experience tells us that LOOP only events have never breached or challenged the pressure integrity of the piping or components.

. Quantitative pressure measurements are on the order of the magnitudes of the waterhammers anticipated. j

. Lack of specific data, configuration, instrumentation details, and system operating parameters makes it difficult to utilize the data for direct comparison to analyses. .

• Results of testing will be provided and a l qualitative comparison of the measured l

pressures to those expected will be provided. l 4

Plant Specific System Data Additional Data is Required

- Draining Velocities

- Others??

I 5

Qualification Steps

'rescriptive Beginning anc. Enc.

• Event Definition

- LOOP and LOCA events assumed to occur concurrently and to begin at the same instant in time e Support / Component Qualification

- The resulting loads on piping and supports should be evaluated using j

" currently accepted methods and acceptance criteria" 2

1

Support / Component Qualification

" Tying it all together"

. Items to be qualified structurally

- Piping

- Supports

- Nozzles

- Fan Cooler Tubes

- Anchor Bolt / Foundation

. Qualification Steps

- Event Definition

- Event Combination

- Thermal / Hydraulic Conditions l

- Pressure Transient Loads (magnitude and duration)

- Applied Loads in the Pipe (APs in pipe segments)

- Bending Moments, Axial Loads, and Loads in Supports

- Define Load Combinations

- Define Acceptance Criteria

- Support / Component Quaification 2

How Can We Achieve " Appropriate Conservatism"

)

. Areas to be Considered

- Waterhammer Pressure Transients (magnitude and duration)

- Applied loads in the pipe (APs in pipe segments)

- Bending Moments, Axial Loads, and I Loads in Supports

. Doing this effectively requires an assessment of the Certainties and Uncertainties 3

]

i l

Assessment of Certainties and Uncertainties

. Start with a clear determination of how the systems really performs, then define the

" appropriate" conservatism.

• Risk considerations should be a part of an effective integrated program.

• Individual plant considerations of LOCA and LOOP would show that the probability ofloss of SW/CCW function following the LOOP and LOCA does not need to be extremely low to meet the regulatory guidelines.

I 4

I Stochastic Factors Associated with Loss of Pressure Boundary 1

l

. Some plant parameters e Magnitude and duration of the waterhammer e Pressure attenuation in pipe e Calculation of the stress in the pipe and loads in the supports e Failure occurrence given a stress in t:1e pipe e Others?

5 i

1 L

1 i Pressure Boundary Integrity I

l 1

e To achieve pressure boundary integrity, two failure mechanisms need to be evaluated for the '

effects of waterhammer loads.

l

- The capability of the pipe to withstand the l' overpressure spike produced by the waterhammer event. 1

{

l

- The capability of the pipe and supports to j withstand the dynamic effects induced by the travelling shock wave.

e Structural response predictions of a piping system caused by short duration pressure spikes are almost universally performed using elastic time history l analysis methodology or an equivalent elastic static method. These methodologies tend to over-predict the pipe stress and support loads.

l l

l 1

6 i

Proposed Roadmap Generic Activities e Assess certainty and uncertainty to define methods to treat stochastic variables e Define methods to be utilized by a plant to obtain necessary plant parameters, including those that vary (air-in water)

. Define methods to calculate air that will be released during boiling

. Define method to utilize plant parameters to determine appropriate waterhammer pressure and duration

. Define acceptable method to account for pressure attenuation through the pipe

. Others?

7

o i

Plant Specific Activities

. Plant specific thermal hydraulic analysis to achieve closing velocities

. Plant specific analysis to achieve air released during boiling

. Calculation of waterhammer pressure magnitudes based on Joukowski and appropriate mitigation factors that depend on air content and void size

. Calculation of waterhammer durations based on system geometric considerations

. Calculation of attenuation of the pressure wave through the pipe (if desired)

. Qualification  ;

- Linear elastic structural dynamic analysis followed by evaluation of component ,

stresses, pipe stresses, and support loads using FSAR-defined load combinations and acceptance

- Comparison ofloads to those in successful plant tests with larger loads 8

I

I -

Table 1: PIRT Component Draining Phase Refill Phase -

l Phenomenon Phenomenon 1 Pumps 1.1 Coast down 1.2 Restart 1.3 Head 2 Check Valve 2.1 Valve slam 3 Piping f;/f,.

3.1 Roughkss 3.2 Layout 3.3 Ficw Resistances 4 Horizontal Piping 4.1 extent of void 4.2 pipe heat capacity 4.3 external heat transfer 4.4 internal heat transfer 4.5 draining flow regime 4.6 air accumulation 4.7 refill / impact loading 4.8 refill flow regime 4.9 released air 4.10 pressure wave attenuation 5 Vertical Piping 5.1 extent of void 5.2 pipe heat capacity 5.3 external heat transfer 5.4 internal heat transfer 5.5 draining flow regime 5.6 air accumulation

' 5.7 refeil/ impact loading 5.8 refill flow regime 5.9 released air 5.10- pressure wave attenuation 6 Fan Cooler coa 5dd e 6.1 heat transfer 7"" ' "'" "Mb*'

6.2 inventory /dryou.

6.3 heat capacity 64 ttapInnggeq_ y 6.5 l%([,f c. l - ef -cloA/ filling pattern

-o 6.6 # pressure drop 6.7 carry over 6.8 pressure wave attenuation 6 Control Valve

,, 7.1 resistance 7.2 effect on pressure wave

6 I

' Nuclear Engineering 2

ELsEVIER Nuclear Engineenng and Design 186 (1998) 23 -37 and Desinn 3 The role of the PIRT process in experiments, code development and code applications associated with reactor safety analysis s

' i Gary E. Wilson ", Brent E. Boyack *

• IJak %atoonal Demarsag and Enuronmental Laharatory. Idahn falls. ID 8.14l5. USA l

a

• Los 4Lnws Natwnal Laboratnes. Los 4Iamos. NM 97545. l'S4

\

Recened 27 November 1996, aaepted il December 1997 l

l i

Alstract in September 1988. the United States Nuclear Regulatory Commission issued a revised emeriency core cochng system rule fer hght water reactors that allows. as an opoon, the use of best estimate plus uncertainty methods m safety analysis To support the 1988 hcensmg revision. the Umted States Nuclear Regulatory Commission and its contractors developed the code scahng, apphcabthly and uncertainty evaluauon methodology to demonstrate the feasibihty of the best estimate plus uncertamty approach. The phenomena identificauon and ranking table (PIRT) process. Step 3 m the code scahng. apphcabihty and uncertamty methodology, was ongmally formulated to support the best estimate plus uncertainty hcensmg opuon Through further development and appucanon, the PIRT process has shown additional unhty as a robust means to estabhsh safety analysis computer code phenomenological requirements m their order ofimportance to such analyses The generic PIRT process, mcluding typical and common tilustranons from pnor apphcations that promoted further development of the process, are desenbed Analysis of the I results of the pnor apphcations is also desenbed. The analysis results provide information that can help guide future ,

appl. canons of the process in a graded approach based on phenomena relative importance. C 1998 Elsevier Science '

SA All nghts reserved.

