ML19341D113
| ML19341D113 | |
| Person / Time | |
|---|---|
| Issue date: | 02/28/1981 |
| From: | Ross D NRC - LOFT SPECIAL REVIEW GROUP |
| To: | |
| References | |
| NUREG-0758, NUREG-758, NUDOCS 8103040978 | |
| Download: ML19341D113 (100) | |
Text
- ---
NUREG-0758 Report of the LOFT Special Review Group
- 3..
U.S. Nuclear Regulatory Commission LOFT Special Review Group f **'* %
- ..... If
'f0 1
0 n.,g.~
o g Ap
~
3 m
B1030 40
i l
Available from GP0 Sales Program Division of Technical Information and Document Control U.S. Nuclear Regulatory Commission Washington, DC 20555 Printed cop.' price:
$4.50 and National Technical Information Service Springfield, VA 22161
. : :2 - - - _ - -
Report of the LOFT Special Review Group i
- _-~ ::
Manuscript Completed: February 1981 Date Published: February 1981 LOFT Special Review Group U.S. Nuclear Regulatory Commission Washington, D.C. 20555
,y...,,,
s....f.
.e
ABSTRACT In July 1980 the Advisory Committee on' Reactor Safeguards (ACRS) published NUREG-0699, " Comments on the NRC Safety Research Program Budget for Fiscal Year 1982." On page 19 of NUREG-0699, the ACRS made the following recommendation to the Commission:
LOFT represents the largest single expenditure in the safety research budget so that its program must be considered with special care.
We recomhiend that the tests through FY 1982 be adequately funded and that following the 1982 tests the facility be decommissioned 4
unless it is taken over by the nuclear industry.
The final tests to be run to the completion of the program should be carefully scrutinized and evaluated by RES to obtain the most useful final series.
We would also wish to contribute to the choice of these tests.
Efficient operation of the facility appears to require the requested level of support and therefore we endorse that level.
As a result of this recommendation and pursuant to the Commission's decision on SECY-80-398, " Creation of a Panel of Experts to Review the LOFT Program and Report to the Commission", the NRC Executive Director for Operations established the LOFT Special Review Group (LSRG) on November 14, 1980.
The LSRG was established to review the LOFT program and report its findings to the Commission. The primary purpose of the group was to consider whether LOFT should be decommissioned in FY 1983, as recommended by the ACRS.
- This report represents the results of the LSRG evaluation of the LOFT program and is submitted to the Commission as an aid in its decision whether to continue NRC support of the LOFT project beyond FY 1982.
The principal consensus reached by the LSRG recommends continued NRC support of the LOFT program through FY 1983.
t I
i y
4 iii
CONTENTS I
Page ABSTRACT................................................................
iii 1 INTRODUCTION.........................................................
1-1 1.1 Purp-]se.......................................................
1-1 1.' 2 Scope of Review and Information Sources.......................
1-2 1.3 Review Group Activities.......................................
1-3 1.4 Report Organization...........................................
1-4 1.5 Options Considered............................................
1-4 1.6 References....................................................
1-5 2
SUMMARY
AND FINDINGS.................................................
2-1 2.1 Summary.......................................................
2-1 2.2 Fi ndings and Recommendation...................................
2-3 2.2.1 Findings...............................................
2-3 2.2.2 Recommendation........................................
2-4 i
3 LOFT HIST 0RY.........................................................
3-1 3.1 P,cckground....................................................
3-1 4
3.2 The Facility..............................................
3-3 3.3 Summary of Completed Tests....................................
3-4 3.3.1 Large-Break Nonnuclear tests...........................
3-4 3.3.2 Large-Break Nuclear Tests..............................
3-4 3.3.3 Small-Break Nuclear Tests.............................. 4 3.3.4 Anticipated Transients.................................
3-6 3.4 References...................................................
3-6 4 -THE ROLE OF LOFT IN LWR SAFETY ANALYSES..............................
4-1 4.1 Introduction..................................................
4-1 4.2 LOCA Goal.....................................................
4-1 4.3 _ Development of the LOFT Testing Program.......................
4-2 4.4 Scaling and Code Assessment...................................
4-3 4.4.1 Scaling Considerations.................................
4-3 4.4.2 Role of LOFT in Licensing.........
4-6 4.4.3 Code Assessmerit........................................
4-6 4.4.4 Identification of Phenomena............................
4-8
\\
.~
CONTENTS (continued)
Page i
4-8 4.5 Additional-LOFT Testing Needs.................................
4-8 4.5.1 Large-and Intermediate-Break L0CAs....................
4-9 4.5.2 - Small-Break L0CAs......................................
4.5.3 Transients and Non-LOCA Accidents......................
4-10 i
4.5.4 Summary of Recommened Testing with L0FT................
,4-10 4-11 4.5.5 Impact of Reduced Testing Matrix.......................
4-12 4.6 Findings......................................................
4-12 4.7 References....................................................
5-1 5 OPERATIONAL ASPECTS..................................................
5.1 Introduction..................................................
5-1 5.2 Human-Machine Interface.......................................
5-1 5.3-Development of Instrumentation...............................
5-3 5-4 5.4 Findings......................................................
5.5 References....................................................
5 6 DEGRADED CORE C00 LING................................................
6-1 6.1 Introduction..................................................
6-1 6.2 Related Studies...............................................
6-1 6.3 Related LOFT Test Sequences...................................
6-1 6.4 Evaluation....................................................
6-2 6-3 l
6.5 Findings......................................................
6.6 References....................................................
6-4 7 RISK-REDUCTION PERSPECTIVES..........................................
7-1 7.1 Introduction..................................................
7-1 7.2 Risk Significance of Transients and Accidents.................
7-1 7-1 7.3 Large LOCA Risk...............................................
7.4 Transient Risks...............................................
7-2 7.4.1 L6 Series - Anticipated Transients.....................
7-2 7.4.2 L8 Series - Core Uncovery and Fuel Damage..............
7-3 7.4.3 L9 Series - Transients with Multiple Failures..........
7-3 7-4 7.5 Degraded Core Risks...........................................
7-4 Operational Risks.............................................
7.6 7-5 7.7 Findings......................................................
7-5 7.8 References....................................................
vi a
i l
1 CONTENTS (continued) fate APPENDIX A - Letter from F. C. Finlayson (Consultani, LSRG) to D. F. Ross, Jr. (Chairman, LSRG),
Subject:
Recommends Continuation of LOFT Program through FY 1986, dated January 28, 1981.
A-1 APPENDIX B - Nuclear Steam Supply System Vendors' Comments Regarding the LSRG Studies.
B-1 vii
L l
FIGURES 1
i l
,1.1 Charter of the LOFT Special Review Group...........................
1-7 3.1 LOFT Major Components...................................
3-7 TABLES Page 1.1 -LOFT Special Review Group Members and Consultants...................
1-8 4.1 Original LOFT. Test Series...........................................
4-14 4.2 Current LOFT Test Plan (FY 1980-1986)...............................
4-15 4.3 Some Major Scaling Distortions in L0FT..............................
4-17 4.4 LOCA ECC Experimental Program Matrix................................
4-18 4.5 Recommended Test Program that Supports LOFT Decommissioning at the end of FY-1983........................................
4-19 4.6 Recommended Test Program that Supports LOFT Decommissionin at the End of FY 1982.....................................g 4-22 6.1 Current LOFT Tests Related to Degraded Core Cooling.................
6-5 7.1 Risk-Significant Sequences - PWRs...................................
7-6 7.2 Risk-Significant Sequences - BWRs...................................
7-7 ix
-1 INTRODUCTION 1.1 Purpose In July 1980, the Advisory Committee on Reactor Safeguards (ACRS) published NUREG-06991, " Comments on the NRC Safety Research Program Budget for Fisce Year 1982." On page 19 of NUREG-0699, the ACRS made the following recom-mendation to the Commission:
LOFT
- represents the largest single expenditure in the safety research budget so that its program must be considered with special care.
We recommend that the tests through FY 1982 be adequately funded and that following the 1982 tests the facility be decommissioned unless it is taken over by the nuclear industry.
The final tests to be run to the completion of the program should be carefully scrutinized and evaluated by RES to obtain the most useful final series.
We would also wish to contribute to the choice of these tests.
Efficient operation of the facility appears to reqcire the requested level of support and therefore a
we endorse that level.
As a result of this recommendation and pursuant to the Commission's decision on SECY-80-398, " Creation of a Panel of Experts To Review the LOFT Program and Report Their Findings to the Commissionu2, the NRC Executive Director for Operations established the LOFT Special Review Group (LSRG) on November 14, 19803 The charter of the LSRG is shown on Figure 1.1.
The LSRG consists of seven persons from the NRC staff and other federal agencies, supplemented by eight consultants.
The members and consultants are listed on Table 1.1.
As expressed in the LSRG charter, the primary purpose of the LSRG is to perform a technical review of the overall LOFT program and report its findings to the Commission.
The report is to serve as an aid to the Commission in its decision whether to continue NRC support of the LOFT program beyond FY 1982, or whether LOFT should be decommissioned in FY 1983, as recommended by the ACRS.**
In financial terms, LOFT costs the NRC approximately $50 million per year.
This figure does not vary much as a result of the number or types of tests conducted at the facility.
As discussed later in this chapter, the LSRG
^ The Loss-of-Fluid-Test (LOFT) Facility is operated by EG&G Idaho, Inc.,
under contract to the Department of Energy (DOE).
The experimental program is directed and funded by the NRC.
- The ACRS, at its January 8, 1981 meeting, indicated that it was considering recommending that NRC support of the LOFT program be terminated at the end of FY 1981, a year earlier than expressed in NUREG-0699.
considered three specific options for the LOFT program These options and their costs are:
Option A:
LOFT program continues thre, ugh FY 1983
$150 million Option B:
LOFT program continues through FY 1982'
$100 million Option C:
LCFT program continues through FY 1981
$50 million Thus, the LSRG has reviewed the benefits and limitations of spending between
$50 million and $150 million to obtain experimental results from the LOFT facility.
To put these figures in context, it should be pointed out that from the program's inception in 1962 to the end of FY 1980, LOFT expenditures have exceeded $350 million.
1.2 Scope of Review and Information Sources As specified by its charter, the scope of the LSRG review included:
(1) reviewing the benefit expected from the program during the FY 1981-1983 period; (2) considering LOFT in perspective with regard to the overall NRC research program and in terms of the needs of reactor regulation; (3) assessing LOFT contributions to specific regulatory needs; and assessing whether it is likely that the program will provide the expected (4) information and whether the program will maintain reasonable flexibility.
To resolve some of the issues identified in the LSRG tasks cRed above, the group also addressed the following questions:
(1) Are the LOFT data inte,' twined with an anlysis development program, or is i
LOFT merely a link, along with other large experiments, in a code-assessment program?
(2) What are the significant atypicalities of LOFT?
(3) Can the results te obtained elsewhere, in whole or in part?
Given that the operational elements of LOFT (operators, control room layout, (4) instrumentation) are not prototypic, to what extent can LOFT assist NRC and the industry in this area?
(5) To what extent could information obtained from LOFT aid in reducing risk, l
actual or perceived?
l During the course of its review, the LSRG utilized many sources of information upon which this report is based. Where specific material has been relied upon in writing this document, it is cited and included in the reference section of the 1-2
individual chapters.
In general, the sources of material reviewed by the LSRG included:
{1) previously published literature regarding the design, construction and operation of LOFT; (2) results of LOFT experiments including selected Quick-Look Reports and Research Information Letters published by the Office of Nuclear Regulatory Research (RES);
(3) user-need lette s from the Office of Nuclear Reactor Regulation (NRR) to RES; (4) the RES Draft Long-Range Research Plan for FY 1983-1987; (5) presentations conducted by RES and EG&G Idaho; (6) discussions with management personnel associated with the LOFT project at RES and EG&G Idaho; (7) written responses to specific qunstions asked by the LSRG; (8) discussion with the ACRS in an open public meeting held on January 8, 1981; (9) written comments solicited from the nuclear steam supply system (NSSS) vendors (included as Appendix B of this report); and (10) physical examination of the LOFT faci.lity by the LSRG, including the sitnessing of LOFT Tests L3-6 and L8-1.
The scope of the LSRG activities was necessarily constrained by the limited period of time allotted to perform its mission (approximately 2 months) and by the part-time nature of the assignment for its members and consultants.
- Thus, not all of the above sources were used by each of the LSRG members and consultants.
- 1. 3 Review Group Activities Following its creation on November 14, 1980, the LSRG h91d its first organiza-tional meeting on November 17, 1980 in Bethesda, Maryland.
Subsequently, the group met in Idaho Falls, Idaho on December 9-11, 1980, during which time the LSRG toured the LOFT facility and witnessed tests L3-6 (a small-break loss-of-coolant accident (LOCA), with the reactor coolant pumps left running) and L8-1 (intentional core uncovery).
To effectively evaluate the overall LOFT program, the group divided into four subcommittees:
(1) scaling and code assessment; (2) operational aspects; (3) degraded core cooling; and (4) accidents / transients and risk reduction.
Various portions of the report were assembled by the subcommittees.
After an initial integration of tca report, the group met again as a whole at EPRI' in
- EPRI:
Electric Power Resea Oh Institute 1-3
1*
l Palo Alto, California on January 22-23, 1981 to develop and review the final report.
As discussed in Chapter 2, the findings made in this report represent the consensus of the members and consultants of the LSRG.*
1.4 Report Organization This report is divided into seven chapters:
(1)
Introduction; (2) Summary and Findings; (3) LOFT History; (4) The Role of LOFT in LWR ** Safety Analyses; (5) Operational Aspects; (6) Degraded Core Cooling; and (7) Risk-Reduction Perspectives.
As their titles imply, Chapters 1 and 3 provide background material for the Chapter 4 provides a discussion and evaluation of what the LSRG considers reader.
to be the major role of LOFT, its use and effectiveness in the licensing of large commercial nuclear power reactors.
Chapters 5 and 6 discuss ancilliary roles of LOFT which have received increased attention in the post-TMI-2 era.
Chapters 4 through 6 provide a deterministic or engineering judgment evaluation Chapter 7 provides additional insight on the effectiveness of the LOFT program.
of the LOFT program from a perspective of its risk-reduction potential.
- Together, the complementary deterministic and risk-reduction approaches serve as the basis for the findings of the LSRG that are presented in Chapter 2.
1.5 Options Considered As will be discussed more fully in Chapter 3, LOFT was at one time considered the test bed for studying fission product dispersal following a LOCA (without ECCS), but it has since evolved into a more complex, multi-faceted facility.
LOFT is now believed by RES and EG&G Idaho to serve the following roles:
Event Role LOCA an integral systems test facility that provides the nuclear ingredi-ent of scaled heat production plus intermediate-scale, thermal-hydraulic interaction Transients and Non LOCA an intermediate-scale experi-Accidents mental facility used to provide measurable plant response to various transients and accidents, including anticipated transients l
" Appendix A of this report contains a letter from one of the consultants of the LSRG whose views regarding the termination of the LOFT program differ from the consensus of the members and consultants.
- LWR:
Light Water Reactor.
1-4
Event Role Operating Simulation a facility for testing opera-tional aspects of transient and accident identification and i
recovery, such as the adequacy of instrumentation and control systems to follow the course of the event, the improvement of operating procedures, and test bed for the technological development and testing of new instrumentation Degraded Cooling a facility for evaluating the effectiveness of core response to conditions approaching and involving inadequate core cooling.
The LSRG provided an independent review of the potential for LOFT to provide technically beneficial information in these areas.
To this end, three options (as mentioned above) were considered:
Option A: Continue the LOFT program through FY 1983.
This option was identified as providing sufficient time to allow completion of the LSRG's recommended test program.