1. Introduction key feature of this licensing option relates to quantification of the uncertainty in BE safety in September 1988, the United States Nuclear analysis and inclusion of this uncertainty in the Regulatory Commission (USNRC) issued a re- determination that a nuclear power plant (NPP) vised emergency core coohng system (ECCS) rule has a ' low' probability of violating the safety (USNRC,1988) for hght water reactors that al- cntena specified in 10 CFR 50. To support the lows, as an option, the use of best estimate (BE) 1988 hcensmg revtsion, the USNRC and its con-plus uncertamty methods in safety analysis. The tractors developed the code scaling, apphcability and uncertamty (CSAU) evaluation methodology

• corresponding auihor Tel - + t 20s 5269511. fat +t to demonstrate the feasibility of the BE plus 20s 5:66971 c.masi gewAinel gov uncertainty approach (Techmcal Program Group, 0029 5493 98 5 - see front matter C 1979 Elseuer Saence 5 A Ait nghts resersed P/I S0029 549h48 >00:16 7

24 G E H%n. 8 E Bmad % clear Ertmeermt .und Desirn 186 (1998) 23-37 1989, 1990). Tlie CSAU is desenbed in detail m 2. PIRT process description the references and the total methodology is not discussed further here. The phenomena identifica. The mformation obtamed through the applica-tion and rankmg table tPIRT) process (Technical tion of the PIRT process identifies the require-Program Group, 1989, 1990, Shaw et al.,1988), ments that will be imposed on analytical tools Step 3 m the CSAU methodology, was ongmally used to simulate accident scenanos. In addition, formulated to support the BE plus uncertainty those requirements are pnontized with respect to licensing option as executed m the CSAU ap- their contnbutions to the reactor phenornenologi-proach to safety analysis. Subsequent work has cal response to the accident scenano. Because it is shown the PIRT process to be a much more not cost effective, nor required, to assess and powerful tool than conceised m its original form. examine all the parameters and models in a best Through further deselopment and application, the estimate code in a uniform fashion (Technical PIRT process has shown itself to be a robust Program Group, 1989, 1990), the methodology means to establish safety analysis computer code focuses or those systems, components, processes phenomenological requirements in their order of and phenomena that dominate the transient be-importance to such analyses. Used early m re- havior, although all plausible efTects are consid-search directed toward these objectises, PIRT re- ered. This screening of plausible phenomena, to sults also provide the techmcal basis and cost determme those which dominate the plant re-effective organization for new experimental pro- sponse, insures a sufhcient and efficient analysis.

grams needed to improve the safety analysis codes PIRTs are not computer code-specific, that is, for new applications. This synergistic use of the PIRTs are applicable to the scenario and plant PIRT process has been repeatedly demonstrated design regardless of which code may be chosen to in applications to BE plus uncertamty analysis of perform the subsequent safety analysis. This also a small break loss-of-coolant accident (SBLOCA) adds to the etEciency and generality of the pro-m a current generation NPP (Ortiz and Ghan, cess. A typical application of the PIRT process is 1992), development of expenmental programs and conceptually illustrated in Fig. I and desenbed m safety analysis requirements for two production subsequent sections.

nuclear reactors (Hanson et al.,1992; Wilson et al., 1992a), a proposed new research reactor 2.1. Define prob /cm (Wilson,1992b), two proposed advanced light water reactors (LWRs) (Wilson et al , 1996, All PIRTs have a common basis in that they 1997a, Boyack,1995; Kroeger et al.,1995) and are deseloped to address plant behavior m the support to resolution of a boiling water reactor context of identifying the relatise importance of IBWR)licensmg issue (Wilson et al.,1997bt Van- systems, components, processes and phenomena ations on the PlRT process have also been in- in dnving the plant response. However, details of cluded in several intemational studies related to PIRT development may vary depending on the BE plus uncertamty analysis (Glaeser et al.,1993, specific problem to be resol ed. In pnor practice, Hoffer,1990, Holmstrom et al.,1994). P!RTs have been directed toward both:

The pnmary purpose of this paper is to desenbe (1) research onented more to code development the genenc P!RT process, including typical and associated with design confirmation; and common illustrations from pnor applications (2) activities directed more to resolution oflicens-(Section 2). The secondary objective is to provide ing issues (Wilson et al.,1997b). Accordingly, it is guidance to future applications of the process to important to determine and define in some detail help them focus, in a graded approach, on sys- the specific problem for which resolution is de-tems, cornponents, processes and phenomena that sired. Further, it will esentually be necessary to base been common m several pnor applications develop an early match between the level of re-(Section 3). The paper is summanzed and con- sources available and the level of efTort to be cluded in Section 4. expended in the PIRT development. It is fool-I l

I

1 1

. G E. Hibon. 8 E 80 sack i Mclen Entmeermt and Design 186 (19981 23-37  :$

r 3 r r

) 3 r 3 r 3 DeGae Denne PIRT --e=-

+ 0** Define problem Objectives % Potential potendal parameterts) 4 Pl ant designs scesanos oflaterest (1) (2)

L J L J Q J v N ) q @ j p

r 3 ,

1. Identify ' # I '

Partition ' # Parution ' btal $

, pta usible plant design scenario Denne high review all phenomena + into + into  % level basic

• available by phase & e composeats e conveniest e system ,, ,,,g,,

composest (subsystems) time phases Processes & analytical q (10) j g (9) j g (8) j O j data L @ J P

I 3 I T F T F '

Develop r 3 Finalize aad Apply AllP Perforts g ;,, g , to determine Revtew Al{P selected document rsassing of phenomena results; PIRT PIRT for component A %

nWu modify if

• connrasuos  % nNut j phenomens seasidvky mah &

"D ports au nquind importance studies Plant 4:g3 (12) (13) designs L J k ) ( j g (14) j g

I (I5)

Ranking approaches, other than AllP, are available Fig i illustration of :pcal apphcatson of the PIRT process.

hardy to set out to develop a broad based, highly This function is, therefore, the basis for the detailed P!RT with insufficient funding, particu- PIRT development, that is, to serve as plant larly if the chosen level of P!RT detail is not performance indicators. Further, but subsequent required for problem resolution. Thus, it is neces- I sary to derine the mmimum achiesement that has Pinis have one pnmary and trua adjunct a high probability of effecting problem resolution. funcuons depending on coniert of use That is, specifically rust what problem must be SET &IET i resolved and to what level. The references noted + tspense. cal )

abose can serse as guides based on pnor G"'d***'

expenence.

Fiast Code 2.2. Defuse PIR T objectnes terformance  : Devesopment ledica:ar Caidance The objective of a PIRT deselopment is assis tor rtRT strongly oriented to the intended use of the PIRT. d'"8 P'"*"' co4, PIRTs have one pnmary and three adjunct func- + Uecertaiory C "'d'""

tions depending on context and use (Fig. 2).

The primary function of a PIRT is to define U.c of rini plant behavior in the context of identifymg the PRIMARY ADJUNCT relatise importance of systems, components, pro- FUNCT10N USE l cesses and phenomena in dnving the plant re-sponse. Fig 2. Conceptual use of PIRTs i

l I

l i

l .

1

's <. I' H dm. o I. Umad wirar tnumarms and veuros In ot% lt u to the primary purpose, a PIR 1 may serse one or sentation of the containment feedback as a 3et more adiunct lunctions IFig 2) That n. to of boundary condinons should he demon-proude guidance m estabhshmg the requirements strated to be satisfactory before such decou-

m. phng n employed This requirement arnes e <parate and miegral etfects (SEl. lE l i esperi- because ot the gravity drnen ECCS conduton mental procrams. where the objectne is to help that exists m many of the passne plant designs, msure the experimental data fully retlest what may he expected m the piant; 2 -1 Dc/mc potentia / << cmu me e code deselopment and unprovement. whcie the objett ve n to help msure the code o capable of The relaine im port a nce of phenomena , pro-modehng the plant twhavior, and cesses n scenano dependent Although there are e code untertamt) quanohcation, where the ob. eseral systemesubsystems that aie actae m all lettne is to help msure the urious contnbutors plant deugns m all scnanos, full undentandmg ta notertainty are identihed and treated m a of the plani hchauor .dso depemis on other sys-manner a pps opr ia te to their importance to tems subsystems ihai are speahc to a particulai plani beha uor .ind. t h u s, io the oserall atodent scenano m a particular plant deugn. And una n.nnty .n already noted. n n the combmahon of plant Imphcanons of the abmc three tiems are f urther speafic designs and scenano speutic features ihai dcwnbed m the rankmp defmioon dntucion in udt dwtate the system subsystem cinclop that Nettion 211 must be tenudcied m a PtRT desclopment. Wre Jet.nled guidance m this regaid b .n.nlable m the 2 / A /m. v.m nnol pl.un ./. ucrn R ef erences lhe t ela t n e n npot tance of phenomena pro- 2 5 /)c/mc parameres t si n/ miceco s ews n pla n t deuen dependein to sarymg de-a ces t hus. H n netessars to estabbsh the NPP T his sicp m the PIRT piotess de.ds a nh estab. {

coselop to winth the PIR IN wdi apply. Thn step inhmg the pnmary evaluanon entena chai will be m the pmcess muu be coupled wnh the scenar- used to iudge the relatne importance of phenom-not s) selettam. dneussed m the followmg sectmn, ena piocesses m the plant behanos of mieret to estahhsh those pl. int systems Nuhsystenn that Prnnary evaluauon ontena f or cntenon) are nor-sull pi.n a ugmhtant pan m the plant beh.nior ol mally based m regulatory safet) s egmremenb mierest i orther detads m ihn regard h.ne been such as those related to restneuons m peak dad dot umented m the Reterences. That experience temperature (PCTL hydrogen generatmn. cic. t he i mdicates two general hndmgs rank of a system, component, process or phenom-e P!R 1 s aJdresung postulated acadent t ra n- ena is a measure of its retaine miluence on the aaib iLUIJ DCA. SitLOCA. SG I R. etc ) m prunary entena (entenon) In t IS apphcanons ol cuump plant designs can often ignore the con- the PIRT process, pomary ev.dnanon unena are t a nnuen t. t he torced ECCS miection m these often denved from the regulatory requirements of plaub normally allows decouplmg of the con- 10 Cf R 50. as dlustrated m Fig 3 i.unment f rom the semaining plant systemx rhe columns m Fig.1 aie explained as f ollows ,

t hat n. the feedback Irom the contamment to e Lesel: an arburary designator for the purpose l ihe primary. secondary and ECCS systems can of diwussion, but related to ihe degree el deiad be sa t nfactonly replesented by Imundary m the mdicated safety requirement tsee uiteria condinons below L e PIR I s addressmg postulated acudent tran- e Source: the documented source ol the unhcated uenb m the ne4 passne plant designs should salety requirement.

consider the intesactions between the contam- e Cnteria: the cnteria for sale operation ot a ment and the semanung plant systems R epre- reactor. At Lesel I this n a sery general. rathei l

i l

l

e e .

G E Wolwn. 6 t Bnack % clear ljitmerrrat .ind Urut*i 1R6 t 199M) 2 t 87 2' Conser- Charac-t eved Source Cniens vousm tensue 1* t oCFR 1 11 Proted put*c healh & safety ( ,gg gag,gy 2' 10C FR 100 t.mt fisseon oroduct r**== lematung y App,widur A t mt fued fasture Lamn RCS bream tarrut contasnmeru bream j( h SRP 6 2 Limn containment Contaart. P & T, leakage.

hydronen. etc DNBR (PWR), RCS & steam SRP 15.14 MCPR (BWR), energy systems 4* to 15 61 deposstion, fuel term &

M OCA immn <<=<unn eram nr** = nee

, 10CFR PCT. omdation.

So 46 & hydrogen generatsort SRP 15 6 5 tong term coohng, tma main n.wn, 5 R Vessel anventorf" 6 Intenm trnportant PtRT narameters 7 Intenm Dommant 3f pts:7 processes Y 8 trm wn Ranked More PIRi phenomena limiting Behavior

• Le,es i . 4 se ennemnes e senc copsseans or =casany gunsnom Lew s e may snare respanne cresnr.a imod as satoCA csAU ennemerns a Fig 3 Illustration of the basis for development of pnmary evaluation entena ISBLOCA example) non speci6c, statement that the public will be parameters are likely (for example LBLOCA), they ,

protected. However, as one progresses down- become pnme candidates for selection as pnmary ward in the table, the enteria becomes more and evaluation entena. In practice, if it can be shown j more specific as to conditions that will insure that the probability of PCTs above 982*C (1800'F) l safe reactor operation. Levels i through 4 relate are low, then clad oxidation and hydrogen genera- i to specifications provided m the Code of Federal tion are insigmficant and these parameters hase Regulations, that is, " requirements m law" reduced usefulness as primary evaluation critena.

Levels 5 through 8 are denved from the require. While PCT, clad oxidation and hydrogen gener-ments of Lnel 4 in the context of creatmg useful ation regulatory limits remain in force for I PIRTs. SBLOCA, analysis of this class of transients are e Conservatism: an mdicator of the degree of more normally directed toward insuring the fuel control of safe reactor operation imposed by the rods remain essentially covered by a two-phase entena. mixture. Thus, it is common to use minimum core e Charactenstic: an indicator of the type of the inventory, or a related parameter such as onset of safety requirement imposed by the cnteria. At significant voids, as the primary evaluation ente.

Level I these tend to be broad, general state- non. The onset of signi6 cant voids may be inter-ments related to safe reactor operation. How- preted as transition from single phase liquid in the ever, as one progresses downward the entena core to a two-phase mixture with sufficient vapor become more and more statements of allowable content to signincantly inhibit the cooling of the reactor behavior. fuel rods. PIRTs developed for containment anal-PCT, clad oudation and hydrogen generation ysis codes may use the regulatory requirements of are speci6cally addressed m the regulatory require- Standard Review Plan (SRP) 6.2 as a basis for j ments. Therefore,in accident transients were these establishing the primary evaluation critena. J l

f l

l l

8 G E WJwn. 8 E B.nack % clear Engmeermg and Des,g, ggs ggngg ;).37 2 o Idents, obtam and renew all aiariable e The team members should hase a demon-experimental and analytical data strated capabihty to work in a team environ-ment, particularly the abihty to tuppress Past expenence shows that PIRT deselopment individual agendas in the mterest of the team is best accomplished using teams having broad objectn es.

base knowledge and expertise m relevant expen- e Experience has repeatedly shown that teams  !

ments, system analyses and plant operations. Fur- composed of more than six individuals become ther, because most PIRT process applications increasmgly less efTectne as the number of address state-of-the art problems having fmite re- members increases. Howeser, the effectiveness sources, it is necessary the team exercise its collec-of the team increases with the degree of techni-tive engmeenng judgement (particularly in the cal support staff available to the team. Assign-initial development). Accordingly, the team must deselop a collectne knowledge base that repre-ment of the team coordinator role to an individual having direct access to technical sup-sents the state-of-the-art understanding of all fac-port staff is an effective technique. Hershe ' hen tors apphcable to the problem of interest.

can call upon the technical support stafT to Therefore, as early as possible in the PIRT pro-cess apphcation, all relevant information should execute the details of studies needed and be identified, obtamed and resiewed, individually de%ed by the team m the PIRT development and collectnely, by the team. Sensitivity studies-a; - confirmation. The team coordinator howeser hmited, hase. also prosed worthwhile t should also have demonstrated skills in inter-help better determine the relatne mfluence of personal relations in the context of etTectively separatmg the wheat from the chafT in the phenomenon identified early in the PIRT develop-ment. Such mformation may come directly from vanous individual team member inputs, with-expenmental data analysis and or code simula- out suppressing the mdividual innovation of tions of the expenments and of the NPP: transient those members.