The program was developed by the LSRG based on a review of the current test program, current user needs, current regulatory issues, and expected risk-reduction benefits.
Option B: Continue the LOFT program through FY 1982.
This option was the recommendation which was advanced to the Commission by the ACRS in NUREG-0699 and which the LSRG was originally charged to evaluate in its charter from the Commission.
Option C: Continue the LOFT program through FY 1981.
This option was first presented by the ACRS in its meeting of January 8,1981.
- 1. 6 References The references cited below are available for inspection and copying for a fee in the NRC Public Document Room (PDR), 1717 H St., N.W., Washington, DC 20555 The document marked with an asterisk also is available for purchase from the NRC/GP0 Sales Program, U.S. Nuclear Regulatory Commission, Washington, DC 20555 and/or the National Technical Information Service, Springfield, VA 22161.
(1)
U.S. Nuclear Regulatory Commission, " Comments oa the NRC Safety Research Program Budget for Fiscal Year 1982," USNRC Report NUREG-0699, July 1980.*
(2)
U.S. Nuclear Regulatory Commission, " Creation of a Panel of Experts to Review the LOFT Program and Report to the Commission," SECY-80-398, August 25, 1980.
1-5
r' p
(3) Memorandum from William J. Dircks (NRC) to Harold R. Denton,. Victor Stello, Jr., and Robert B. Minogue (NRC), " Establishment of the" LOFT Special Review Group," dated November 14, 1980.'
[
k' d
4 h
b i
I
/
l 1
l.
1-y I
4 1
i.
l.
1 1-6
...,,,--.,,,..._._..,.n.
~
CHARTER OF THE LOFT SPECIAL REVIEW GROUP This group is established for the purpose of reviewing the LOFT' program and reporting on.their findings to the NRC Commissioners.
The review shall be technical-in nature, focusing on, but not limited to, the benefits expected from the program planned for the FY 1981 to FY 1983 period.
The primary purpose of the group is to consider whether LOFT should be decom-missioned in FY 1983, as recommended by the ACRS.
The group would be expected to consider the LOFT prog' ram from the perspective-of NRC's overall research
- program and in terms of the needs of reactor regulation.
To facilitate this work,-NRC and the INEL*, where LOFT is located, would provide presentations, reports, and tours and interviews.
Also, the group would be welcome.to' attend any tests performed in the LOFT reactor or related facilities.
.The report would be intended to aid the Commissioners in their decision whether
.to continue NRC support of the LOFT project beyond FY 1982.
The report should address s~ ecific regulatory needs and describe how the results 'of the LOFT p
program are expected to meet those needs.
Furthermore, based on the performance and responsiveness of the program to date, the report should indicate the like-lihood that the planned program will provide the expected information and that it maintains reasonable flexibility to address changing regulatory issues.
A final report would be issued by February 3,1981, and after follow-up
.-discussions with the Commissioners, the group would be dissolved.
"INEL:
Idaho National Engineering Laboratory.
Figure 1.1 Charter of the LOFT Special Review Group 1-7
~
I I
l Table 1.1 Members and Consultants of the l
LOFT Special Review Group J
MEMBERS AFFILIATION D. Ross (Chairman)
Office of Nuclear Reactor Regulation, NRC Office of Nuclear Regulatory Research, NRC R. Bernero Office of Nuclear Reactor Regulation, NRC R. Capra National Aeronautics and Space Administration L. Jaffee A. Pressesky Department of Energy Office of Nuclear Reactor Regulation, NRC B. Sheron R. Woodruff Office of Inspection and Enforcement, NRC 1
CONSULTANTS AFFILIATION I. Catton University of California at Los Angeles F. Finlayson Aerospace Corporation P. Griffith Massachusetts Institute of Technology H. Isbin University of Minnesota H. Kouts Brookhaven National Laboratory R. Pack Institute f Nuclear Power Operations A. William Snyder Sandia National Laboratories Electric Power Research Institute B
,tm 4
1-8
2
SUMMARY
AND FINDINGS 2.1 Summary As directed by its charter, the LSRG examined the role of LOFT with respcct to specific regulatory needs in order to ascertain whether current safety issues might be better illuminated by data obtained from LOFT.
The group examined in particular the expected benefits to be derived from the program during FY 1981 through FY 1983.
In examining the role of LOFT in regard to regulatory needs, the LSRG found that LOFT is viewed differently by several segments of the nuclear community.
RES and EG&G Idaho consider LOFT to be a multi-faceted experimental device that will be needed through FY 1986 to perform research on a variety of diverse tasks including:
(1) providing data for the assessment and improvement of computer codes intended to predict PWR behavior under a variety of transient and accident conditions; (2) providing an understanding of PWR behavior under various transient and accident conditions to serve as an aid in identifying operations needed to stabilize and recover the plant; (3) serving as a test bed for the interpretation and improvement of plant instrumentation; (47 providing research on human-machine interface aspects; and (5) serving as a faci.ity with which to conduct degraded and molten core experiments.
The much-needed formal viewpoint of NRR regarding the usefulness of LOFT for most of these tasks is not well defined.
The ACRS views the LOFT program as less effective with respect to the overall NRC mission than other high priority programs within the con-straint of the annual RES budget (approximately $200 million).
In addition, tha industry, as evidenced by letters from reactor vendors (Appendix B), does not foresee or endorse any long-term role for LOFT.*
Utilizing these initial inputs, the LSRG evaluated the role of LOFT in the 4
following specific areas:
(1) Safety Analysis:
including transient and accident code assessment and simulation; (2) Operational Aspects:
including human machine interface and instrumentation assessment and development; and (3) Degraded core Cooling:
including both mild and severe core degradation testing and the pending degraded core cooling rulemaking efforts.
These roles were examined through both engineering judgment and risk reduction potential perspectives.
The principal role of LOFT, as seen by the LSRG, is assisting in the completion of the LOCA mission.
The LSRG sees the other roles of LOFT as somewhat peripheral, possibly distracting, and not well-defined by NP.R need:-
This position is developed in detail in Chapters 4 through 7 of t.fs repcet.
In parallel with the experimental programs for LOCA is an analysis development l
program.
The LSRG believes that an end product, or NRC goal, is to produce a realistic assessment. of the LOCA.
However, it appears to the LSRG that the
- Tests that could potentially yield useful additional information in the short-term are specifically identified in the Westingbrise letter.
2-1
assessment portion, which consists of comparing analyses and experiments, is not proceeding fast enough.
The LSRG believes that, with due attention to the overall LOCA program, LOFT can provide key information that will enable the NRC to close the books on LOCA uncertainties.
As one of the consequences of the Emergency Core Cooling System (ECCS) hearings, the Commissioners (then Atomic Energy Commission) made a public commitment to quantify in a reasonable manner the safety margins embedded in the ECCS acceptance criteria and the associated prescriptions for evaluating emergency core cooling systems.
Further, it was intended by the Commissioners that the ECCS licensing methods were to contain a degree of flexibility so that contents of the evalua-tion models could be responsive to new information.
Changes which are warranted on the basis of better understanding of the important phenomena associated with LOCAs should be a driving force which leads to improvements in ECC systems and optimization of parameters for present procedures and systems, as well as to stabilization and increased efficiency in the licensing methods.
Throughout its deliberations, the LSRG considered the need to recommend whether operation of the LOFT program should continue for the indefinite future or Based upon what it perceives whether the program should have a defined endpoint.
to be the current regulatory needs, the LSRG has selected by consensus the latter.*
Howe,ar, in exploring the various technical options available with LOFT, RES and EG&G Idaho presented only two possible LOFT facility decision alternatives:
no LOFT facility or a LOFT facility fully manned for continuous operation at full test capabii!ty.
Given the value of the LOCA results to date and the potential for the future value of the LOFT facility to safety issues, an afterna-tive option of operating with reduced scope and operating costs strikcs the LSRG as viable. Justification for this approach lies in the fact that the current rate of tests (and the data generated) outpaces code development and assessment for licensing applications.
A stretched-out schedule, at a total cost no higher than now contemplated, could result in extended availability of the LOFT facility to resolve proble.ns which are row unforeseen.
The LSRG believes that the budgetary, staffing and use planning management of the LOFT facility should be reexamined and structured consistent with a testing program which is paced to NRC code development and assessment and which can accommodate multiple programs and multiple users.
For example, the LOFT facility operating policy might require that operating expend;tures be recovered fully from user (s) and program (s).
Thus, the composite of user schedules of programs, independently justified, would determine LOFT facility usage and staffing levels.
The decision to operate or decommission the LOFT facility would be determined by the magnitude of usage.
Should the need arise, the schedule could be accele-rated again to the full performance level.
In the time available, the LSRG was not able to develop this thesis.
Nevertheless, we believe that the operation of stretching the schedule, while reducing staff, and maintaining the total budget to complete mission should be studied furtSer.
TNote the exception to the Consensus in Appendix A.
i 2-2
2.2 Findings and Recommendation 2.2.1 Findings
- 1) Regulatory needs dictate additional LOCA and non-LOCA tests should be per-formed with LOFT prior to its decommissioning:
(a)
In order to resolve the remaining LOCA uncertainties, large,
intermediate, and small-break test results are needed.
These data will be utilized:
for code assess u nt, confirmation of the complete-ness of FCCS models, and the quantification of safety margins associated with the present ECCS acceptance criteria.
It is expected that the ultimate use of this research will lead to modification of 10 CFR 50.46 and Appendix K.
Small-break LOCAs are recognized as dominant contributors to risk, but the large-and intermediate-break LOCAs
' can be risk-dominant sequences if the analyses upon which the ECCS design is based are.in significant error.
(b) Anticipated transients occur on operating plants on the order of one or more times per year. Mild anticipated transients, as addressed in the LOFT L6 test series, are not considered signficant contributors to risk.
However, the transient codes upon which the analyses are performed must be assessed.
The acceptability of using LOFT for this purpose has not been well established.
The LSRG believes well-instrumented commercial reactors that incorporate a data retrieval system (similar to a flight recorder on aircraft) are the ideal source for providing these data.
However, until this data base is obtained, LOFT may play a useful role in helping to understand phenomena that have actually occurred in operating facilities.
LOFT is highly instrumented for this purpose, and tests of this nature can be cenducted quickly with minimal impact on other needed testing.
(c) Severe transients that involve multiple equipment failures and/or operator errors are signficant contributors to risk.
These transients have not traditionally been analyzed in Safety Analysis Reports.
The LSRG believes that these types of tests are needed for both ATWS code assessment and for validation of inadequate core cooling analyses and guidelines developed pursuant to the TMI Action Plan.
Core uncovery tests, which can be add-ons to other tests ad are calculated to have a peak cladding temperature (PCT) of less than 1500 F, can provide needed confidence in the ability of reactor systems to accommodate inadequate core cooling.
Tests with a potential for moderate core damage (for example, loss of all feedwater and LOCA with no ECC that have a calculated PCT of approximately 2000 F) are worthwhile but have the undesirable potential for contaminating the facility, impacting the schedule of future tests, and adversely affecting decommissioning costs.
These tests should be studied for their practicality and for their cost benefit.
If the results of these studies prove favorable, such tests should be planned for the end of the LOFT test program.
(2) HRR must take a formal and more active role in identifying LOFT testing needs.
2-3
-(3) RES code-assessment efforts must be improved to make better use of LOFT.
test data and to provide NRR with useful results on a more timely basis.
l (4) The LSRG sees the role of LOFT in evaluating new instrumentation for use in commercial reactors, such as pressurized water reactor (PWR) reactor vessel liquid level indication, as highly desirable and responsive to NRR l
needs; however, any decision to continue or terminate the LOFT program should be made independent of this role.
(5) Core melt accidents dominate reactor risk, but the severe fuel-damage experiments presently scheduled for FY 1985-1986 are peripheral to the LOFT mission.
Moreover, the facility does not have the necessary equip-ment to cope with severely damaged cores.
The additional cost of procuring the needed equipment is not included in the present LOFT budget.
(6) The entire area of human error is risk significant.
However, the human-machine interface research presently being conducted at LOFT is peripheral to the LOFT mission.
Because of the limitations inherent in a test reactor, LOFT is not the appropriate facility for conducting this research.
2.2.2 Recommendation Based upon what it perceives to be the current regulatory needs, the LSRG has identified in Table 4.5 of this report a series of tests that should be performed with' LOFT in order to meet those needs. While the LSRG has indicated a basic priority for these tests, it is recommended that the detailed planning and scheduling for these tests be performed by RES and EG&G Idaho in a manner that is consistent with decommissioning at the end of FY 1983.
In addition, the LSRG recommends that further studies be initiated to evaluate the feasibility of providing a stretched-out schedule for LOFT, at a total cost no higher than that now contemplated through FY 1983.
Such a study could result in extending LOFT facility availability to resolve problems that are now unforeseen.
2-4
3-' LOFT HISTORY
)
3.1 Background
In 1962 the Atomic Energy Commission initiated a test program at INEL (then NRTS*) to study the consequences of a LOCA in a commercial LWR caused by a rupture in a major primary pipe.
This! program was primarily intended to be a study of the release, transport, and deposition of fission products released from a reactor core meltdown'and was known as the " LOFT-U Program."
It was to be limited to a few nonnuclear blowdowns (simulated pipe rupture) and a reactor core meltdown.
The facility was intended to house a single, 55 MWt PWR.
The experimental test facility was to be located at and use si.ielded facilities originally constructed for the Aircraft Nuclear Propulsion Project.
In the mid-1960s, the thermal power levels of commercial nuclear power plants rapidly increased from one license application to the next, ultimately reaching 3000 MWt (1000 MWe).
As a result, in 1967, large PWR plant designs added new safety-features, such as ECCS for prevention of a reactor core meltdown in the event of a LOCA.- These new engineered safety features were provided to give greater assurance of safety.
In 1968, the importance of protecting the contain-ment structure was recognized, and so attention was shifted from studying fission product release and dispersal to assuring the efficiency of ECCS.
The AEC, therefore, redirected the LOFT program toward the complexities of studying the new ECCS design features for preventing core meltdown in large reactor systems and cores, rather than studying the sequence of events in an unprotected reactor core meltdown. This new direction resulted in the need for an integrated test facility to study the nuclear, thermal, hydraulic, and structural processes associated with a PWR-type LOCA, which is the design basis accident postulated to be the most demanding for the protective systems.
Consequently, a complete redesign of LOFT was begun in 1969 so that the system would model, as nearly as possible, the ccnditions present in the primary coolant system and core of a typical large PWR, as well as the associated ECCS performance during a postu-lated large-break LOCA, and at the same time it would provide systems to contain the possible release of radioactive fission products.
In addition to the normal PWR systems, LOFT has a special suppression tank to contain the mass aad energy of the blowdown and prevent the release of fission products.
Because the LOFT-U design was essentially complete and the facility was partially built in accordance with the earlier objectives, significant design changes and almost a new start were required in 1969 to satisfy the revised requirements.
Figure 3.1 shows a diagram of the present layout of the major LOFT components.
With the formation of the Energy Research and Development Administration (ERDA) and the NRC in 1975, the responsibility for the LOFT experimental program was assigned to NRC, and responsibilities for completion of the facility and its operations were assigned to ERDA.
The first test series (which in some respects represented commissioning tests) was nonnuclear and was run from 1976 to 1978.
" NRTS:
National Reactor Test Station
The first test in LOFT as a fully functional nuclear plant was run on December 9, 1978.
From that time on, the facility was considered complete, and the NRC has been responsible for funding the entire program.
Though the NRC has accepted the facility, it should be noted that the fission product cleanup system was never completed. This reduces the scope ol possible experiments substantially.
The LOFT test program was originally formulated during the 1970-1974 planning period and represented a skeletal structure of various test series.
Subsequently, as testing has been done, experience acquired, and licensing issues have arisen, the specifics of the tests to be performed in each series have been better defined.