,g ,.or & PIRT d im g k of mterest. Otten this type of analysis can chmi-nate phenomena by showmg it does not exist, or represented directly or indirectly on the team.

is insigmficant. Of more importance, the effectiveness of the Expenence has shown success in developing a team increases pr portional to the degree of PIRT, in a suflicient and cost effective manner, is technical expenise, access to customer manage-strongly dependent on the composition of the ment and program management skills pos-team. Team attnbutes that. promote success in- sessed by the customer representative.

clude the following.

e The collective expertise of the team should be 2.7. Defne high letel basic system processes extensne m depth. Team members should have extensne and current knowledge in their field The mitial development of the PIRT process of expertise, as may be reflected m mdividuals and the first several process applications, fociued havmg well deserved mternational reputations. pnmarily on the relative importance of compo- -

e The collective expertise of the team should be nents and phenoniena. However, in more recent broad based. The team make-up should have at applications (Boyack,1995; Wilson et al.,1997b) least one member having expertise appropriate the efficiency ofintroducmg the identification an6 to the problem ofinterest in each of the follow. ranking of high level system processe nas been mg fields: expenmental programs, code deve;. introduced. Specific identification of high level opn ent, code applications (safety analysis), system processes has proven most valuable in plant operations and PIRT development efficiently organizing and focusing the subsequent methodology. There should be a clear team identification of the components and phenomena coordinator, preferably the team member hav- that shovd be considered in the PIRT deselop-1 ing expertise in PIRT development methodol- ment. Further, perspectises of the relative impor-

! ogy. tance of the high level system processes is a t

l I

y .. . . . . . . . . _ .. .. . . . . . . .. _ . _ . . . . . . _ . . . . .

s G E. Wdson, B E Basack i helear Entmeerwr and Des >en 186 (1998) 23-37 :9 significant and m ranking the components and gories: 1) those local (within) to a componente phenomena. Often it can be shown a specific high subsystem and 2) higher level system interactions lesel system process is of httle consequence m a (integral) between components / subsystems. Thus, particular time phase (Section 2.8), thus, further partitioning of the NPP mto components /subsys-consideration of the associated components and tems is a sigmficant aid in organizing and ranking phenomena is not necessary; a significant resource phenomena / processes. Component partitiomng savings. High level system processes that have common to prior PIRT process applications prosed useful in pnor PIRT process applications include:

include:

Accumulators Break Cold legs Depressuruation Inventory reductic'1 Core Downcomer ECCS Inventory replacement Short-term dynamic Fuel rods Hot legs Pressunzer core cooling Pumps Lower Upper plenum Long-term evaporause Gasivapor plenum core cooling transport Steam generator Separator Other unique W ater Suspended water (PWR) (BWR) systems depletioniaccumulanonis transport urface transport For PIRTs directed toward analyses in which Debns transport Debris depletion there are strong interactions between the primary systems and the containment, additional compo.

• Y # #

2.8 Partstson scenario s:sto contement time phases Containment Containment Surnp The re!ative importance of phenomena is time dependent as an accident progresser Thus, it is  %,etwell Suppression Vent (BWR) conscruent to parution accident scenanos into time phases m which the dominant phenomena! E ' I processes remam essentially constant. All, or ]

nearly all pnor applicauons of the P!RT process i to LBLOCA have partstioned those transients 2.10. Identtfy plaussble phenomena by phase and into blowdown, refill and reflood ume penods.

This partitioning has a basis in the regulatory ###"##"#"'

requirements. Other time phase partitioning m As n ted m. Section 2.6, PIRT development is prior PIRT process applications have included:

best based on the collective expertise of broad e SBLOCA in conventional PWRs: High pres-based teams and in the initial rounds tends to rely sure blowdown and long-term recovery e Most transients in advanced PWRs: High pres- n the use of engincenng judgement. The some-sure blowdown, ADS blowdown, short term what subjective nature of the process at these

, recovery and long-term recovery Points in time introduces the question of com-e Most transients in advanced BWRs: Pre-isola. Pl eteness. That is, 'how do the team members tion, isolation, depressuruation (ADS), short, discover what they do not know' with respect to ,

term recovery (refill) and long term recovery expanding state-of-the art knowledge, ne best 1 techruque discovered, to date, has been to formal- l 2.9. Partstion plant design into components ize the process into two substeps:  ;

e First, identify ' plausible' phenomena! processes  ;

From a modeling perspective, phenomena / pro- with stnct adherence to not allowing ranking l cesses important to a NPP response to an accident discussions during this substep. In efTect this l scenano can be grouped in two separate cate- becomes a ' brain storming' session (s). In this i i

I

, i l

i l

l

e e 30 G E It dwn. B E B,nack .%ucieor Engineerune and Desic., tu tI9% li 17 context, pi.iusible phenomena processes a re High, phenomenon should be exphcitly ex.

those that hne wome >igmficance to the plant hibited and well measured; phenomenon behauer Only after all team members are sat- should be prototypical.

afied that the question of completeness can be o Adjunct use in the context of code modeling:

defended, should the work proceed to phenom- Low, phenomenon has smal! effect on the ena procen ranking as subsequently descnbed. pnmary parameter of interest. Phenomena should be represented m the code, but al.

2.11 Rank ,umponents and phenomena most any model will be sutTicient; unportan< c Medium, phenomenon has moderate indu-ence on the primary parameter of interest.

This step is the heart of the deselopment of a Phenomena should be well modeled; accu-PIRT. Fig. I indicates a three step process specifi- racy maybe somewhat compromised; cally reflecting the use of a proven subjectise High, phenomenon has dommant impact on decision making tool, the analytical hierarchy the primary parameter of mterest. Phenom-process ( AHP) (Saaty, !982). The use of this tool ena should be exphcitly and accurately is highly recommended to formalize subjectise modeled.

decision making into a product that is defensible, e Adjunct use m code uncertamty quantification:

scrutinizable and complete. Howeser, application Low, combined uncertainty of phenomena of the AHP is not without resource penalties and may be determined in a bounding fashion, or other less costly means are asailable Accordingly, may be ehminated when justified; the followmg descussion is focused on the genenc Medium, phenomena should be evaluated to features of component and phenomena ranking determme af uncertainty should be treated regardless of whateser aids may be used to help m individually as are high ranks, or m a com-makmg subjectise decisions. bined manner as are low ranks; The ranking process is dependent on the use of High, phenomena uncertamty should be in-the resulting PIR T(>) as already discussed in Sec- dindually deterrnined and then combmed tion 2 2 and Fig. 2. In the context of that pnur statistically with other uncertamty sources.

discussion, the simplest basis for rankmg of a Ranking scales based on just low, medium and phenomenon. process regarding its relative impor- high are attractive in that they the easiest to apply tance to the pnmary evaluation criterion is to use and have a simple numerical counterpart, i.e. I, a scale of low, medium or high. That is, w.th low; 2, moderate; and 3, high. However, it is often respect to the pnmary and adjunct uses summa- helpful to difTerentiate between the lowest of the rized in Fig. 2, the rankmg defirutions correspond low and highest of the high ranks. Prior experi-to their use as desenbed below: ence suggests a numerical ranking scheme of I-5 e Pnmary use as plant performance mdicators: is a better scale, with the following meaning: 1, Low, phenomenon has small influence on lowest of the low in importance; 2, low impor-pnmary parameter of interest; tance; 3, moderate importance; 4, high impor-Medium, phenomenon has moderate influ- tance; and 5, highest of the high in importance.

ence; The fmal product of application of the PIRT High, phenomenon has controlling influence. process is a set of tables (PIRTs) documenting the e Adjunct use with experiments that may provide ranks (relative importance) of phenomena and data for code development and validation: processes, by transient phase and system Low, phenomenon should be exhibited, but component.