NRR has been involved throughout this procedure.
However, NRR input has been primarily informal, and most of the test program identified through 1982 has been formulated by RES, utilizing informal and indirect NRR input and " anticipated" licensing needs.
Furthermore, as new needs have been identified, especially following the TMI-2 accident, the LOFT program has been greatly expanded in scope and extended to include the development and assessment of instrumentation to be used by opera-tors for the identification of accident conditions and recovery of the plant.
In additi n, an advanced operator diagnostic and display system is now under o
development as an overlay to control systems used during accident conditions established during LOFT tests.
NUREG-0660, "NRC Action Plan Developed as Result of the TMI-2 Accident,"1 identi-fies many individual items for which the spinoff results from LOFT tests are related.
These include:
(1) training and qualification of operating personnel; (2) use and development of simulators; (3) operating procedures; (4) testing a vent path for noncondensibles; (5) a fault-diagnostics system for training operators in the control and mitiga-tion of accidents which may lead to severe core damage; and (6) identification of steps in the evolution of severe damage to a core.
The LOFT program has been redirected by RES in an attempt to provide useful results which relate to these tasks.
Thus, the revised LOFT objectives that have been formulated by RES could be stated as follows:
To establish conditions ie a nuclear reactor that are broadly charac-teristic of a variety of transients and accidents postulated for a large PWR, for the purposes of testing and developing methods for analytical description of these events, for recognition and identification of specific aspects of transients and accidents, and for identification of manual and automatic steps to stabilize the core and recover from these events.
The possible end products would be:
(1) data for the improvement and assessment of computer codes intended to predict the behavior of PWRs under a wide variety of accident conditions; I
3-2
(2) following the behavior of PWR under accident conditions and developing the operator actions needed to stabilize and recover the plant; (3) the interpretation and improvement of plant instrumentation needed to identify and diagnose accident conditions; and (4) an advanced plant data display system developed and tested on a PWR under
-actual accident conditions.
As information becomes available from LOFT testing, it is communicated to NRR and the nuclear industry via " Quick-Look" and Experimental Data Reports.
Signi-ficant findings and conclusions are communicated formally in the form of Research Information Letters (PILs).
3.2 The Facility The major component of LOFT is the mobile test assembly (MTA), which consists of a 55 MWt reactor and reactor cooling system mounted on a double-width test dolly which is designed to be transportable by rail.
The MTA is installed in a containment building.
It has auxiliary systems for plant support and a conti-guous underground control room.
At the completion of its test life, or if a major unexpected problem were to occur sooner, the MTA theoretically can be moved by rail from the containment building into a nearby large " hot shop" for remote maintenance or disassembly.
It should be noted, however, that the LOFT containment building had been largely completed before the 1969 revision of LOFT test objectives and the reactor system design.
After redesign, the steam generator of the reactor cooling system was too tall to go through the equipment access door of the containment without disassembly from the MTA, thus precluding easy transport of the MTA to the hot shop.
The reactor core is 5.5 ft high and 2 ft in diameter, and it contains 1300 PWR-type fuel pins. The core is instrumented with high-temperature thermo-couples and other specially developed instrumentation to measure temperatures, flows, pressures, and coolant levels inside the reactor vessel and in attached piping.
The reactor coolant system has one active, heat-dissipating, operating loop
/
that models the three unbroken loops of a typical Westinghouse four-loop plant.
A second loop, called the blowdown or " broken" loop, models a single loop in a four-loop plant and contains quick opening valves that can simulate the rupture of a large single reactor coolant pipe in a commercial reactor.
The blowdown loop discharges reactor coolant into a suppression tank, which is pressure-and temperature-controlled to simulate the conditions calculated to exist in a containment su11 ding following a large PWR LOCA.
The LOFT facility's ECCS has the same types of components as a large PWR, and it is scaled to simulate the performance of the ECCS in a large reactor following a LOCA.
Three systems are provided for emergency coolant injection:
(1) a large accumulator containing water under gaspressure that can passively inject its contents into the reactor system quickly; (2) high pressure injection pumps to provide a moderate flow of high pressure water for mitigating small breaks; and (3) low-pressure injection pumps that inject a large flow of low pressure water continually to cool the core after the primary system has sufficier.tly 3-3
depressurized.
The primary coolant system and ECCS are extensively instrumented, and the simulated break size, location, and ECCS injection points and flowrates can be varied for experimental purposes.
Since early 1980, a diagnostic and display system has existed in the uM T control This system is being used to develop systems to diagnose faults, develop room.
improved procedures to be used in accident situations, and to improve techniques for display and interpretation of pertinent reactor and plant data during the accident.
The input of operators and the experience of simulated accident condi-J tions with LOFT are being used in the development of the system.
3.3 Summary of Completed Tests
- In the five years of testing at the LOFT facility, four distinct types of tests have been conducted:
(1) large-break LOCA experiments without nuclear power; (2) large-break LOCA experiments with nuclear power; (3) small-break LOCA experiments with nuclear power, including follow-on tests involving core uncovery; and (4) anticipated transients initiated at full power.
3.3.1 Large-Break Nonnuclear Tests The large-break nonnuclear tests were initiated at approximately full reactor coolant pressure and temperature.
They were designed to study the effectiveness of the ECCS in delivering coolant to the core and to provide experier.ce in handling the LOFT nuclear reactor under acc k at conditions.
3.3.2 Large-Break Nuclear Tests The large-break nuclear tests were designed to test certain conservatisms in the NRC ECCS rule.2 The results showed that the ECCS water is delivered more quickly to the core, that more reactor coolant remains in the reactor vessel after blowdown, and that less ECCS water flows out the break than is predicted by codes which are based on the ECCS rule.
These tests also demonstrated that the ECCS (which are scaled to those in commercial reactors) were effective in cooling the LOFT core.
The tests showed that early in the accident, even before the ECCS is actuated, the transient flow of water through the LOFT core signifi-cantly lowers the temperature of the fuel, leaving the core more readily cool-able when the ECCS water enters the core than ha'i been expected.
The details of the observed thermal-hydraulics are still unoer study, but the same cooling phenomena are predicted to occur in commercial reactors.
3.3.3 Small-Break Nuclear Tests The small-break test series is designed to study:
(1) the effectiveness of different types of heat sinks available; (2) the effects of varying the break size, location, and depressurization rate; and (3) the characteristics associated with running or not running the reactor coolant pumps during a small-break LOCA accident.
- Material extracted from the " Draft Long Range Research Plan Fiscal Years 1983 through 1987, USNRC Report NUREG-0740, November 1980.
A copy of the Final Plan will be publicly available on February 17, 1981.
3-4
The first test (L3-0) was completed two months after the TMI-2 accident.
It was initiated at full pressure and temperature, but to obtain data as fast as possible, nuclear heat was not used.* This test was performed to study the system response to a stuck-open pilot operated relief valve (PORV), as had occurred at THI-2.
The second test (L3-1), Jimulating a 4-in. diameter break in the pump discharge piping **, caused a slow, continuous depressurization, followed by eventual actuation of the ECCS to refill the primary system before the core was uncovered.
The third test (L3-2), simulating a 1-in. diameter break in the pump discharge pipe, caused a very slow reduction in pressure, which stabi-lized at an intermediate value.
Operator intervention then reduced the pressure sufficiently to actuate the ECCS, and the plant was stabilized without uncovering the core.
Of special interest was the indication that the primary coolant in the U-tube steam generator first underwent a transition from liquid natural circulation to liquid-vapor natural circulation, then possibly to reflux cooling (or condensate fall-back), and then back to liquid natural circulation, with no evidence of flow instability.
Another discovery was that flow paths in-the LOFT reactor which bypass the core have an important influence on the capability of LOFT to simulate the thermal-hydraulic behavior predicted in commercial PWRs during small-break LOCA accidents.
(See Chapter 4 for a more detailed discussion of bypass flow.) The fourth small-break test (L3-7) was designed to examine the effectiveness of various heat sinks available to PWRs.
The fifth and sixth small-breaks tests were identical except for operation of the reactor coolant pumps.
In the fifth test (L3-5), the pumps were shut down immediately after the test began; in the sixth (L3-6), they were left running.
A preliminary comparison of liquid inventories indicated that shutting the pumps down early in a small-break accident reduced the rate of mass loss through the break.
In a combined (piggy-back) test with L3-5, the liquid level was loucred far below the hot leg, and the flow conditions in the bot leg were monitored.
All flow measurements showed that flow continued in tha positive direction through the intact loop.
In a piggy-back test with L3-6, the core was entirely uncovered ***, and the cladding temperature was allowed to increase by about 300 F before the ECCS was initiated.
This additional test provide data for evaluating heat-up rates for uncovered cores and recovery by actuation of ECC systems.
The program is now being planned to combine tests in this way whenever possible.
Except for the first small-break LOCA test, the remainder of these tests were initiated with the core gererating power.
- Breaks in the pump discharge piping are considered the worst, because the primary system coolant inventory loss is the greatest prior to ECCS actuation and recovery.
- The planned test, based on code predictions, did not involve core uncovery at the time of reactor coolant pump trip.
Instead, the core was expected to be covered at the time of the pump trip, be allowed to slowly uncover in a controlled manner, and recover once the cladding thermocouples increased by 300 F.
Because of the significant error in the code prediction for this test, the core uncovered in its entirety when the pumps were tripped.
Cladding heat-up rates and ECCS recovery were much faster than planned.
3-5
3.3.4 Anticipated Transients In four anticipated transients run to date*, behavior of the pressurizer, controls, The data and valves and how they affect the system response have been studied.
have been used to assess transient computer codes.
3.4 References The references cited below are available for inspection and copying for a fee in the PDR ar.d for purchase from NTIS and/or the NRC/GPO sales program.
(1)
U.S. Nuclear Regulatory Commission, "NRC Action Plan Developed as a Result of the TMI-2 Accident," USNRC Reg. ort NUREG-0660, Rev.1, August 1980.
U.S. Code of Federal Regulations,10 CFR 50.46, " Acceptance Criteria for (2)
Energency Core Cooling Systems for Light Water Nuclear Power Reactors."
- Anticipated Transients Completed:
LG-1:
loss of steam load LG-2:
loss of reactor coolant system flow LG-3:
excessive load increase LG-4:
loss of main feedwater 3-6
. Broken iooP intacttoo)
^
w r
7 St.1 Ouick opening experimentat valve (2) measurement steam generator I
N\\
g stati.
PC-2 simulator u
\\
}
meas eet station
' )
Pressurizer O
p t:d S
ene ator isolation
,j'y mu atOf h
'.f,.
, gnmental
- ECC miection measurement location station Vented to umps perimental supression vessel L'
j measurement f
V
/
~
~
PC 3 k%/
station
'i rf experimental Reactor measurement station vessel Supressiori vessel Downcomer
,i Core -
Lcwor plenura i
INEL.A.16 383 Rear.cor vW Figure 3.1 -
LOFT major conponents.
m 4 THE ROLE OF LOFT IN LWR SAFET7-ANALYSES 4.1 Introduction The purpose of this chapter is to discuss the role of LOFT in LWR safety analyses.
Although this topic can be broad and far reaching, in this chapter it is restricted to the-thermal-hydraulic area, and only the role of LOFT in assessing the thermal-hydraulic _ behavior of.large PWRs in transients and accidents is evaluated.-
In considering the role of LOFT in this area, the LSRG identified these key
-questions:
\\
(1) What is the LOCA goal of the Commission, and how is LOFT keyed to that role?
(2) What are the scaling atypicalities, and how do they affect the results?
(3) How is LOFT used in the code-assessment process, and what impact would
-termination of the program have on the assessment process?
Eo.
of these questions will be addressed in the following sections.
4.2 LOCA Goal At present, the LOCA is analyzed according to the requ red and acceptable i
features defined in Appendix K to 10 CFR 50.1 The acceptance criteria for ECCS performance are defined in 10 CFR 50.46.2 The criteria, required features, and LOCA evaluation models developed to comply with these require-ments have been in existence since 1974.
When the ECCS rule was promulgated, uncertainties existed in the expected thermal-hydraulic performance of ECCS because of an insufficient experimental data.
To accommodate the uncertainties, conservative features were incorporated to ensure that the overall predictions would also be conservative.
It was anticipated that as more and batter data became available, some of the co'.,ervative features of the Appendix K rule could be relaxed.
As stated on ;.r.gu 30 of the " Opinion of the Commissioners" regarding the rulemaking hearing on the acceptance criteria for ECCS:
Continuing research and development will provide a more extensive data base for scch items as heat transfer coefficients during blow-down and during spray and reflood cooling, oxidation rates for zirconium, fission product decay heat, steam-coolant interaction, oscillatory reflood flows, fuel densification, pump modeling and flow blockage.
With the additional data it may become practical to assign a statisti-cally meaningful measure of precision to the calculation.
It is probable that, with a better data base, some relaxation can be made in some of the required features of the evaluation models.
- However, the Commission believes that any future relaxation of the regulations should retain a margin of safety above and beyond allowances for statistical error.
1 i
i
l Since the promulgation of the ECCS rule, NRC research efforts have been predominantly oriented toward resolving the LOCA uncertainties.
Moreover, the staff has always assumed that the ultimate use of this research would be a modi-fication of the Appendix K rule.
In fact, before the TMI-2 accident, efforts were underway by the staff to incorporate into the Appendix K rule research results in two significant areas: decay heat and metal / water reaction rates.4 LOFT has been considered as an important link in the overall effort to resolve j
some LOCA issues.
The LOFT design and operation have provided a strong focus l
for the development and use of calculational models which yield realistic or "best estimate" answers. Models used in licensing methods use bounding or conservative assumptions which yield conservative results.
Mor9 recently, LOFT has also been considered as a test vehicle for examining ie typicality 8
the.ystem response to transients and accidents other than LOCAs.
and t.sefulness of the facility for these purposes must also be examined.
4.3 Development of the LOFT Testing Program Before beginning any discussion of the technical merits or drawbacks of the LOFT facility, it is useful to describe the genesis of the LOFT testing program.
Seven series of tests were identified and agreed upon between the research and regulatory staff offices of the NRC in 1976.
The series were first published by INEL in 1978 and are shown in Table 4.1.*
Since these series of tests were formulated, the LOFT testing program has not char ged substantially, except that as a result of the accident at TMI-2 the testing priorities have been changed to emphasize small breaks and transients.
The current LOFT test plan proposed by RES is shown in Table. 2.**
This plan calls for testing through November 1985, with decommissioning to follow in December 1986.
Since the test program was originally developed, much has been learned about LOCAs as well as about transient and non-LOCA accident behavior, and regulatory priorities have shifted.
Some phenomena uncertainties, once considered of major importance, are no longer considered as significant.
Examples of these include ECCS bypass and blowdown heat transfer.
In addition, some phenomena and pro-cesses, previously rega-ded as having a low lmpact by the staff, have gained renewed interest.
Exa.ples of such processes are two phase natural circulation behavior and thermal-hydraulic behavior of system components.
Although RES has rearranged the LOFT testing program and testing details have been revised to better address current licensing issues, it is generally recognized that for the LOFT data to be of maximum benefit, detailed needs have to be better identified by F.RR, which is the prime user of LOFT data.
Since the original testing progrim wts developed, NRR has not formally identified current licensing nee.fs and priorities.
Because of this, much of the basis for present LOFT testig is predicated on " anticipation" of NRR needs. Approximately three months before the LSRG was established, NRR had initiated efforts to identify a comprehensive list of user needs for the LOFT facility; this list was expected to form the basis for future test planning.
^I.0. Burtt, LOFT Experimental Program Document," EG&G Idaho, Inc.,
Idaho National Engineering Laboratory, review copy dated December 8,1978.
- " Draft Long-Range Research Plan Fiscal Years 1983 through 1987," USNRC Report NUREG-0740, November 1980.