accurate measurements and prototypicality are of low importance; 2.12. Perform selected PIRT confirmatwn Medium, phenomenon should be exhibited; senstimtry studies measurements may be denved; prototypical-ity may be somewhat compromised; As reflected in Fig. I and the precedmg discus-

F G E Wohan. 8 E Bma,k sclear Enemeerme and Deuen 186 Iv48) :). .it 3i sions. PIRT deselopment is an iteratne process to deselop genenc PIRTs by and large hase with significant feedback between the various proven to have limited usefulness. The move to-elements. At such time as may be deemed cost ward genenc P!RTs by:

efTectise by the deselopment team, key sensitiv- e lumping'seseral reactor designs together (for ity studies should be performed by way of adding example all PWRs);

additional con 6rmation of the validity of e 1 umping' several scenarios into a ' class of the PIRT results. These studies may be a natural scenanos' (for example all LOCAs); and

, continuation of sensitivity studies performed e denning phenomena at higher levels (for ex-early in the development (Section 2.6). Continu. ample heat transfer rather than the speci6c ing assessment of the moderately and highly modes of heat transfer),

ranked phenomena is of particular interest. has proven to be of limited value because the Snal results nearly always consist of everything 2.13. Finali:e and document PIRTfor subject being of high importance, or the phenomenolog-scenanos and plant designs ical ' classes' being to imprecise to be very useful to code development and/or improvement. Nev.

PIRT tables alone do not convey all the infor- ertheless, within 1imited objectives', features mation that will be needed by the developers common to all, or several prior PIRTs, can be and users of the P!RTs. It is emphasized that denved. Two 1imited objectives' have been se-excellent communication (documentation) of the lected for this paper.

PIRT process results is of paramount impor- e Identification of highly and moderately tance to the team members dunng PIRT devel- ranked phenomena that have been common opment, to the team members and the customer in several of the prior PIRTs (Section 3.1).

dunng peer review of the PIRT process results e Identification of the relative importance of and to the customer and/or other pnmary user components in prior PIRTs by an arbitrary 1 of the PIRTs, following deselopment. Experience normalization scheme based on the frequency strongly indicates success in developing useful of occurrence of low, moderate and high phe-PIRTs is a direct function of the degree to nomena ranks and weighted by ranking im-which supplemental products are well docu- portance (Section 3.2).

mented. Information typically important include desenptions of the plant (s), scenario (s), ranking 3.1. Phenomena ranks common to prior PIRT scales, phenomena and processes defimtions, process applications evaluation criteria and the technical rationale for each rank. Wilson et al. (1997a) and Boyack Table I summanzes phenomena / processes that (1995) are good ecamples of the level of docu. have typically been found to be of high impor-mentation that has proven successful. tance during vanous general reactor response pe-riods in the course of accident scenarios.

Similarly, Table 2 summarizes phenomena / pro-3 Typical results from prior PIRT process cesses that are typically of moderate importance.

applications The data rxluction, in both tables, is considered to be appropriate to the objectives of this paper, The degree of phenomenological detail con- but it is recognized the following data reduction tained in a PIRT and, thereby, the direct appli- procedures are somewhat arbitrary:

l cability of the PIRT, is directly proportional to e Phenomena / processes and their ranks that are the specincity of the operating envelop consid- common to only one plant design or one scc-cred in the PIRT. That is, PIRTs that convey nario have been omitted from the tables.

the most phenomenological informatio : tend to e Cells that are shaded indicate that the time  ;

be Sery design and scenans depeMent. Attempts phase of interest does not exist in the subject I

32 G E' Wdson. 8 E Bmack %dar Eagarermg and (Jeugn I86 glM8) 23-3 Table 1 Phenomena processes of typicas high irnportance in prior P!RT apphcations Erwwncy oflush rank throughout l

the rdant systems of mierest I 8 .c Z c )

g < j l g 5f 5!! }  !$g Raged beowdows Cntacal now

!slill!

I I I t i 2j

~

! !!z S I frorn high pressure Flashing i through large Subenbcal 6ow 12 1 1 4J 4 1 2 2j 4 breaks andor ADS Vo ding & distnbunon 10 3j 3 Condensanon 10 j 4 Entraanmenroe-entrainment i I 4 4 Convubw heat transfer '

I 2j i i

!nwntarvievel l I 2 Conductaan heat transfer 2 1 1 Decav heat {

1 lj_I l Phase separance 2 1j i Naturatintenal circulataan j i Evaporshon 1j i Stored encryv Parallel channel Gow i_ l 1

Energy transport bv 1. quad I _

Noncondensbie effects Core inventory Decev heat i i

]

I tj i reconry. enduding Conwchw heat transfer I i _

reGil and re0and Toew. phase Sowidettey $ l Ij i inventarvievel 4 2 i Cnncal Gow 4 1 1 2 Conductece beat transfer 4 I I 2 '

Dow remstancer fnetendfonn ton J l7 I Subenocal Sow 2 I I Small break slow Cntical Sow 2 I I depressuruanos Decay beat I toeg4 ens Subentical 6ow l 4 2 1 2j 2 recovery Noncondensbic e6ects 3 3 3 laventarv4evel 5 1j Convechve heat transfer 5 l now reestance/ 4 14 I

IJ l I I I

[nrtens form loss _

Chsation 4 i jj i Conduction heat transfer 4 Decav best I

IJ l 4

Naturafintenal circulat.on I

I] I 3 I I I

G E 54 tison. B E Smack % clear Engmeermt and Drutn 186 419981 '-37 31 Tabic 2 Phenomena processes of t>rwal moderate import.snce m pnor PIRT appsw.ations Frequetwy of snoderste rank thre Ebout l the rilant res of mt res l 4

h i

g i 8 s

i sag g 1

=

._ .,_ =

i s ag i i

!g m

g g s

i r s 8i.li Rapid tdowdows Conden sation d

24 2 I 7

4 7

7 f

frorn high pressure Flashmg 23 4 4 4 2 4 4 I tbmugh large Voeding & distnbution 19 1 1 6 2 4 4 I breaks and'or ADS Shwed eseergv 2 _

2 f 6 6 Sutznacal Sow 17 5 2 4 3 i Conduction bene transfer 16 2 2 a 3 3 2 Converove heat transfer 15 2 2 2 3 3 Nancandensitae effects 11 2 2 3 2 2 inetorvievel 9 1 4 J,jj i Fatrainmentdeetrainment 7 I 2 2 2 Thermal stratifn:auon 6 1 1 1 Natural /mtenal cerculanon 5 1 2 2 Cnucal Gow 4 1 2 i flow feestance/ 4 2 I I fncamTorm loss Tem-phase Gowidcha-p 3 1 1 1 Parafiel channel Sow 3 l l l CHF/drvout 3 2 Phase separation 2 1 1 Bailms 2 1 1 Care mventary Subcnocal oow I4 1 3 1 2 3 2 2 recovery,includ ng Convocave heat transfer 9 3 1 2 re611 and refkral Vending & distnbunon 8 2 l Conducuan heat transler 8 2 4 1 1 Nancondensible effects 7 2 4 I leventory4ewi 6 1 I I condemsanos 5 1 2 l I Countercurmni SowCCF1. 4 1 3 NaturaViotenal circulauon 3 I I Flashing . 3 I I I flow ressannos/ 2 I .I fnetionTorm loss Entrainment'de entramment 2 I I Stored energy 2 I I long e Condensation S 2 2 2 recovery Subentxas Oow 5 2 i Noncamdenable effects 5 I I I I I Inventarv4evel

]_ 2 1

ta l tl t li'dwn, H I Hea.i > Aartrar bremrerme and (Morn IRh tIMI 2i 47 i transient plant deugn, or b combmed mio another importanw. For certam scenanos and time phases l time phase shown m the table, or thal a PlRT did these components may and probably will have not exist for the time phase of mterest m the associated phenomena of high rank. The lesser indicated transient / plant design. relatne amportance m Table 3 is better character.

Because of the above factors, the reader is

! i7ed as meanmg these components have fewer high cautioned that the tables should not he mierpreted ranked phenomena for shorter penods of tirne.

as indicatmp the only highly important phenom. Thus, it should be recognized the normahzed rela.

ena,proccues m NPP responses to accidents. The tne component importance shown m the table is-imponance of phenomena / processes tends to be only of an approximate nature. For example the plant design specific. Thus, other phenomenal pro- ellort to achieve more commonalty in the data base cesses may also be highly important to specific by lumpmg a number of components in the listed j

designs. These tables only indicate the phenomenal IICCS component results in a large number of processes that are highly important in two or more entries m each of the rank categones llow, moder-plant designs, transients. ate and hight. This tends to inflate the normalized importance of the ECCS component. Similarly, 11 Rc/arne importantc of components m pnor because accumulators do no exist m all the designs P/RT pro,cu appheations hsted, the relatne importance of this component may be under-estimated. Nevertheles;s the data in Similar to the phenomena / processes importance Table 3 can be used as general guidance in future shown m Tables I and 2, the retaine mportance PIRT development. In addition, there is sufficient of :omponents m poor PIRT process applications data in Table 3 that a reader could refine the can be denved through appropriate data collection analysis, say to just PWR designs, sf so incimed.

and reduction technique >. Such miormation is provided m Table 3 based on the followmg data l reduction procedures: 4 Conclusion j

e The data focuses on components that are com-mon to the majority of plant designs in the data The PIRT process was onginally formulated to base (for example core, fuel rods, downcomer. support the BE plus uncertaintylicensing option as etc.). executed in the CS A U approach to safety analysis.

e To help promote commonalty in the PIRTs, Subsequent work has shown the PIRT process to l certam component groupmgs have been used. be a much more powerful tool than conceived in i

For etample seseral of the plant specific compo- its original form. Through further development and l l nents m the AP600 and SBWR, such as the apphcation, the PIRT process has shown itself to CMT, IRWST, suppression pool, etc. have been l be a robust means to establish safety analysis grouped into a single ECCS entry (see later computer code phenomenological requirements in comment). their order ofimportance to such analyses. Because e The data has been normalized to a scale of one it is not cost efTective, nor required, to assess and I

. to ten as shown in Table 4. examme all the parameters and models in a best The relative component importance given in estimate code in a uniform fashion, the methodol-Table 3 is a measure of the relative importance of ogy focuses on those processes and phenomena the indicated components in all the pnor PIRTs which dommate the transient behavior, although identified. For example, based on the total data all plausible effects are considered. This screening available, the downcomer, break and core compo- of plausible phenomena, to determine those which l- nents are shown to be of pnmary priority m the dominate the plant response, msures a sufficient i

_ plant behaviors. Similarly, the hot legs, pressurizer, and efficient basis for identifying needed code pump and containment exterior components are improvements and subsequent code development.

shown to be of lesser importance. Lesser impor- While the usefulness of PIRTs tend to increase tance should not be interpreted to mean of httle with their specificity to each particular apphcation 3

1(lj

• E * [ - r $ ',E { &4R'j7Q b k = h : 3 7,- ~ U 1 3 1 1 H

I I 3 3 2 2

@o I1a l M

. 4 5 3 2 2 I

I 5 I 2 4 4 1 H

n 3 I o 6 2 3 2 9 ea g M i t l

i 3 2 2 I 2 1 1 1 3 2 L

4 3 6 7 0 3 H 1 l

6 g 4 2 4 l 5s g@ M 1 l

6 3

3 I 3 a 2 L 1 2 2 1 4 3 al 7 0 3 H 1 l

6 9 21 2 4 I.j 1 g@ M 1 l

1 3 3 3 a 2 L 3 1 1 2 2 1 t

n 4 7 3 4 7 3 e

r H 1 de l 6 7 2 0 7 3 t

a c

ss g@ M 1 a

d 1 s 8 4 1 4 n

i I 3 t 2 2 1 is 1 2 1 1 2 6 1 I H

A.s l

e 4 2 7 7 9

. ss g M l 1

1 1 1 1 6 7 m l l

01 2 1 4 9 1 2 2 3 1 2 1 2 1 ta L o

T 8 1 5 1 1 2 1 1 1 I H

l I

3 9 7 4 2 2 2 1 3 3 4 1 5ag M I I

L 5

1 7 4 4 2 a 1 3 4 9 1 1

3 3 1 3 3 2 2 1 H

l s 2 1 1 3 2 1 n.

i g EI M l

t a 9 1 4 1 3 a L lp p 5 3 2 7 1 1 2 2 a H 1 I

T i 9 1 9 0 3 9 1 I 7 1 2 1 I R 3sg M 1 1 j-PI I i

A 4 7 3 2 n 9 1 2 5 1 5 r L l 1 1 i 1 o

n p 6 3 2 4 2 3 I 2 1 3 2 H .

n i

l l

I 2 1 R 1 2 1 1 t

n s <s5 ij M 1 1 _

I e A 6 R  ?

n L i 4 7 2 4 9 2 2 _

o p

1 t _

m o

c I{l o t a 7 i 7 K 4 A 3 1 1 1 1 2 1 f

o {ii) e t

n e .

a t

t n mT po.so m n s p

r o e n

duS mp d t n

e

- m t

n e

r o

o lcWune 3 a r e

t a

m i

p m nR S.as 5 I.

a.  :

- m n,

ea

., d a

a h S is ,

hr .

r n

u m

n, i a lu m

e o R ms k - n t .

. ta i, 3 r . e e i v

C C CMRT N pn c a e

e r

in a a o

t t

e R.

W n s

e e

n n

. n ,

o u

c u n n C-r a a u c _

bl ta l

a e ECPsa Bw F C a. r F I P U i

l P A _

TR _

36 G E. ishon. 8 E. Boiack , % clear Engmeermg and Design 186 (1993) 23 37 Table 4 Component normahzanon procedure for Table 3 was performed for the Ofhce of Nuclear Regulatory Research, US Nuclear Regulatory Commission (Componeni connahzed rank), - t Washington, DC 20555 under DOE Idaho Opera-

' (Weighted rank sum) -(Weighted rank sumi.,,,, _ tions Office Contract No. DE.AC07 941Dl3223.

• [(Weighted rank sum) ,-tWeighted rank sumi,,,,,,,

, g9 References where Boyack. 8 E.,1995 AP600 LBLOCA phenorr,ew idenunca-(Weighted rank sum). non and ranking tabulation. LA UR 95-27tt. les Alamos Nauonal bboratory, (distnbution at the ducret on of the USNRC)

. { Low ranks + 2 = { Moderate ranks + 3 Glaeser, H , Hofer E., Kloos, M., Skorek, T.,1993. Uncer.

tainry and sensitivity analysis of a post expenment calcula-

= V H@ rds

- uon in thermal hydraulics. SMIRT 12, Post Conf Sem.

No 15. Heidelberg, Germany and Hanson. R G , Wdson. G E., Ortiz, M G , Gngas. D P .1992.

Desclopment of a PtRT for a postulated double ended (Weighted rank sumL.. .,, are for the total data base gudloone break m a production reactor. Nuct Eng Des 136, 335-346.