A copy of the Final Plan will be publicly available on February 17, 1981.
l 4-2
P p
'Although NRR has not prepared an up-to-date, comprehensive, user-need letter-specific to LOFT, a number of user-need letters have been prepared and trans-mitted to RES; these can and most likely do require information from LOFT.5 9 These user-need letters, for the most part, have purposely not identified any specific facility from which the required information should be obtained.
This is because one of the primary responsibilities of RES in responding to user-
- needs is to ascertain the optimum method by which to obtain the necessary data.
It is worth noting that user-need letter RR-NRR-77-5,5 "NRR Requirements for Loss of-Coolant Accident Analysis Computer Programs," specifically identified the NRR need for-a "best-estimate" evaluation capability.
LOFT has been identi-fied to provide.much of the experimental data base for code development and assessment to meet that need.
.he LSRG believes that the availability of a definitive user-need letter by NRR--documenting the tests needed by NRR and the basis for-the need--would have greatly increased the ability of the L54G to critically assess the role of LOFT in licensing.
The preparatien of a formal user-need 1 st of testing needs by NRR was not feasible within the LSRG schedule.
How Jer, one member of the LSRG was also a member of Reactor Systems Branch in NRR one of the prime users of. LOFT thermal-hydraulic performance data.
A prelim;ntry list of optimum LOFT testing needs was developed by this LSRG member for consideration by the LSRG as a whole.
'This list, although modified by the LSRG, is considered to reflect to some extent NRR data needs from LOFT.
This list is discussed in Section 4.5.
4.4 Scaling and Code Assessme.nt
~4.4.1 Scaling Considerations The LOFT facility is scaled to a typical four-loop Westinghouse PWR with a 15x15 fuel assembly array.
The scaling basis is to maintain the same power / volume ratio between the LOFT facility and the prototype.
The reactor power in the LOFT. facility is approximately 55 MWt, compared to the prototype power of about 3200 MWt, for a~ scaling ratio, of approximately 60 to 1.
For comparison, the only other U.S. integral system test facility capable of performing testing similar to LOFT is the Semiscale Facility.* This is a 2.5 MWt nonnuclear facility which uses electrically heated rods to simulate the nuclear fuel rods.
The scale of this facility relative to a 3200 MWt PWR is about 1500 to 1.
Moreover, the scaling is such that Semiscale was originally considered to be a "one-dimensional" facility, with no thermal-hydraulic effects in the radial direction.
Testing has shown, however, that three-dimensional effects exist even at this scale.
"Semiscale is an integral systems test facility, also located at INEL.
Semiscale is a volume-and power-scaled, thermal-hydraulic test facility whose core is simulated by ele.rically heated rods.
It is designed to simulate typical-Westinghouse fo loop PWRs.
4-3
7 In any scaled system, scaling distortions will occur.
For example, if volume scaling is preserved, then surface area will not scale.
However, the main objec-tives in scaling are to mininize the effects of scaling distortions and to try to properly represent most key thermal, hydraulic, and neutronic phenomena expected to occur in the prototype.
Many of the scaling distortions in LOFT arise not only from the direct scale reduction, but are also the result of the need to use available equipment, financial considerations, and additional requirements (such as maintaining prototypic elevations).
Some of the major scaling dis-tortions are listed in Table 4.3 along with comments on possible impacts of the distortion.
Scaling distortions are not necessarily detrimental even if they result in some distortion of the thermal, hydraulic, or neutronic phenomena.
This is because the results of scaled tests are never directly applied to licensing decisions, but rather ere indirectly used via computer codes.
Scaling distortions should be, for the most part, appropriately accounted for in the analysis codes and should not impact large PWR predictions.
A recent study conducted by RES has compared the large-break LOCA behavior of LOFT to the predicted behavior of a typical Westinghouse four-loop PWR.20 This comparison was made using best-estimate computer codes.
The RES conclu.sion was that the thermal-hydraulic behavior of LOFT was very similar to that of the four-loop PWR for the first 20 seconds of the blowdown phase of a large-break LOCA.
At present, the typicality of LOFT to a PWR during the reflood portion of a large-break LOCA has not yet been formally evaluated.* Although the conclusion drawn in Reference 10 has not been universally accepted, it is encouraging in that, at least for the blowdown portion of the large-break LOCA, no known scaling distortions appear to significantly distort the thermal-hydraulic benavior.
Continued study is warranted, however.
For accidents and transients other than the large-break LOCA, the effect of the scaling distortions and their significance have not been thoroughly and systematically evaluated by either RES or its contractors.
The LSRG looked at the usefulness of LOFT to accurately simulate small breaks.
The length and position of the LOFT core may preclude partial core uncovery for scaled small breaks in which core uncovery would be expected for the prototype under similar conditions. Thus, a significant aspect of small-break simulation--
namely core uncovery--is not properly represented in LOFT.
Recently, RES pro-posed a number of tests which produce slow core uncovery similar to that expected during small breaks.
LOFT is not instrumented to provide an adequate claracter-ization of the steam generator.
In addition, flow instrumentation that would make it possible to evaluate what is happening at the low flowrates experienced during small breaks is not in place.
Also, a recent small-break test (L3-1) did not produce an expected small-break phenomenon, namely liquid clearing of the pump loop seal."
This behavior has been attributed te a bypass flow path which exists between the vessel upper
- Since the LOFT core is less than half the heignt of the core in a modern PWR, typicality will be hard to demonstrate.
4-4
head and the downcomer annulus,* as well as another bypass between the hot-and cold-leg pipes.due to leakage through the reflood assist valves.
Large PWRs also have the. upper-head-to-annulus bypass, but in those cases it has been predicted that the loop seal will clear.
EG&G calculations indicate that this leakage path in LOFT is approximately 3 percent of the core flow, or comparable to prototype values.
However, the actual leakage path cannot be measured directly but only indirectly. inferred by assuming a value which leads to the prediction
- agreement with the test;-therefore further evaluation is necessary to resolve this problem.
The LOFT system lacks the complications present in the four-loop PWRs so that a variety of phenomena characteristic of multiple-loop behavior cannot be dis-covered.
Loop-to-loop hydraulic instabilities cannot occur with a single, active loop. Water inventory and location are uncertain, and core bypass through inactive loops with the reactor coolant pumps running cannot occur with a single, active loop.
Adequate flow instrumentation for determining hydraulic behavior in the vicinity of the core is not in place.
The " broken" loop in LOFT does not contain active components (that is, pump and steam generator), but rather they are simulated by pipi have flow resistances equivalent to that of the active compng flow paths that onents.
For large-break LOCAs, where the hydraulic behavior is primarily dominated by inertial effects, this is acceptable.
However, for many small breaks which depend upon natural circulation for decay heat removal and are signficantly affected by steam generator heat transfer, the inactive components are probably inadequate.
Only recently has LOFT been seriously considered for simulating anticipated transients and other accidents in addition to LOCAs.
The LSRG is not aware of any evaluations that have been performed to show the effects of scaling dis-tortions and differences in the balance of plant for these events.
The key-consideration, however, is whether the tests proposed with LOFT would provide sufficient challenges to the mathematical modeling features of computer codes to test their applicability to PWRs.
More work is needed to provide assurance of the adequacy of these LOFT tests.
A systematic evaluation of scaling distortions is believed necessary as part of any testing program for anticipated transients.
Despite the lack of this evaluation, however, the LSRG has recognized the unique features of the LOFT facility which support its desirability as a test vehicle for certain antici-pated transients.
This is discussed in more detail in Section 4.5.
In summary, the major scaling distortions of LOFT, as they apply to and affect large-break LOCAs, have been identified and their effects have been evaluated for the blowdown phase of a large break LOCA. While some additional work in this area is still needed, for the most part, the scale distortions do not appear to significantly or adversely affect the large-break LOCA behavior in such a way to render the results questionable for either general applicability to PWRs or for code assessment.
The capability of LOFT to adequately simulate small-break
- In B&W-designed reactors, a bypass flow path between the vessel upper-head and downcomer upper annulus has been included in the design.
These flow paths are refe. ed to as' vent valves.
4-5
behavior in PWRs is presently in question ana.equires further evaluation before a final conclusior can be drawn.
If the typicality issue is ignored and an evaluation made solely on the use of LOFT as a code-assessment tool for small breaks, it can be said that LOFT provides useful information in this area.
The capability of LOFT to adequately simulate anticipated transients (including ATWS) and non-LOCA accidents in'PWRs has not been evaluated.
This is a necessary part of any test program designed to look at anticipated transients and accidents.
Despite the lack of this evaluation, it is believed that both the scale of LOFT and the fact that it is a nuclear facility would make LOFT desirable for selected testing of anticipated transient and non-LOCA accident behavior.
4.4.2 Role of LOFT in Licensing Loft plays no direct role in licensing.
This has been and continues to be the practice of the NRC.
LOFT is dedicated to studying postulated transients and accidents, and has the mission of augmenting the experimental data base to enhance the understanding through analysis of the interactions of the thermal-hydraulic and neutronic features in a scaled PWR system.
4.4.3 Code Assessment NRC-sponsored safety research includes multimillion dollar programs for code development and improvement. These programs involve the evolution of best-estimate system and components codes.
The predictive capability of the codes is tested by comparing predicted and experimental results.
Code developers are extending the state of the art in modeling complex, thermal-hydraulic features.
This includes local and overall responses in scaled and full-size geometrical components as well as in scaled integral systems.
The need to disassociate "the step-by-step improvement" in codes by code developers from efforts to evaluate how well the code performs has long been recognized.
In response to this need, the NRC has developed a program for independent code assessment. The code assessment program requires the assessors to work independently of the code developers.
This~ separation provides a check on how a code is used as well as an independent evaluation of the capabilities and performance characteristics of the code.
Code assessment begins when a version of a best-estimate code is released for use and the code features are " frozen."
Four National Laboratories are involved in code assessment, with the specific tasks assigned to each being controlled by RES.
Obviously, it is necessary to ensure that the code assessors not only have access to the experimental data base used by the code developers, but that they also have access to data not previously used which can provide further indepth studies to test features of the code.' These new challenges are used to ensure that the modeling has been based on a thorough understanding of the basic principles and that any empiricisms used'have not been " tuned" just to match previous experimental results.
The experimental elements used in the development and assessment of best-estimate LOCA codes include basic model studies as well as experiments involving small, intermediate, and nearly full scale.
Five separate features of the LOCA are amenable to detailed studies in what are known as separate effects tests.
For example, the first portion of the loss-of-coolant-accident involves the depres-f surization of the primary system and is called " blowdown." For this phase of the accident, modeling of the respont9 of the fuel elements requires detailed l
4-6
knowledge of the thermal-hydraulics involved; these separate effects are listed under blowdown heat transfer in the LOCA ECC. Experimental Program Matrix as shown in Table 4.4.*
Separate effects test data are derived from both United States and foreign facilities.
A full-scale PWR system is not part of the experimental program.
- Instead, several facilities have been built both here and abroad to simulate the primary system including the steam generators.
These systems are called integral systems.
As noted previously, LOFT is the only integral test facility which uses nuclear fuel.
The purpose of the integral tests is to provide an overall and detailed checkout of the performance of models derived from first principles and separate effects tests in a complete simulated PWR system.
Boundary conditions must be supplied to carry out separate effects testing, and the interactions and coupling of system features to the integral tests provide additional infoiination that the modeling has been adequate to result in realistic predictions.
The goal of code assessment is to quantify, in a reasonable manner, the applicability of the best estimate code to PWRs by quantifying the uncertainties in the figures of merit for the specific analysis and discovering any systematic code errors.
RES has sought national and internatior 'l cocperation in the development and use of integral and separate effects f.
elities and in advancing methods for code assessment.
The program has been bold in terms of international arrange-ments and innovative in regard to special features.
Unfortunately, as presently envisioned, RES will not complete the code assessment program for another four or five years.
At this time, no version of any best-estimate code has undergone a complete code assessment.
Partial evaluations, however, are expected in the near future.
LOFT has played an important role in providing integral system data for the RES code development and assessment data base.
LOFT testing has provided evidence that the present licensing methods for large-break LOCAs (using evalua-tion models) appear to proviae conservative results.
On the other hand, the NRC commitmant to the public3 to quantify the LOCA safety margins has still not been achieved.
In fact, the applicability of best-estimate codes to PWRs has not been independently assessed.
The LSRG concludes that:
(1) the present approach to code assessment must be expedited; (2) the National Laboratories should initiate a more coherent program of code assessment; (3) the assessment of the current version of best-estimate codes should be completed promptly for limited LOCA conditions which do not result in any si: ificant fuel damage; and (4) NRR must play a more active role in code-assessment.
LOFT data are presently used by NRR in the assessment of its working calculational tools and those of licensees, nuclear steam supply system (NSSS) vendors, and reload fuel suppliers.
Completed code assessment for LOCAs will define code uncertainties and should effectively identify the type and need for any further LOCA testing.
Any abrupt or severely curtailed program for LOFT LOCA testing could well place the completion of code assessment in jeopardy.
Flexibility must be maintained in LOFT LOCA testing to provide the necessary data base to fulfill the need for code assessment.
- Table 4.A is extracted from a Presentation by L. S. Tong entitled, "USNRC Research Program," before the International Atomic Energy Agency, International Conference on Current Nuclear Power Plant Safety Issues, October 20-24, 1980.
4-7
4.4.4 Identification of Phenomena An important part of code assessment is revealing the presence or absence of
~
phenomena important to the analysis but not properly accounted for in the coder,.
LOFT has served to identify some of these phenomena.
Examples are three-dimensional flow behavior in the downcomer and clad rewetting, which were observed in large-break blowdowns.
Still under review are the significance of two phase natural circulation and steam generator heat transfer rates on small-break behavior.
i For some transients, the response of the pressurizer requires further attention.
In addition to. tests that uncover new phenomena, tests that confirm previous understanding and proceed as predicted with no " surprises" are equally important in understanding plant behavior.
4.5 Additional LOFT Testing Needs In the preceding sections of this chapter, W cole of LOFT in the code-assessment This process and the potential impact of scal % distortions were discussed.
section will address the additional LO" tasting believed to be necessary to resolve assessment issues, either for weral assessment purposes or for specific licensing needs.
It will also consider the expected impact on code assessment and/or licensing if the data are not available.
For the purposes of this discus-sion, the testing needs will be grouped in three separate categories:
large-break LOCAs (intermediate-size or transition-size breaks are included in this category);
sm611-break LOCAs; and anticipated transients and non-LOCA accidents.
The LSRG endorses additional testing with LOFT for those tests in which the unique features of a LOFT-type facility (nuclear and larger scale) are considered highly superior and/or necessary compared with other facilities (such as Semiscale).
As discussed in Section 4.3, testing needs were developed from a variety of sources; these included:
(1) NRR; (2) National Laboratories involved in code assessment; (3) RES and EG&G Idaho; and (4) LSRG.
The LSRG believes that the additional testing needs from LOFT must be accurately identified.
Once the needs are known, an optimum decommissioning date can then be identified.
It was with this assumption and approach that the LSRG evaluated the three options identified in Chapter 1.
4.5.1 Large-and Intermediate-Break LOCAs A number of large-break LOCA simulation tests have already been run with LOFT.
These include L1-1 through L1-5 (nonnuclear) and L2-2 and L2-3 (nuclear).
However, these tests do not cover a wide range of initial conditions and break characteristics.
Because the range of assessment is only as good as the range of the test data from which the assessment was performed, it is believed that at least one test is needed with initial conditions sufficiently different from previous large-break tests to provide information on the range of predicted pheno _ena.
In addition, the LOFT large-break LOCA tests run to date were per-m formed with unpressurized fuel.
Because today's large PWRs have prepressurized fuel, at least one additional test may be needed with prepressurized fuel and a high linear heat generation rate.