Hoffer, E.,1990 ne GRS programme package for oncer-tainty and sensiuvity analysis Proc. Sem on Methods and Codes for Assessing the Off Site Consequences of Nuclear and tts objectnes, there are certam commonalties Accidents EUR 13013, Commission of the European Conununi in all such apphcations. The more important of g,; , Ge r, ,Yickett, A., wilson, G ,1994.

these common features have been desenbed m this Status of code uncertainty evaluauon methodologies, Proc.

paper in the context of typical results from pnor lot. Conf. on New Trends in Nuclear System nennohy-applications of the P!RT process. This information draulics. Pisa, Italy.

may be exFCted to help future applications of the Kr eger. P G et al,15/5. Prehmmary phenomena ident:6ca.

tion and ranking tabies for SBWR LOCA scenanos. Tech.

PIRT process, particularly m. identtfying typical: ,, cal utter Report W-6092 5 6,95, Brookhaven National e pnmary evaluation critena for judging the im- Laborsiory, June (distnbuuon at the discrenon of the portance of phenomenaiprocesses, including USNRCE their licensing basis; 0"'z, M G; Ghan. I S.,1992. Uncertainty analysis of nuni-e partitioning of accident scenanos into time mum vessel hquid mventory dunng a SBLOCA in a Bab-cock and Wilcon plant, NUREG'CR-5818, EGAG, Idaho.

phase; Saaty, T ,1982. Decision-Making For haders Belmont. CA.

e partitionmg of NPP designs mio components / Lifetime Learning Publicauons, Wadsworth Inc.

subsystems. ' Shaw, R A., larson, T K ., Dimenna, R K.,1988 Develop-

$ phenomena! processes ranking approaches, in- ment of a phenomena identincation and ranking table (PIRT) for thermal-bydraube phenomena dunns a PWR cluding the meaning of the ranks for guidance LBLOCA, NUREG/CR.5074, EG&G Idaho to expenmental, code development and code Techrucal Program Group,1989 Quantifyics reactor safety .

uncertamty quantification work; marsms: application of CSAU to a LBLOCA, NUREG/

e phenomena / processes commonly of moderate C 249 EGAG Idah and high importance in prior applications of the margms: application of CSAU to a LBLOCA. Nuct. Eng PIRT process; and Des li9,1-117.

e desired elements in the organization and use of USNRC.1968. US Federal Register, Acceptance entena for expert teams to develop PIRTs. emergency core coohng systems for hght water reactors,10 CFR 50 Wi lson, G E , Wadswonh, D C., Mdler. B G., Lommers. L ,

Kroeger. P., et al.,1992a Phenomena based thermal by-Acknowledgements drauhe modehns requirements for systems analysis of a modular high temperature gas-cooned reactor. Nuct Eng This paper is based, in the main, on work that Des 136,319-333 i

1 l

a .

G E Wdson. 8 E Bosack , % clear Engineereg and Design 186 II998123 37 33 Wdson G E ,199 b Statmically based uncertamty analysis Boucher, T J .1997a Phenomena ident:6 cation and tank-for rankmg of component importance in the thermal hy- ing tables for Westmghouse AP600 SBLOCA. MSLB and drauhc safety analysis of the advanced neutron source SGTR Scenanos. NUREGiCR4541, INEL. LMITCO.

reactor. EGG SE 10078. EG&G. Idaho June l Wilson. G E., Fletcher. C D. Eltawda. F 1996 Use of PIRT Wilson G E.. Boyack, B E , Leonard, M T . Walsams. K A .

- process in research related to design artahcation of the Wolf. T W .1997b Final Report: BWR drymeu debns AP600 advance passive LWR. Int. Conf on Nuckar Engt- transport phenomena identincation and ranlung tables l neering, vol 2. ASME. (PIRTs) INEEL EXT 97&894, Idaho National Engineer.

Wdson, G E., Fletcher. C D Davis, C B . Burtt. JD. ins and Environmental laboratory, September.

O t

n

- ~

O +

Letter Report JCN W-2427 Task Order No. 4 RESOLUTION OF WATERHAMMER AND TWO-PHASE FLOW ISSUES Hossein Nourbakhsh Engineering Technology Division Department of Advanced Technology Brookhaven National Laboratory Upton, NY l1973 Prepared for the U.S. Nuclear Regulatory Commission OtTice of Nuclear Reactor Regulation January 1998

1. INTRODUCTION NRC Generic Letter 96-06 (GL 96-06)" Assurance of Equipment Operability and Containment Integrity During Design Basis Accident Conditions""Iincluded a request for licensees to evaluate cooling water systems that sene containment air coolers to assure that they are not vulnerable to waterhammer and two-phase tiow conditions N1 ore specifically. the issues of concern are *

(1) Cooling water systems sening the containment air coolers may be exposed to the hydrodynamic etreets of waterhammer during either a loss-of-coolant accident t LOC A) or a main steam line break (NISLB). These cooling water systems were not i designed to withstand the hydrodynamic etiects of waterhammer and corrective actions may be needed to satistV system design and operability requirements i (2) Cooling water systems serving the containment air coolers may experience two-phase j tlow conditions durmg postulated LOCA and NISLB scenarios. The heat removal assumptions for design-basis accident scenanos were based on single-phase flow conditions Corrective actions may be needed to satisfy design and operability requirements The waterhammer and two-phase flow concerns discussed in GL 96-06 are primarily associated with low-pressure systems a ed the existing methodology discussed in NUREG/CR-5220 R may be oserly conservative for low pressure applications. This view has been expressed by Consumers Power Company based on work that was done by Fauske and Associates Inc Associated with the Palisades Plant On October 16.1997, the NRC met with Consumers Power Company and Fauske and Associates in order to better understand the analysis that has been performed and to determine whether generic application of this approach is feasible for low-pressure systems The NRC also participated in a joint NRC/NEl workshop, held in Gaithersburg, Nianland on December 4.1997, where resolution of the GL 96-06 waterhammer and two-phase flow issues were discussed Brookhaven National Laboratory (BNL)was requested (Task Order No 04 under JCN J-2427) to assist the NRC stafTis reviewing resolution of the GL 96-06 waterhammer and two- j phase tiow issues BNL participated in the October 16.1997 meeting of the NRC statTwith Consumer Power Company and Fauske and Associates BNL also attended the Joint NRC/NEl workshop on GL 96-06.

This letter repon provides an assessment of the information discussed during the October 16.1997 meeting and the Joint NRC/NEl workshop. Section 2 provides an assessment of the adequacy and generic feasibihty of the approach that has been used by Consumer Power Company and Fauske and Associates. The adequacy and feasibility ofinitiatives that were suggested during NRC/NEl workshop for resolving waterhammer and two-phase flow issues are discussed in Section 3 Section 4 provides a brief summan together with conclusions 1

E

2. CONSUSIERS ENERGY WATERHA3DIER EVALUATION SIETHODOLOGY Dunng the October 16.1997 meeting, Fauske and Associates Inc. presented the Consumers Energy Waterhammer Evaluation Methodology for the Palisades Plant. First an assessment of bubble collapse during pipe re611 was disussed M This was followed by a presentation on important conditions controlling condensation - induced waterhammer. W An assessment of those two presentations are provided in this section 2.1 Assessment of Bubble Collapse During Pipe Refill To quantify a mechanistic assessment of collapse of voided region Fauske and Associates.

Inc. presented a modeling approach based on dynamics of a slug-bubble trapped between two liquid columns. Initially the two liquid columns are assumed to be at rest, but upon service water pump restan, one liquid column will begin to re611, push and compress the bubble Bubble collapse rate is quantified by solving the momentum equations of the two surrounding water columns. The etrect of vapor condensation in the bubble in the presence ofinert gas is included it was stated that gas comes out of solution when the water was boiled The followmg assumptions have also been used in the model (l) Non-condenseable gas and vapor obey the perfect gas law (2) Temperature ofliquid and pipe wall contacting bubble remain constant and equal mitial liquid temperature.

(3) Initial liquid temperature equals boiling point at internal pipe pressure.

(4) Rate of condensation is controlled by purely turbulent mass transport within the bubble.