4-8
Furthermore, the two-loop Westinghouse plants have an ECCS design feature called
-upper plenum injection (UPI) in which ECCS water is injected into the upper plenum of the reactor vessel during a LOCA.
A test with UPI has been run in Semiscale; however, the small size of Semiscale may have masked the three-dimensional flow behavior expected.
Because of the LOFT scale, at least one
-UPI large-break LOCA simulation test should be run for code assessment and exploration of alternate ECCS.
In addition, at least one large-break LOCA replicate test is recommended to confirm the reproducibility of test data.
Jk) intermediate-size break simulations have been performed with LOFT.
Because an intermediate-size break LOCA represents the crossover between the small-break and large-break LOCA, at least one and preferably two ir.6ermediate-break tests are considered necessary.
In summary, four large-break LOCA and two intermediate-break LOCA tests are considered desirable to complete the large-and intermediate-break LOCA testing necessary for code assessment.
(See Table 4.5.)
4.5.2 Small-Break LOCAs Six small-break simulation tests have been run with LOFT:
(1) L3-0:
run shortly after the THI-2 accident to simulate the event (however, the core was not generating power);
(2) L3-1:
simulated a small, cold-leg break and was used as a required problem for industry code assessment; (3) L3-2:
simulated a very small cold-leg break and was used to evaluate the effect of slow system depressurization; (4) L3-7:
evaluated the effectiveness-of various heat sinks available to PWRs; (5) L3-5:
simulated a small, cold-leg break in the intact loop with the reactor coolant pumps (RCPs) off, designed as a counterpart to L3-6; and (6) L3-6:
simulated a small, cold-leg break in the intact loop, with the RCPs operational.
L3-6 was also used as a required problem for industry code assessment.
Since test L3-6 was conducted in December 1980, some consideration has been given to the need for a second similar test with LOFT.
The principal reason is that the RELAPS and TRAC predictions of L3-6 showed large differences from the test data.
The initial post-test analyses identified a number of code
" improvements" necessary to better predict the data.
These included correction of a " mass balance" error and a significant reduction in the RCP two-phase flow degradation curves (head and torque).
It cannot be ascertained from one test if these improvements represent a general improvement in the codes' capability, or if these improvements are merely " tuning" the code to predict a specific test.
A similar test with different initial conditions (that is, initial power, temperatures, break sizes, and so forth) might resolve this question.
4-9
The LSRG believes test L3-3, which is designed to simulate a locked-open PORV, may be needed to study a vendor proposed accident mitigation system
- and to explore the relieving capacity of PORVs.
Other than the two small-break LOCAs identified above, the LSRG believes that no other small-break testing is necessary with LuFT, pending progress on the code-assessment work.
(See Table 4.5.)
4.5.3 Transients and Non-LOCA Accidents A key use of LOFT is to examine and test for new phenomena that may occur in operating plants.
Recent examples of how LOFT can be and is used in this process are the olanned simulation of the Arkansas Nuclear One, Unit 2 (ANO-2) turbine trip of Januacy 29, 1980 and the St. Lucie natural circulation cooldown event of June 11, 1980.
It is planned that LOFT will simulate these two events in order to better ascertain the actual phenomena that occurred; this will be done through LOFT's instrumentation, which is better and more extensive than that on the large PWRs.
The results of these tests will both help establish the typicality of LOFT to large commercial PWRs, and help confirm previous staff assessments of the significance of these events.
Presently, both LOFT and Semiscale are used for transient simulation.
Therefore, early termination cnd decommissioning of LOFT would not totally eliminate this simulation cspability, but it would obviously limit the ability to assess future events in operating reactors.
The LSRG evaluation concludes that additional testind is needed in the transient and non-LOCA accident area.
The testing needs cover : cpectrum of events and scenarios.
The LSRG believes that seven to eight separate operational transient tests should accommodate the present NRR-identifiable needs.
These tests are identified and described in Table 4.5.
It is recognized that some unusual operational transient conditions may arise from daily operations of reactors.
Consequently, situations may occur that require tests with LOFT to clarify the phenomena in more detail to assist NRC in confirming licensing decisions. When such needs are justified, the LOFT program should have the flexibility to accommodate additional tests or to replace some previously planned experiments.
4.5.4 Summary of Recammended Testing with LOFT This section summarizes the testing program recommended by the LSRG that should be completed with LOFT before decommissioning.
Chapter 6 of this report addresses the need for and desirability of degraded core cooling tests.
For convenience, Table 4.5 incorporates the recommended testing in the area of degraded cere cooling.
AThe " Accident Override System" is still under development by Westinghouse.
The system under evaluation is similar in concept to the automatic depressuri-zation system (ADS) used today in boiling water reactors (BWRs).
1 4-10 E
The'LSRG has identified four additional large-break LOCA~ tests necessary to complete the' data base-for verification of large-break LOCA models.
Two intermediate-break LOCAs havelalso been identified as necessary for code
. assessment.
Two additional small-break tests should be performed with LOFT to complete the RCP on/off testing needs and to, study a vendor proposed accident mitigation system and PORV relieving capacity.
Based on the unique attributes
-of the LOFT facility, an estimated seven to eight transient'and non-LOCA accident tests with LOFT should accommodate present NRR needs.
.The LSRG believes that.the testing identified above can be completed by the end of FY 1983.
However, because it is'possible that certain of-these tests may not be completed during_this time, the LSRG evaluated test priority. -This evaluation was based on test needs as perceived '
the LSRG and on the benefits
~
. expected from each test.
As shown in Table 4.'i, three priorities were assigned:
.high, medium, and low.
The LSRG believes that as many as possible of-the tests identified in this section should be completed consistent with LOFT decommis-sioning at-the'end of FY 1983.
The tests should be scheduled according to priority, with the high priority tests being_done first.
Within each of the priority groups, the test sequence will require further evaluation by RES and EG&G Idaho.
The need for core changeout following~these tests was not evaluated.* Therefore, the feasibility of performing the tests in the recommended priority groups must also be examined.
4.5.5 _ Impact of Reduced Testing Program The LSRG optimum test program (identified above) is designed to derive optimum benefits from the unique features of LOFT before it is decommissioned.
- However, in case this recommended program cannot be realized, an abbreviated program has been' developed.
This abbreviated program would be compatible with Option B in Chapter 1 (that is, decommissioning LOFT at the end of FY 1982).
This section identifies the key tests (previously identified in Sections 4.5.1 through 4.5.3) which should be retained as part of this program.
It is was 4
assumed that decommissioning at the end of FY 1982 would allow approximately six.to eight additional tests (two to three in the remainder of FY 1981 and four to five in FY 1982) depending on the type of test and the degree of com-plexity in running the test.
Table 4.6 lists eight tests which the LSRG would recommend be run with LOFT before decommissioning at the end of FY 1982.
In general, it is believed that the large-break tests identified will provide the minimum data necessary to complete the assessment process.
The confidence level i
of the ascessment would be reduced by untimely decommissioning, and this reduced-l-
confidence could ultimately restrict or reduce the degree to which the safety I
margins of Appendix K could be quantified.
Failure to complete all of the recommended transient and other non-LOCA accident
. tests would also reduce the degree to which models and methods used to analyze these' events can be tested. To make up for a lack of experimental information,
" Existing externally mounted incore thermocouples may not read fuel clad
_ temperatures correctly; therefore, internal thermocouples should be installed in any new fuel rods before they are used for LOCA tests.
4 '
~~
it has always been NRC staff practice to impose conservative features on analysis models to bound uncertainties.
The ultimate result of a reduced data base could be (but not necessarily would be) accommodated by retaining the present, or possibly imposing additional, conservative margins in the analysis models.
It is also anticipated that some of tests which would not be run with LOFT would instead be run with Semiscale.
This approach would ensure that the necessary data were obtained; however, the applicability nf these data to large PWRs would have to be better assured because of the larger scale differences.
If LOFT were to be decommissioned at the end of FY 1981 (Option C), the LSRG believes that only one or two more tests could be accomplished.
The LSRG recommends that any tests performed bc large-break 'LOCAs.
4.6 Findings
(1) The large-break LOCA program with LOFT has proven to be very successful and should be brought to an orderly completion.
The LSRG has identified what it believes to be the LOFT testing needs.
These tests should be performed in accordance with Table 4.5 and should be consistent with a decommissioning at the end of FY 1983. To accomplish this, further evaluation by RES and EG&G Idaho of both test sequencing and the need for the core changeout is necessary.
(2) NRR must take a formal and more active role in identifying LOFT testing needs.
(3) RES code-assessment efforts must be improved to make better use of LOFT data and to provide NRR with useful results on a more timely basis.
4.7 References The references cited below are available for inspection and copying for a fee in the NRC PDR, 1717 H St., N.W., Washington, D.C.
Documents marked with an asterisk also are available for purchase from the NTIS, Springfield, VA.
(1)
U.S. Code of Federal Regulations,10 CFR 50, Appendix K, "ECCS Evalt:ation Models."*
(2)
U.S. Code of Federal Regulations, 10 CFR 50.46, " Acceptance Criteria fo" Emergency Core Cooling Systems for Light Water Nuclear Power Reactors."*
(3) Opinion of the Commissioners in the Matter of Rulemaking Hearing, Docket No. RM-50-1, " Acceptance Criteria for Emergency Core Cooling Systems for Light-Water-Cooled Nuclear Power Reactors," CLI-73-39, 6 AEC 1085 (1973).
(4) " Proposed Action Plan for Modifying the Emergency Core Cooling System (ECCS)
Rule in 10 CFR Section 50.46 and Appendix K to 10 CFR Part 50," SECY-78-26, dated January 18, 1978.
1 4-12
i I-
[
.(5).. Memorandum from Edson G. Case (NRC) to Saul Levine (NRC),
Subject:
'NRR-Requirements for. Loss-of-Coolant Accident. Analysis Computer Programs (RR-NRR-77-5), dated June 23, 1977.
(6) Memor'andum.from E. G. Case (NRC) to S. Levine (NRC),
Subject:
Request for Confirmatory Research-Project Related to the Behavior of Pressurizer Safety /
Relief Valves During Subcooled Discharges (RR-NRR-79-02), dated-January 17, 1979.
- (7) Memorandum from Harold 'R.. Denten (NRC) to Saul I avine -(NRC), f abject:
NRR' User Need for ECC Standard Problem Program (RR-NRR-79-20), dated
' August 17,'1979.
(8) Memorandum from Harold Denton (NRC) to Saul Levine (NRC),J ubject:
Request S
for Confirmatory Research on LWR Heat Transfer (RR-NRR-79-20), dated August 17, 1979.
(9) Memorandum from Harold Denton (NRC) to Saul Levine (NRC),
Subject:
User-Need Concerning Two-Phase Natural Circulation and Pump Performance (RR-NRR-79-30), dated December 27, 1979.
(10); Winters and Lambert, "Large Break Transient Calculations in a-Commercial
~
'PWR and LOFT Prototypicality Assessment," EGG-LOFT-5093, dated April 1980.
(11) Letter from N. C. ~ Kaufman (EG&G) to-R. E. Tiller (DOE),
Subject:
Suitability' of L3 1 as a Standard Problem, dated August 12, 1980.
4-13 u
n. m
m.. - _, _ -,
-.m j.
e r
Table 4.1 -Original-LOFT Test-Series-Test Series Designation Description L1-Large-break tests (nonnucle,ar)
Large-break-(cold leg) test's L2-
-L3 Small-break and intermediate-break tests i
L4 Alternate ECC system tests-L5 Large-break (hot-leg) tests t
L6 Non-LOCA transients L7 Steam generator _ tube ruptures 4
\\
t
[
9 l
l r
4-14
Table 4.2 Current LOFT Test Plan (FY 1980-1986)
D:signation Target Date Description / Comments f
L3-6 12/10/80 I
Small-break (2.5%), intact-loop, cold-leg-pumps on.
(Complete)
Pumps tripped at end of experiment to measure water remaining.
L8-1 12/10/80 Core uncovery without ECC at low decay-heat level.
(Complete)
Add on to L3-C.
L9-1 3/4/81 Loss of all feedwater (multiple failures) with scram on high pressure; PPS setpoints representative of LPWR (PORV challenged); mild ATWS L3-3 4/8/81 Small cold-leg break (0.16%), HPIS flow approximately equal to break flow; dry steam generator secondary.
Determine the boundary between break-heat removal and PORV heat removal.
Needs further justification.
CV Leak Test 6/81 Required test of containment leak integrity.
L6-7 7/81 LOFT typicality to Arkansas Nuclear One turbine trip test.
L9-2 7/81 Rapid cold water accident, upper plenum voiding.
Add on to L6-7.
(St. Lucie natural circulation cooldown event.)
LS-1 8/81 Intermediate-size break (accumulator line).
Determine if large-break and small-break models continue to predict intermediate-break results.
Also check out liquid level device.
L8-2 8/81 Core uncovery at high decay heat level.
Reflood with with degraded ECC capability.
Add on to L5-1.
Whole Core 10/81 F1 center bundle at 350 psi (BOL).
Large peaking factor Changeout if only CB changed.
L2-5 1/82 Worst prototypic hydraulic conditions in core.
Investi-gate fuel behavior at BOL fuel pressure (no fuel damage expected).
Replace 3/82 F2 will be pressurized to 700 psi.
Center Bundle L2-6 5/82 Same as L2-5, with 700 psi fuel pressure (E0L).
Fuel damage and fission product release expected.
Replace 7/82 Al in, unpressurized.
Only minimal fuel damage experi-Center Bundle ments can be done until F1 is examined for damages.
LS-2 9/82 Intermediate-size break on hot leg.
Pressurizer surge line.
Needs further justification based on L5-1.
L6-4 9/82 Uncontrolled rod withdrawal at power.
Investigate v1rst b
ase moderate frequency accident.
L9-3 12/82
^"WS.
Loss of feedwater is initiating event (multiple failures).
4-15
b Table 4.2 (Continued)
'D:signation Target Date Description / Comments
'L9-4 3/83 ATWS.
Loss of offsite power'is initiating event (multiple failures).
R: place 5/83 Put F1 back in at 350 psi.
(F1-inspection completed and fuel is assumed not damaged.)
Canter Bundle L8-3
-8/83 Small break with slow core heatup (1*F/ min).
Uniform clad swelling and blockage of flow channel.
Investigate potential initiating events.
(Candidate:. loss of feedwater)
I' R: place 8/83
.F1 out, A3 (unpressurized) in.
Install well-instrumented steam generator (SG).
Center Bundle L7-1 12/83 Large break with SG tube rupture at start of reflood/ refill (> 25 tubes ruptured).
Provides upper bound of envelope on effect of ruptues.
Critical number of tube ruptures resulting in extreme core temperatures I
expected to be between 10 and 25 based on Semiscale results.
.L7-2 2/84 Large break with SG tube ruptures at start of reflood/ refill (<10 tubes ruptured).
Provides a lower bound of envelope on effect of ruptures.
L7-3 should be inserted if possible which has critical number of ruptures.
4 L4-1 5/84 200% cold leg break. Accumulator injection into upper-plenum.
Investigate topdown core quench.
Applicability to UPI plants.
L4-2 8/84 1200% cold leg break.
Upper plenum LPIS injection.
Investigate W two-loop plant phenomena.
Replace 12/84 A3 out.
F3 (pressurized) in.
Center Bundle L8-4 3/85 Severe core damage.
Investigate potential. initiating events.
(Candidate:
Whole Core 4/85 F4 center bundle.
Changeout L10-1 7/85 Override test. Override of L8-3 transient.
L10-2 9/85 Override test.
Override of L8-4 transient.
L8-5 11/85 Severe core damage.
Investigate potential initiating events.
(Candidate:
steam line rupture.)