(5) Effective friction loss coerTicients account for obstacles to the flow of the two water columns.

(6) Variations in the liquid column masses during the bubble collapse transient are ignored.

The details of the model formulation has not yet been provided. However, the modeling approach and the above assumptions, with the exception of assumption (3), seem, appropriate.

The nonconservative assumption ofinitial liquid temperature to be equal to the boiling point at internal pipe pressure may not be valid specially during non-LOCA (e g., LOOP only) and smaller break LOCA events when the energy transferred to the containment fan coil units (CFCUs) from

- containment atmosphere may not be sutlicient to cause boiling in CFCUs. For such situations voiding could occur due to system drainage (column separation).

2 L -_

i

F The sample results presented indicate (as expected) a small over pressure (10S of pso for l

l situations where the bubble is surrounded by high temperature water However. the sensitivities l of the results to the degree of subcooling and to the initial inert gas mass fraction in slug-bubble.

l assumed in the calculations. were not provided i

2.2 Important Conditions Controlling Condensation-induced Waterhammer l

l Fauske and Associates. Inc also presented a discussion on important condition controlling condensation induced waterhammer including the basic considerations for the condensation process Much of the discussions were qualitatise at and no plant specitic evaluations were presented at the meeting The imponance of the Jacob number as a scaling parameter were also discussed The Liange-Griffith chugging criterion (deris ed based on their experimental data) and its application to j a ditTerent set of experimental data reported as a pan of an N11T thesis were also discussed Some observations from the waterhammer experiments in a seruce water con 6guration performed by Fauske and Associates, were also presented However, the biases due to scale distonion or due l to non-prototypical conditions of these experiments have not yet been quantified l l

l

3. WATERilA.\l. \lER EVALUATION 31ETilODOLOGIES PRESENTED IN Tile JOINT NRC/NEI WORKSilOP ON GL 96-06 Consumers Energy Northern l'tilities and Baltimore Gas and Electric presented their evaluation methodologies for resolving waterhammer and two-phase now issues in the Joint NRC/NEl Workshop on GL 96-06 FI l For the operability assessment of Palisades Plant, simple thermal hydraulic models have been utilized to estimate the amount of steam generation. to track the location of steam bubble and to determine boundary conditions for start of re611 transient Void collapse has been assumed in piping downstream of air coolers The waterhammer load has been estimated using NUREG/CR-5220. Also, pipmg analysis has been performed to demonstrate operability. The details of their evaluations were not provided at the meeting and no discussion on the assumptions and input parameters used in their calculation were provided in thcir presentations For design basis assessment, the modeling approach and the result of a scoping study that was discussed earlier (see Section 2.1) was also presented at the workshop Nonhern Utilities presented the result of an integrated thermal hydraulic model for evaluation of waterhammer issues for Millstone Unit 2, A detailed computer modeling of the reactor building containment cooling water system (RBCCW) and cooler units components has been developed for the analysis The results of the study indicated a pressure surge of = 250 psia in cooler unit inlet headers for a LOCA/ LOOP study. No details of the assumptions and input l

parameters used for the analysis was provided at the workshop However, the applicability of the one dimensional integrated thermal hydraulic model in addressing some of the important 3

1 f

I l

e .

phenomena of concern to waterhammer issue (e g , water steam stratification in horizontal lines) is questionable Baltimore Gas and Electric. teaming with Creare. have developed an analvtical model for the s oid collapse for evaluation of waterhammer issue for Cahen Clifl's Nuclear Power Plant.

This new modeling approach. as stated in their presentation. has been reviewed by Dr Grillith formerly of NilT The details of this modeling approach have not been made available for review

4. St!.\l.\lARY AND CONCIESIONS An assessment of the information discussed during the October 16,1997 meeting and the joint NRCSEI workshop on GL 96-06 has been provided Although the imponant physical processes and phenomena ofinterest to GL 96-06 waterhammer and two-phase flow iuues hase been identined, no systematic approach has been adopted for resolving these technical issues An integrated experimental and analvtical methodology somewhat similar to the one developed for ses ere accident technicalissues resolution M seems, also appropriate for resolving the GL 06-06 waterhammer and two-phase flow issues The basic component of this physically based methodology is illustrated in Figure 1 The integration is achieved by specifying the safety issues and prioritizing the physical processes which need to be considered to resolve an issue aComponent 1) and by expressing them in terms of specincation for both experiments (Component II) and model deselopment.modeling improvement (Component III). Technical issues resolution is achieved by means of bounding as.,umptions and calculations (Component V) or by means of more realistic calculations and their uncertainty quantifications (Component IV)

This approach assures that the analytical methods used to resolve an issue is comprehensive, systematic, auditable, and traceable. In addition, the proposed approach assures that all important features of an issue are fully addressed and the analytical models (and their assumptions) used to resolve a safety issue have the capability to scale up processes to plant relevant conditions 4

\

( i

-i w

T t

n g e o n r t

i i a

a d r nd r u

g n e u n e.

i o u o a s m f

ni sB s n s m o ta f i h n o 4 a twt oia i l Cic f i

c i t tr t

pl d c u ne a p 6nn h oiulmc t e

nS ct a 4 e u s e

s T lo sC sA W

i e

e e O D V R -

%n l.o G i t

,' r u l o o f

s*

y.

q R i

,d e

s s n n lote e s a e i di s u

ss e n a n n y nt a o s t ee la nt ed h I i

lah a

e mta o s l 4 At ai a l t e l nk a g nV M a ic rc n el ne h e v h me d ic h p aid c S P E c re o e s hn e S pM o T x p c o e l

E r T s P w

." e lo h

t F

< t e

s O e a s hP t

y n t

e ci n t

n n ti a o e

/

t nm mte r p h n e ev e la c n m t

a io mo u e no i C

o Pt a p r sR Ut a lo p ish e d ic c t

ic 4 e m l ni f nfi e vi a c wat n i

s d c

_ e g ic ep in ns a a Dn i hct enu o 1 1

AcS l

l e e el t uiQ d do T oa o s luc 1 MM Re la ii..

N C e r

. u 1

q i

1

r y 4

5. REFERENCES 1 Nuclear Regulatory Commission ( NRC) Tssurance of Equipment Operability and Containment integrity Dunng Design-Basis Accident Conditions." NRC Generic Letter 96-06.1966

Izenson. N1 G . P H Rothe and G B Wallis. " Diagnosis of Condensation-Induced Waterhammerf NUREG CR-5:20. October.1988 3 Fauske and Associates. Inc . ' Assessment of Bubble Collapse During Pipe Retillf presentation to L' S NRC. October 16.1997 4 Fauske and Associates. Inc . "Important Conditions Con

• rolling Condensation-Induced Waterhammer. ' presentation to U S NRC. October 16.1997.

5 C ensumers Energy, Surgent & Lundy and Fauske & Associates," Potential Waterhammer in Seruce Water System. Palisades Plant," presented at Joint

. NRC NEI Workshop on GL 96-06. Gaithersburg. NiD. December 4,1997 6 Kalyan K Niyogi. "Niillstone Unit : Experience With GL 96-06 Waterhammer I Issue," presented at Joint NRC NEl Workshop on GL 96-06, Gaithersburg, NID.

December 4.1997 7 J Todd Conner, "Calven ClitTs Nuclear Power Plant Generic Letter 96-06 Activities." Baltimore Gas and Electric, presented at Joint NRC/NEI Workshop on GL 96-06. Gaithersburg, NID, December 4,1997.

8 U S Nuclear Regulatory Commission. 'An Integrated Structure and Scaling Niethodology for Severe Accident Technical Issue Resolution," NUREG/CR-5809, Dratl for Public Comment November 1991 l

l l

l

> l l 6 l

i l

l