Decommission 12/86
- Designation:
L2 Series - large breaks L6 Series - anticipated transients l
L3 Series - small breaks L7 Series - LOCA with steam generator (SG) _ tube rupture L4 Series - alternate ECCS L8 Series - core uncovery and fuel damage L5 Series - intermediate breaks L9 Series - transients with multiple failures L10 Series - accident override 4-16 l
Table 4.3 Some Major Scaling Distortions in LOFT Item Distortion comment Core height LOFT core.is 5.5 ft high, LOFT core should quench earlier commercial PWR cores are during reflood; EG&G maintains 12 ft high.
PCT not function of core height.
Core Elevation of core with LOFT core lower in vessel than respect to other loop commercial cores.
Small elevations.
breaks calculated to produce core uncovery in large PWRs, would not do so in LOFT.
Hot-leg to cold-Recent small-break test Does not affect large-break leg bypass (L3-1) indicates bypass behavior; may have signifi-flow larger in LOFT than cant effect on small-break in large PWR.
behavior, particularly inventory behavior.
Externally mounted Provides " fin" effect on Cladding temperatures measured thermocouples on fuel surface; may promote in LOFT may be significantly fuel rods quenching and cooling.
different from cladding temperatures than would be measured without the fin effect.
Inactive com-Broken loop cannot repre-Thermal-hydraulic stability ponents in sent pumped flow or steam effects due to parallel loops broken loop generator heat transfer.
during small breaks and anti-cipated transients not represented.
Steam generator Lower (800 psi) than in LOFT steam generator will be secondary side large PWRs (900 psi).
more effective heat sink than pressure large PWR steam generator.
Core Core shape and size, Affects reactivity feedback coefficients, leading to atypical nuclear kinetic response.
This reduces the value of tests such as ATWS.
Wall surfaces Atypical surface to Leads to stored energy in walls mass ratio to be transferred to blowdown and ECCS water.
Must be compensated for by insulation.
Downcomer Gap between vessel and Atypical downcomer thickness flow skirt could cause wrong flow patterns during blowdown.
4-17
I LOCA ECC EXPERIMENTAL PROGilAM M ATRIX Tabic 4.4:
(
I s....
....i ei. i, intrep el 4 nicas S ole System filow luwn. c.it fiellemal 48.. s I CC llyp n l' nno mal Stren Uppe Plenum Me nirl Seic r
De entr ainnwns t e mnire le.nntre UI in & llel.II t.rurt aiin Japanese SC', F
'tt -i t'lll FRG UPTF OllNI Upsli t Hr. upil Japan SCTF I
tJeevly Full
GE 1L1 A-I
)
INELtOFT ltall iUI I
" I I UI I l'"'util Ja an CCTF loieni eiteate INEL LOF T INELLOF1 Jeanese CCTF J.apei CCil 8f 81 i L Of i S (>
State Jepsi CClF IICI ;*/18. M..s. 4
' W' I. * ""w d I.'""I'l F L EClli St ASE T UCI t/15 Miiiti i Ifil l Sennu.s8e thi l. Senusr.se INL L Senusale Cara r i. th M'al. f W IllCllT St ASET Ital i Ses Lippe Plenum Sn. it
- 1 pmi HOSA
- ISPH A UDill sc,le UE ILTA 2 S G.
ANL Feron Law se
- (W sulm IlDilli Mil t arene O AH IMou Hl
- t..
ee ri PHI flaswell time
- lMII IasminLsm.pl Al CL Mo let Orv g,g g M. lel
, lilaeunives I scunt IIPl M.ntcl l ess 5.,3y I clingli Memlet irst i I Cuenparteil Pnwp n l l Pros ned Psopom Nini NitC Fusaluess l
l
l Table 4.5 Recommended Test Program Supporting LOFT Decommissioning at the End of FY-1983 In order to meet current regulatory needs, the LSRG recommends that the following series of tests be performed with LOFT.
While this table indicates a basic priority for these tests, the detailet planning and scheduling (including the need for core or center bund e changeout) be performed by RES and EG&G Idaho in a manner that is consistent with decommissioning at the end of FY 1983.
A.
Test Priorities Large-Break Intermediate-Break Small-Break Transients and Priority LOCA LOCA LOCA Non-LOCA Accidents High L4-2 L5-1/L8-2 L9-2/L6-7 LA-1 LA-2 L9-3*
LA-3*
Medium L2-6 LA-5 LA-4 L3-3 Lcw L6-4 L9-4 LA-6 LA-7 LA-8 Note: Where the current LOFT test program specifies a test scenario ttat is endorsed by the LSRG, that designation is used (for example, L4-2, L5-1, and so forth). Where the LSRG has identified testing needs which do not fall within the scope of the present LOFT test program, these tests are designated by LA (for example, LA-1, LA-2, and.so forth).
4-19 u
Table 4.5 (Continued)
~
B.. Test Descriptions
-Designation Description-Comments High Priority L4 LBLOCA with UPI Designed to verify ECCS~models with UPI; most desirable to run this test in LOFT because 3-D flow distribution effects should be retained.
-LA-1 LBLOCA with offnormal
~Needed to ensure that data base is initial conditions adequate for code-verification.
L5-1,.LA-2 IBLOCA Needed to complete the break spectrum being examined with LOFT.
Flow-regime transients may be extremely important for this type of LOCA.
Tests are best performed at LOFT because of nuclear heat and scale.
L8-2 Core uncovery test with
-Fuel response is considered dependent upon high decay heat heatup rate.
L8-1, conducted in 12/80, examined the low decay-heat area.
This test is scheduled as a piggy-back test with L5-1.
L9-2 ANO-2 turbine trip Designed to simulate the ANO-2 transient transient in which the plant experienced a stuck-open steam dump valve and pressurizer spray valve following a. turbine trip test.
L6-7 St. Lucie cooldown Designed to simulate the St. Lucie cooldown transient transient in which the plant experienced void formation in the reactor vessel upper head during natural circulation cooldown.
Scheduled as a piggy-back with L9-2.
- L9-3, LA-3 ATWS induced by loss-L9-3 is designed to be the initial ATWS of-feedwater test and therefore should be a " mild" test.
If required, LA-3 would provide a follow-on and "more severe" test.
Medium Priority L2-6 L8LOCA with Designed to be conducted at high power prepressurized-fuel (16 KW/ft) with prepressurized fuel.
Some fuel damage and fission product release is expected.
Efforts should be made to obtain Zr-H 0 reaction data.
2 f
LA-4 LBLOCA repeat test Needed to confirm the repeatability of i
thermal-hydraulic behavior during LBLOCAs.
f 4-20
i Table 4.5 (Continued)
Designation Description
-Comments LA-5 SBLOCA with RCPs Needed as a follow-on to L3-6 for code i
operating.
assessment.
Test should be performed in a similar manner to L3-6 howeve*, initial conditions should be changed.
L3-3 SBLOCA with locked open Designed to evaluated the concept of using-PORV a depressurizing accident mitigation feature similar to the automatic depres-surization system (ADS) used in boiling water reactors and to evaluate PORV relieving capacity.
Low Priority L6-4 Continuous rod withdrawl Designed to produce a departure from nucleate accident (at power) boiling (DNB) condition on the fuel to help quantify the conservative margin imposed by present assumption that all fuel that enters the DNB region fails.
Because the charac-teristics of this event rely on the nuclear response of the fuel, LOFT is the only facility considered suitable for this test.
L9-4 ATWS induced by loss-of-Designed as a follow on to L9-3.
The offsite power response will differ from L9-3, due to loss of RCPs upon loss of-offsite power.
LA-6
" Uncontrolled" boron Because characteristics of this event rely dilution event on the nuclear response of the fuel, LOFT is the only facility considered suitable for performing this test.
LA-7, LA-8 Steam line break Needed to examine the return to power phenomenon and mass release from steam generators (steam separator effective-ness).
Two tests are considered, with and without offsite power.
l 4-21
Table 4.6 Recommended Test Program Supporting LOFT Decommissioning at the End of FY 1982 Designation Description Comments L4-2 LBLOCA with UPI See Table 4.5 LA-1 LBLOCA with offnormal See Table 4.5 initial condi^. ions L5-1/L8-2 IBLOCA/ core uncovery See Table 4.5 test with high decay heat LA-2 IBLOCA See Table 4.5 L9-2/L6-7 ANO-2 turbine trip See Table 4.5 transient /St. Lucie cooldown transient L9-3 ATWS induced by loss-See Table 4.5 of-feedwater LA-5 SBLOCA with RCPs See Table 4.5 operating L2-6 LBLOCA with See Table 4.5 prepressurized fuel 4-22
5 OPERATIONAL ASPECTS l
1 5.1 Introduction In an emergency situation invol.ving the operation of a nuclear power reactor,
. licensed operators must be able to comprehend quickly from control room instru-mentation what is happening and why.
One of the potential emergency situations with which licensed operators must be prepared to cope is the LOCA.
Because of the many reactor years of operating experience that the industry has had and because of the use of simulators in training reactor crews, it had been assumed that if a LOCA did occur, operator response would be adequate.
Never-theless, during the event at THI-2, the time required for operators to compre-hend the problem was long; hence, operator response was initially misdirected and the appropriate operator response was late.
Reasons for the delay in comprehension included, among other things, the design of the control room human-machine interface and deficiencies in operator training.
As discussed in Chapter 3 of this report, the LOFT program was initiated long before the accident at TMI-2.
Before that event, the objective of the program was to obtain data from the LOFT reactor to be used to assess the validity of computer codes use to predict the performance of commercial PWRs during a LOCA.
After the TMI accident, it became important to RES to determine how LOFT or LOFT results could be used in research on the control room human-machine inter-face as it relates to LOCAs and other transients and accidents (hereafter col-lectively referred to as transients).
Further, RES determined that LOFT uight also be used as a test bed for the development of instrumentation to be used in diagnosing conditions of inadequate core cooling.
Work in the areas of human-machine interface and instrumentation development and testing have been initiated at LOFT.
This chapter addresses the question of whether or not this work should be continued.
5.2 Human-Machine Interface The LOFT reactor and its normal and emergency cooling systems are designed and scaled to physically model those aspects of large PWRs that are important to system performance during a transient.
Results of experiments performed with LOFT can be used as integral tests of analytical models utilized for predict-ing the behavior of commercial PWRs during transients.
These experiments are performed every few months, after test procedures are written, calculations to determine expected results are performed, and the operators for each specific experiment are trained.
The LOFT control room is smaller than some control rooms at commercial plants, and it has fewer control and instrumentation panels.
The control room is arranged so that key operating personnel are positioned a few feet apart.
Flanking the reactor cor. trol panel on the left and the rignt are control panels for the secondary cooling system and the ECCS.
During test L3-6/L8-1, which the LSRG observed, an operator was positioned at each of these three panels; the shift supervisor was positioned nearby, where he could observe the control panels and direct the' actions of the operators.
This situation differs from that in the control room of a commercial reactor, where the reactor, secondary cooling system, and ECCS control pane:1s are not necessarily immediately adjacent at d r:nly the operator may be present.
5-1
During test L3-6/L8-1, six persons were monitoring readouts from thermocouples and electrical conductivity probes located in the core, reactor vessel, and i
downcomer; a seventh person was monitoring the performance of the reactor coolant The test generally proceeded according to the countdown and checklist pumps.
schedule', with most actions taken in sequence and at the expected time.
)
To enhance the capability of LOFT operators, an augmented operator-capability (A0C) program has been initiated.
An important part of this program is the design, installation, operation, and continued development of the operational i
~ diagnostics and display system (0DDS).
This is a computer-based system which acquires data from the existing plant computer.
The ODDS stores data and on demand, displays current or_ previous data in various formats.1 Development of additional information displays and supporting system software is proceeding concurrently with. a functional analysis of operator tasks for various plant conditions.
Criteria for selection of information displays include:
(1) frequent use during. normal operation; (2) support for specific plant evolutions; and (3) potential use during operational transients.
Further, a new display mus: complement the conventional process instrumentation.
Use of predictive displays has been considered, and one will be installed in 1981 for steam generator water level.2 Water level will be plotted on a video screen that will show the actual level for the last 60 seconds and the predicted level for the next 60 seconds (The predicted values will be obtained by fast-time simulation).
In addition, ODDS has the capability for displaying on video screens simple process system schematic diagrams which show values of parameters and the status of components.
Color computer graphic techniques developed at '.0FT may be useful at commercial reactors.
Indeed, color computer graphics are being independently developed by some-licensees and NSSS vendors to enhance the effectiveness of operators at fossil and nucler power plants during normal operations and,-if necessary, during an emergency.
RES has expressed that the LOFT reactor can be used for researching and optimizing that part'of the human-machine interface which is bounded by control room instrumentation necessary for operator reaction to transients.
The basic require-ments for a facility to do this work are:
i l
(1) a realistic operational environment and the ability to reproduce the ' tests using a variety of subject operators; (2) the capability to generate realistic values of commercial plant parameters in real time for a' spectrum of transients; and i
(3) a control room with data displays and system controls which can be recon-
_ figured in a relatively short time.
l l
i I
5-2
Given a facility that meets these requirements, the responses of many operators in several control room configurations to various transients can be tested.
Because of the variability of human response, many such tests need to be run.
In principle, the LOFT reactor can be used for this purpose; however, for the following reasons that approach is not considered to be practical:
(1) an appreciable amount of time is needed between Msts to prepare for the next test, thus limiting the number of operator responses per unit time that can be observed; (2) a serious operator error during a test could conceivably damage the system beyond repair; (3) training operators before testing would bias the test results; and (4) the control room does not appear to be amenable to easy reconfiguration.
Ex. ting operator training simulators should be upgraded to incorporate LOFT transient results.
Responses of operators in the various simulated control rooms should be compared to identify those human-machine interfaces that lead to better operator responses to transients.
Further, if a special purpose simulator is designed to simulate the response of a commercial PWR to a spectrum of transients and if the design includes a control room that can be easily reconfigured, then much more human machine interface research could be done than is possible with LOFT.
The LSRG has been informed that the human-machine effort (A0C) in the LOFT program is funded at $3.8 million for FY 1982 and FY 1983.
Independent of LOFT, the plant operational safety decision unit in the long range research budget has a human machine interface subelement which has projected funding of $39.7 million for FY 1982 through FY 1987.*
5.3 Development of Instrumentation As a result of the accident at TMI-2, the staff required:
(1) that licensees for PWRs install subcooling meters and (2) that additional instrumentation for detection of inadequate core cooling be designed and installed.
As part of this latter requirement, licensees were asked to consider installation of both thermocouples in fuel-cooling channels and PWR vessel liquid level instrumentation.3 Thermocouples in fuel-cooling channels have been used in various reactors for many years to indicate fuel performance.
Reactor vessel level instrumentation, however, has not been used in PWRs because the vessel is normally co:apletely filled with water.
The industry is now developing level instrumentation for this purpose.
Measurement techniques that are being pursued include the use of electrical conductivity probes, pressure transducers, heated thermocnuples, neutron detectors, and waveguides.4'S As part of the lessons learned from Three Mile Island, the NRC staff has been developing and refining the concept of a liquid level signal for the vessel of a PWR.
Several concepts have emerged from the regulated industry and from NRC contractors.
Proof of principle for these devices is needed, and this can best be done in an integral test facility such as LOFT.
RES has in place a vigorous and responsive plan to provide the verification of some proposed vendor and utility devices in LOFT and Semiscale and the ORNL THTF.**
The use of LOFT
- "Oraft Long-Range Research Plan Fiscal Years 1983 through 1987, "USNRC Report NUREG-0740, November 1980.
A copy of the final plan will be publicly available on February 17, 1981.
- 0RNL THTF:
Oak Ridge National Laboratory Thermal Hydraulic Test Facility.
5-3
for this purpose is highly desirable and responsive to the NRR request.
It appears that such verification can be accomplished in concert with the other test sequences mentioned in Chapter 4, and does not require separate tests solely for instrument verification.
LOFT has developed, and uses during loss-of-cdant experiments (LOCEs), axial strings of thermocouples and electric conductivity probes in the reactor vessel and in the core.
LOFT also uses gamma densitometers in coolant piping.
The conductivity probes and the densitometers are used to measure the quality of the coolant during LOCEs and thus are essential in determining the performance of LOFT.
5.4 Findings
(1) The human-machine interface research presently being conducted at LOFT is peripheral tn the LOFT mission.
Because of the limitations inherent in a test reactor, LOFT is not the appropriate facility for conducting this research.
(2) The LSRG sees the role of LOFT in evaluating new instrumentation for use in commercial reactors, such as PWR reactor vessel liquid level indication, as highly desirable and responsive to NRR needs; however, any decision to continue or terminate the LOFT program should be made independent of this role.
5.5 References The references cited below are available for inspection and copying for a fee in the PDR. The document marked with an asterisk also is available for purchase from NTIS and/or the NRC/GPO sales program.
(1) Memorandum from D. F. Ross, Jr. (NRC) to Members and Consultants of the LOFT Special Review Group (LSRG), dated November 7, 1980, Enclosure 6, "The LOFT Augmented Operator Capability Program."
(2) Memorandum from G. D. McP..erson (NRC) to Roger Woodruf f (NRC), " LOFT Response to Man-Machine Interface Questions," December 23, 1980.
(3)
U.S. Nuclear Regulatory Commission, " Clarification of TMI Action Plan Requirements," USNRC Report NUREG-0737, November 1980.*
(4) William J. Dircks, Information Report for the Commissioners, SECY-80-529, "TMI Action Plan - II.F.2, Additional Instrumentation for Measurement of Coolant Level in Reactor Vessel," December 4, 1980.
(5)
L. D. Goodrich et al; EG&G Idaho, Inc., "Special LOFT Features for Improved Monitoring and Survival of LOCA Transients," LO-87-80-130, January 17, 1980.
5-4
L b
6 DEGRADED CORE COOLING 6.1 Introduction As the LOFT test program is presently defined, there are three test series associated with degraded core cooling:
(l; L-8:
Core unct,very and fuel damage (2) L-9:
Transients with multiple failures (3) L-10: Accident override
- The present LOFT test program would progressively examine core response to inadequate cooling.
Some of these tests include fuel damage ranging from moderate to severe.
However, it is not the objective of these tests to produce a major core meltdown that would result in significant core migration, lower grid plate failure, or pressure vessel interaction.
The series is intended to explore only the fringes of impaired coolability of a degraded core and is not intended to explore the meltdown phenomena.
Thus, the LSRG has used this as the definition of degraded or inadequate core cooling in this report.
It is with this understanding that the LSRG considered LOFT's ability to shed further light on inadequate core cooling.
- 6. 2 Related Studies The LSRG is aware that the Commission has established a Degraded Cooling Steering Group that is intended to provide guidance on needed core melt research.
That I
group, whose charter includes mitigation of the complete core melt scenario, may also express an opinion on the role of LOFT with respect to degraded cooling.
In addition, the "NRC Action Plan Developed as a Result of the TMI-2 Accident" (NUREG-0660)1 includes three items related to degraded cooling.
The.<e are:
(1) II.B
" Consideration of Degraded or Melted Cores in Safety Review" (8 subparts)
(2) II.F
" Instrumentation and Controls" (3 subparts); and (3) 1.C -
" Operating Procedures" (procedures for inadequate core cooling (ICC)).
4 6.3 Related LOFT Test Sequences If the present LOFT test program were continued through FY 1985, it would include 11 tests related to degraded cooling as part of the L-8, L-9, and L-10 series.
These tests are listed in Table 6.1.
The date proposed by RES for each test is also included in the table.
^ Accident override refers to a concept for PWRs whereby the operator could cope with an anomalous transient by use of a system such as the automatic depres-surization system presently used in BWRs.
6.4 Evaluation The LSRG examined the tests listed in Table 6.1 to determine if they addressed specific regulatory needs and whether the information planned to be obtained To this end, the LSRG asked the following would provide the expected results.
questions:
(1) Has the NRC defined its information needs related to degraded core cooling?
(2) Is it important to study degraded core cooling scenarios that can be stoppe.1 short of bulk core melting?
(3) If the answers to (1) and (2) above are yes, is LOFT the correct place to execute these studes?
The following section addresses the responses to these questions.
With respect to question 1, the record shows that NRC has defined information needs related to degraded core cooling.
As listed in Section 6.2 of this report, NUREG-0660 clearly states this interest.
Item II.B describes the NRC plan for developing and implementing a phased program to inciude, in safety reviews, consideration of core degradation and melting beyond the design basis.
The plan includes eight specific subitems related to both short-and long-term actions, goals, and requirments related to accidents involving core damage greater than the present design basis.
Item II.F describes the NRC plan for providing instrumentation to monitor plant variables and systems during and following an accident.
Three of the subitems interface with degraded core cooling.
Specifically, these are: (1) additional accident monitoring equip-ment; (2) identification of and recovery from conditions leading to inadequate core cooling; and (3) instruments for monitoring accident conditions (see Regulatory Guide 1.97).
Finally, Item I.C describes the NRC plan for improving the quality of procedures to provide greater assurance that operator actions are technically correct, explicit, and easily understood for normal, accident, and transient conditions.
One subitem deals with the analyses, guidelines, and procedures associated with inadequate core cooling.
However, the predominant interest of the NRC to date has been in features that could prevent the complete melt (that is, core migration below the lower grid plate).
Additional research to help define and specify mitigation features such as passive removal of heat from the containment, hydrogen control and core 2 wj11 retention devices which are coverd by the Advance Notice of Rulemaking not be dealt with by any specified LOFT test sequences.
Characterization of fission product source terms might be enhanced by some of the terminal L-8 tests provided enough of the core was brought to the melting point.
However, the LSRG sees no evidence of an expressed regulatory user need for heat transfer tests in LOFT on a core which is presumed to be in a degraded (partial melt or slumped) state.
Therefore, the LSRG is not recommending that any test involving substantial core damage be conducted.
There is, however, an expressed regulatory need to know more about detection of inadequate core cooling initiation and progression of this type of condition.
j 6-2
~
With regard to question 2, is thrce is an important need to study degraded cooling scenarios that can be stopped short of gross core migration.
Guided by the potential for overall risk reduction, the answer to question 2 is yes, because the risk of core melt accidents, which dominate risk, is reduced in proportion to those scenarios which can be stopped short of core melt.
There-fore, LSRG concludes that the interest expressed by the NRC in recognizing inadequate core cooling and doing something about it is important.
Because the answers to questions 1 and 2 are yes, it is appropriate to ask whether LOFT is the correct place to execute these studies.
As identified above, the information needs relate to the ability of the operating staff of a nuclear power plant to respond to anomalous events that have the potential for core degradation.
To meet these needs, the LSRG believes that stylized scenarios should be avoided.
Instead, the operator's ability to properly react to symptoms rests upon reliable training and instrumentation.
The most effective training for this purpose is simulators.
Simulators are programmed to respond in a manner consistent with results of more exact calculations performed with a larger computer.
The inexactness of simulator programs persuades the LSRG that several experiments exploring the f ringes of inadequate core cooling are worthwhile for benchmarking these programs.
Obviously, the benchmarking can not be performed with commercial power reactors.
Therefore, the LSRG believes that LOFT is the correct facility at which to perform this type of testing.
be used to supplement this effort. Smaller scale facilities such as Semiscale might
6.5 Findings
I (1) Core uncovery tests, which can be add-ons to other tests and are calculated to have a peak cladding temperature (PCT) less than 1500 F, can provide needed confidence in the ability of reactor systems to accommodate inadequate core cooling.
Severe transient tests with a potential for moderate core damage (for example, loss of all feedwater and LOCA with no ECC, that have a calculated PCT of approximately 2000 F) are worthwhile but have the undesirable potential for contaminating the facility, impacting the schedule of future tests, and adversely affecting decommissioning costs.
These tests should be studied for their practicality and for their cost benefit.
If the results of these studies prove favorable, such tests should be planned for the end of the LOFT test program.
(2) The LSRG sees the severe fuel damage experiments, presently scheduled for FY 1985-1986, as peripheral to the LOFT mission.* In addition, the facility does not have the necessary equipment to cope with severely damaged cores.
The additional cost of procuring the needed equipment is not included in the present LOFT budget.
- Note the degraded core cooling decision making that is involved with the multiple rulemaking ventures may create the need for additional tests in the degraded core area.
6-3
i a
j
-6.6 References-The references cited below are available.for inspection and copying for a fee in the PDR and for purchase from NTIS. The document marked with an asterisk also is available from the NRC/GP0 sales program.
U.S. Nuclear Regulatory Commission, "NRC Action Plan Developed As A Result (1) of the TMI-2 Accident," USNRC Report NUREG-0660, Revision 1, August 1980.*
(2)
U.S. Nuclear Regulatory Commission, " Domestic Licensing of Production and Utilization Facilities; Consideration of Degraded or Melted Cores in Safety Regulation (Advance Notice of Proposed Rulemaking)," Federal Register, 7
i Vol. 45, No. 193, pages 65474 - 65477, October 2, 1980.
4 c
i 4
k i
i i
i 6-4 l
L Table 6.1 Current LOFT Tests Related to Degraded Core Cooling L-8 Series:
Core Uncovery and Fuel Damage Test Description Date L8 Core uncovery with no ECCS until heatup 12/80*
L8-2 Longer core uncovery, intermediate break 08/81 L8-3 Small-break with slow core heatup 08/83 L8-4 Severe core damage 03/85 L8-5 Severe core damage 11/85 L-9 Series:
Transients with Multiple Failures Test Description Date Loss of feedwater with delayed scr$m (mild ATWS) 03/81 L9-1 L9-2 Rapid cooldown of secondary with aggravating failures 07/81 L9-3 ATWS initiated by loss of feedwatei 12/82 L9-4 ATWS initiated by loss of offsite power 03/83 L-10 Series:
Accident Override l
Test Description Date L10-1 Override of test L8-3 07/85 L10-2 Override of test L8-4 09/85
- The LSRG witnessed the conduct of LOFT Test L8-1 on December 10, 1980.
6-5 w--
i
- 7. LOFT TESTING IN RISK PERSPECTIVE 7.1 Introduction The preceding sections of this report have discussed the-uses and potential uses of LOFT. testing in the areas of-thermal-hydraulic performance during-accidents, operational safety, and degraded core cooling.
In this section, the LSRG considers the risk reduction significance of testing.
7.2 Risk Significance of Transients and Accidents The risk to public health and safety posed by an operating nuclear power plant is a composite of the risks posed by possible accidents, natural events, and malevolent acts.
The Reactor Safety Study (RSS), WASH-1400,1 analyzed the risks from accident sequences in PWRs and BWRs as a result of hardware failure and human error.
The PWR analysis, which was based on the Surry three-loop
. estinghouse plant, has greater relevance to the LOFT test program.
The RSS W
analysis addressed accident sequences initiated by large-break LOCAs, small-break LOCAs, transients, and a number of special types of failure.
The results showed that small-break LOCA and transient-initiated accident sequences dominated the risk.
Since the RSS, probabilistic risk analyses on four other reactors have been performed, based on the models in the RSS.
These analyses will be published soon,* when the Reactor Safety Study Methodology Application Program (RSSMAP) is concluded.
The four plants studied were:
(1) Oconee 3, a Babcock & Wilcox-designed PWR (large dry containment);
(2) Calvert Cliffs 1, a Combustion Engineering-designed PWR (large dry containment);
(3) Sequoyah 1, a Westinghouse-designed PWR (ice condenser containment); and (4) Grand Gulf 1, a General Electric-designed BWR (Mark III containment).
Tables 7.1 and 7.2 list the risk-dominating accident sequences drawn from the RSS and RSSMAP studies of these six plants.
However, because the RSSMAP studies are in varying stages of publication and final draft review, the tables must be used with caution.
The results of risk assessments should be used judiciously because of differences in the degree of completeness of the several analyses and because of uncertainties in the reliability data base.
' Nevertheless, the analyses are useful for directing research toward risk-significantsubjects.
7.3 Large-Break LOCA Risk The accident at Three Mi M Island dramatized the RSS warning that risk is dominated by transients, small breaks, and human error.
Large LOCAs do not appear to be risk dominant se.luences, but they could be, if poorly designed ECCS were provided or if the calculations on which the ECCS design is based "To be published as Sandia National Laboratories /Battelle Columbus Laboratories,
" Reactor Safety Study Methodology Application Program," NUREG/CR-1659 (SAND 80-1897), Vols. 1 to 4.
Risk assessments such as the RSS can appraise the are in significant error.
quality of ECCS but not the validity of their design.
Therefore, validation of the ECCS design in test facilities such as Semiscale and LOFT is risk significant as long as there are significant design uncertainties to be resolved.
From The level of uncertainty that can be tolerated is difficult to define.
the basic risk perspective, even a 90% certainty that the thermal-hydraulic One does not need the level of certainty design basis is correct is sufficient.
implicit in Appendix K calculations to three or more significant figures.
7.4 Transieat Risks the L6 series of The LOFT test program addresses transients in four series:
anticipated or operational transients, the L8 series of core uncovery and fuel damage, the L9 series of transients with multiple failures, and the L10 series of accident override transients.
The use of the term transients in these three series is not the same as in the discussion in Section 7.2.
7.4.1 L6 Series - Anticipated Transients The L6 series covers the design basis transients which are, in essence, the transients described and analyzed in a plant's Safety Analysis Report and Some success in analyzing these transients with RELAPS supporting documents.The LOFT project has identified a number of difficulties has been reported.
in relating LOFT analyses and tests for anticipated transients to commercial (Some of these are listed as atypicalities in Table 4.3.)
LOFT has PWR data.
only one operating loop, as compared to the two, three, or four loops in a commercial PWR.
Other factors are differences between LOFT and a commercial PWR in steam generator conditions ai.f pressurizer heater response.
These differences have effects which must be assessed or each specific tran-Whatever the differences are, even if the L6 series results can be sient.
related to commercial reactor transients, the series does not address the Those transients which are significant to risk (discussed in Section 7.2).
transients are sequences involving serious system degradation and large release The L6 series could reduce risk only insofar as it of radioactive materials.
confirms the safety analyses which show that these anticipated transients do not lead to such degradation and release.
The TMI-2 accident brought to light an interesting alternative to LOFT confir-mation of these anticipated transient analyses.
The detailed analysis of plant response in that accident would have been virtually impossible if there had not been a reactimeter installed on the plant at the time of the accident.
(The reactimeter is a multichannel test data recorder whose use is optional.)
Recognizing the importance of this device ftd accurate reconstruction of the s
events in a reactor mishap, the NRC has been u.ged to require the use of such dcvices in all operating plants, just as a flignt recorder is required in large aircraft.
The LSRG recommends that recording systems like the retcti-It is likely that the random occurrence of anticipated meter be required.
transients in day-to-day plant operations will contribute to the data base for direct confirmation or rejection of analytical models that describe transients.
7-2
7.4.2 L8 Series'- Core Uncovery and Fuel Damage The L8 and L9 series do address the risk significant transients and accidents identified in Section 7.2.
The objective of the L8 series is to investigate transients resulting in core uncovery and fuel damage.
The L9 series covers anticipated transients with multiple failures.
Between the two series it is fair to expect that a risk-baseu agenda could be attempted.
This agenda 2
includes exploring the limits of natural circulation cooling, the temporary loss of primary and secondary coolant delivery, the limits of small LOCA mitigation, and ATWS (anticipated transient without scram).
However, the LOFT project briefings F inted up the difficulties of a comprehensive L8 and L9 series test program.
These tests are events which severely stress the core and the system. A single test can easi'y damage the LOFT core--and perhaps even the system--so severely that costly and time-demanding replacement is required before another test can be run.
Moreover, very serious questions have been raised about the typicality of LOFT for tests of this class whose upset conditions go even further beyond the normal than the L6 series.
The results of transient analyses and tests reported to date indicate that for conditions far from normal there is wide divergence between analysis and test results.
Although the LOFT reactor has been analyzed extensively to determine how well its design scale to a commercial PWR for large LOCA tests, there have been no reports of a similar scaling analysis with respect to the risk significant transients being discussed here.
7.4.3 L9 Series - Transients with Multiple Failures The LOFT L9 series planning is still in a relatively early phase.
As reported to the LSRG, some event-tree analysis is being used.
The range of initiating events selected for event-tree evaluation include:
(1) loss of offsite power; (2) loss of feedwater; (3) steam dump; (4) turbine trip; (5) accidental opening of the pressurizer relief salve; and (6) uncontrolled rod withdrawal at power or during startup.
The basis for selecting these initiating events is not clear after event-tree evaluation, tests are to be selected based on:
(1) probability of the sequence occurring; (2) uncertainty in the event sequence; (3) uncertainty in the consequences of the sequence; and (4) the severity of the consequences of the sequence.
Factors 1 and 4 combine to measure the risk significance of the sequence; factors 2 and 3 measure the confidence of the analysis that they are risk significant.
L9 series planning does not seem to consider the many analyses of transients significant to risk which have been available for quite some time.
An adequate justification for the L9 test series has not been presented to the LSRG.
7-3
7.5-Degraded Core Risks The TMI-2 accident points up the possibility of addressing degraded core risks in two general categories:
those accident sequences which terminate in a severely degraded but not melted core and those sequences which proceed to full cora melt.
In general, the RSS and RSSMAP do'not make this distinction.
The criteria for " core melt" in the RSS and RSSMAP are such that failure to prevent any significant core damage is deemed core melt, although informal attempts have been made to segregate the RSS and RSSMAP sequences into these two categories.
Estimates are that from 10 to 90 percent of the sequences might be arrestable (like TMI-2); that is, they might be the result of a failure which can be rapidly corrected.
The sequences which proceed to core melt are the ones which dominate risk.
Therefore, if LOFT testing can increase the likelihood of arresting core melts, it can be risk significant.
The risks of degraded core sequences are bminated by full-core-melt sequences because the consequences of. a full-core-melt are potentially many times greater than those other types of acciderits.
Therefore, research significant to degraded core performance should be centered on full-core-melt sequences.
j That research can be on the conditions leading to core damage or it can be on the physical phenomena of the processes of core melt, fission product release, and fission product transport.
7.6 Operational Risks Studies of nuclear power plant risk indicate that about half of the risk is attrit'otable to human error.
The human error is comprised of three classes:
(1) preincident errors of commission or omission, such as improper maintenance or failure to realign a system properly after a test; errors of omission during the incident, such as failing to shift the ECCS (2) injection to recirculation; and (3) errors of commission during the incident, such as turning off a needed system.
The RSS addressed the first and second classes, but not the third.
Most of the assessed risk that is associated with human error is in the first category.
l In addition, many contend that as a result of pessimism in the RSS, the contri-bution of category 2 is exaggerated.
Therefore, it is important to be specific l
If the in describing a human error research project as significant to risk.
l devices--or procedures--are related to preventing or discovering and correcting category 1 errors, they can be very significant to risk.
If they merely reduce the likelihood of category 2 or 3 errors, they may not be very risk significant.
l The LOFT tests pertinent to operational safety are discussed in Section 5 of l
l These tests are directed toward the human-machine interface, and this report.
The work on the human-machine interface the development of instrumentation.
is heavily slanted toward diagnosis of plant conditions during accident evolu-This work does indeed have potential risk reduction significance.
tions.
However, at the present stage, it does not appear to emphasize aspects which 7-4
will reduce the preincident errors (category (1) =bove).
LOFT would not seem to be appropriate for this work because LOFT is far simpler than in a commercial plant, whereas the goal of the work is to provide a practical status display and review for many complicated systems.
- 7. 7 Findings (1) Large-break LOCA:
10FT testing for validation of large-break LOCA analysis is risk significant as long as there are significant design uncertainties to be resolved.
(2) Anticipated transients:
The anticipated transients in the LOFT LG test series are not very significant from a risk perspective.
Useful data for investigation of these transients could be obtained by requiring a reliable data recorder on all operating plants.
(3) Severe transients:
The more severe transients in the LOFT L8 and L9 test series are risk significant, but there are serious doubts about the practicality of doing. -h tests with LOFT.
(4) Degraded core cooling:
LOFi
'ing which contributes to the ability to arrest core me't sequences ca..
- -k significant.
Severe core damage and fuel melt tests are risk signi....
but do not appear to be practical with LOFT.
(5) Operational rish:
Thare is a wide variety of w
,,,9 related to opera-tions and huma.a error which can be important to reduction of risk.
LOFT is engaged ir, only a few areas of such testing and is not an attractive facility for more.
i
- 7. 8 References The references cited below are available for inspection and copying for a fee in the PDR.
The document marked with an asterisk also is available for purchase from NTIS and/or the NRC/GP0 sales program.
(1)
U.S. Nuclear Regulatory Commission, " Reactor Safety Study - An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants," WASH-1400 (NUREG-75/014), October 1975.*
(2) Memorandum from F. H. Rowsome (NRC) to L. S. Tong (NRC),
Subject:
Accident Sequences for Experimental Investigation, dated October 16, 1979.
1 l
l 7-5
Table 7.1 Risk-Significant Sequences - PWRs Release Plant Category
- Sequence ** Probability Description
-6 Surry 2
TMLB'-6 2x10 Station blackout with loss of AFWS 2
TMLB'-y 7x10 Station blackout with loss of AFWS
~
-6 2
V 4x10 Interfacing-systems LOCA
-0 3
S C-6 2x10 Small LOCA, sprays fail, no water for 2
ECC/ spray recirculation
-6 Sequoyah 2
S HF y 5x10 Small LOCA, ice condenser drains blocked 2
5x10 6 Interfacing-systems LOCA 2
V
-5 3
shy 2x10 Small LOCA, ECCS fails in recirculation 2
-6 3
5)HF-6,y 3x10 See 5 HF6 2
-6 3
TML y 3x10 Total loss of feedwater
~4 Oconee 2
T MLUO' y 1x10 Total loss o' feedwater with HPI failures, 2
spray fails
-5 2
V 7.3x10 Interfacing systems LOCA
-5 2
T)MLU0' y 5x10 See T MLUO' y 2
-5 T)B MLU y 1.5x10 Station blackout with loss of AFWS 3
3 Calvert
~4 Total loss of feedwater, no containment ESFs Cliffs 2
TMLOO'-6 1x10
-5 2
TMLOO' y 1.1x10 See TMLOO'-6
-3 3
TML y 1.6x10 Total loss of feedwater
-4 3
TMQ D y 1.2x10 Stuck-open PORV, ECC injection fails
-5 3
TMLQ y 6.7x10 Total loss of feedwater, stuck open PORV
-5 3
TMQ-H y 6.2x10 Stuck-open PORV, ECCS fails in recirculation
-5 3
TKML y 5.4x10 ATWS with loss of all feedwater RKey to re. lease categories is listed on page 7-8.
QRKey to PWR event sequences is listed in page 7-9.
7-6
l Table 7.2 Risk Significant Sequences-Bh?s
\\
Release Plcnt Category
- Sequence ** Probability Description P :ch'
-6 B:ttom 2
TW y' 3x10 Long-term loss of decay heat removal, ECCS fails on containment failure
-5 3
TW y 1x10 See TW y'
-5 3
~4 Gulf 2
TPI-6 8x10 Stuck open relief valve, RHR failure
-5 2
TQW-6 1.5x10 Long term loss of heat removal
-5 3
- Ksy to release categories is listed on page 7-8.
- Kay to BWR event sequences is listed on page 7-10.
2
.t
)
e T
7-7
.I
a KEY TO PWR AND_ BWR_ _ RELEASE CATEGORIES
- DURATION WMBEING ZIEVATI(.at PRACTION OF CORE INY RE!2ASED PROBABILITY g
M E
ON CATE00RY Reactor-Yr (Hr)
(Hr)
(Hr)
(Noters) (10 stu/Mr) Ee-Kr Org. 1 I
Cs-88:
Te-Sb Sa-sr au La Pwk 1 9 10~
2.5 0.5 1.0 25 520(d) 0.9 6s10 0.7 0.4 0.4 0.05 0.4 3:10"
~
Pwn 2 8 10~
2.5 0.5 1.0 0
170 0.9 7s10' O.7 0.5 0.3 0.06 0.02 4:10~
PwR 3 4x10
5.0 1.5 2.0 0
6 0.8 6 10" 0.2 0.2 0.3 0.02 0.03 3:10 Pwn 4 Salo" 2.0 3.0 2.0 0
1 0.6 2x10 0.09 0.04 0.03 5 10 3x10 ' 4x10
~
~
~
~3 9:10 ' 5 10~
1 10" 6x10"* 7x10~
~
~
7a10
2.0 4.0 1.0 0
0.3 0.3 2x10 0.03 Pwn 5 PWR 6 6x10~
12.0 10.0 1.0 0
N/A O.3 2x10~ 8s10" 8x10 1 10~
9 10 ' 7 10 1:10 '
~
~
PWR 7 4x10~
10.0 10.0 1.0 0
N/A 6x10~ 2x10~ 2x10~
1x10~ 2x10 ' 1x10 1x10~ 2s10 '
~
~
~
~4 Pwn 8 4x10 '
O.5 0.5 N/A 0
N/A 2x10" 5x10~ 1x10~
5:10~ 1x10 1 10 0
0 FwR 9 4x10~
0.5 0.5 N/A 0
N/A 3x10' 7m10 in10 6x10~ 1a10~
1x10" I O
O
~
~3 Dwn 1 lul0' 2,1 2.0 1.5 25 130 1.0 7x10 ' O.40 0.40 0.70 0.05 0.5 5 10
~
~3
~I BwR 2 6x10~
30.0 3.0 2.0 0
30 1.0 710 0.90 0.50 0.30 0.10 0.03 4x10 BWR 3 2x10~
30.0 3.0 2.0 25 20 1.0 7x10~
0.10 0.10 0.30 0.01 0.02 3x10"
~4
~4
~4
~4
~4 BWR 4 2x10" 5.0 2.0 2.0 25 N/A 0.6 7x10 8x10 5x10 4x10" 6x10 6s10 1m10
~
Bwn 5 1x10 3.5 5.0 N/A 1O N/A 5x10"* 2x10 6x10~
4x10 ' 8 10~
Sa10"I4 0
0
~4
~
(1) A discussion of the isotopes used in the study is found in Appendix VI.
Background on the isotope groups and release mechanisme is found in Appendia VII.
(b) Includes Mo, Rh, Tc, Co.
(c) Includes Nd, Y, Co, Pr, Ls, Nb, Am, On, Pu, Np, Zr.
(')
A lower energy rele ase rate than this value applies to part of the period over which the radioactivity is being released.
The effect of lower energy release rates on consequences is found in Appendia VI.
- Table extracted from Reactor Safety Study (WASH-1400) page 78, Main Report.
l 7-8
KEY TO PWR ACCIDENT SEQUENCE SYMBOLS A - Intermediate to large IOCA.
B - Failure of electric power to ESFs.
B' - Failure to recover either onsite or offsite electric power within about 1 to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> following an initiatir g transient which is a loss of offsite AC power.
C
- Failure of the contain:nent spray injection system.
D
- Failure of the emergency core cooling injection system.
F
- Failure of the containment spray recirculation system.
G
- Failure cf the containment heat remov.al system.
H
- Failure of the emergency core cooling recirculation system.
K
- Failure of the reacto* protection system.
L
- Failure of the secondary system steam relief valves and the auxiliary feedwater system.
M
- Failure of the secondary system steam relief valves and the power conversion system.
Q
- Failure of the primary system safety relief valves to reclose af ter opening.
R - Massive rupture of the reactor vessel.
S
- A small IDCA with an equivalent diameter of about 2 to 6 inches.
S2 - A small IDCA with an equivalent diameter of about 1/2 to 2 inches.
? - Transient event.
V - LPIS check valve failure.
- Containment rupture due to a reactor vessel steam explosion.
a B
- Containment failure resulting from inadequate isolation of containment openings and penetrations.
Y - Containment failure due to hydrogen burning.
8
- Containment failure due to overpressure.
l C - Containment vessel melt-through.
0 - Failure of containment heat removal functir...
U Failure of decay heat removal by high pressure makeup and bleed.
l 7-9
~.
KEY TO BWR ACCIDENT SEQUENCE SYMBOLS
- Rupture of reactor coolant boundary with an equivalent diameter of greater than six inches.
A 8 - Failure of electric power to ESTs.
C - Failure of the reactor protection system.
D - Failure of vapor suppression.
E - Failure of emergency core cooling injection.
F - Failure of emergency core cooling functionability.
G - Failure of containment isolation to limit leakage to less than 100 volume per cent per day.
H - Failure of core spray recirculation system.
I - Failure of low pressure recirculation system.
J - Failure of high pressure service water system.
H - Failure of safety / relief valves to open.
P - Failure of safety / relief valves to reclose after opening.
Q - Failure of normal feedwater system to provide core make-up water.
S - Small pipe break with an equivalent diameter of,about 2"-6".
3 S - Small pipe break with an equivalent diameter of about 1/2"-2".
y T - Transient event.
U - Failure of HPCI or RCIC to provide core make-up water.
V - Failure of low pressure ECCS to provide core make-up water.
W - Failure to remove residual core heat.
o - Containment failure due to steam explosion in vessel.
S - Containment failure due to steam explosion in containment.-
Y - Containment failure due to overpressure - release through reactor building.
Y' - Containment failure due to overpressure - release direct to atmosphere.
6 - Containment isolation failure in drywell.
c.- Containment isolation failure in wetwell.
C - Containment leakaga greater than 2400 volume per cent per day.
il - Reactor building isolation failure.
-0
- standby gas treatment system failure.
r t
i I
l 7-10 j.
l a
w
.e
.,m myg-g
APPENDIX A This appendix presents a letter from Mr. F. C. Finlayson (Consultant to the LSRG) to Dr. D. F. Ross Jr., (Chairman, LSRG), dated January 29,1981.
~
The letter states Mr. Finlayson's recommendation that the LOFT program continue at least through the period currently recommended by the NRC/.
LOFTmanagement(FY1986).
l A-1
_ _ _ - _ _ _ _ = _ -
/
T 11 I? A I? 11 O S P A C E C O 11 P O R A T ' O N J,.
' u./
Tmt Office llax 92957. los Anacles California D0009, Telephone:(213) 648-5000 Jan.iary 29, 1981 i
Dr. Denwood F. Ross, Jr., Chairman LOFT Special Review Group Division of Systems Integration office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washing ton DC 20535 e
Dear Dr. Ross :
After contemplating the results of the last LOFT Special Review Group (LSRG) meeting, I have concluded that I cannot support the apparent consensus of the group to recommend termination of the WFT program at the end of FY 1983.
I recognize that programs that have outlived their usefulness can and should justifiably. be terminated.
However, I believe that the evidence strongly supports. the position that LOFT, 'on balance, has not yet fulfilled its mission and is proaucing valuable insights into problems of reactor safety with every test.
These insights into important phenomes.a that were previously unidentified in-over ten years of separate effects testing include in part such etfects as, three dimensional flow behavior in the downcomer, early clad rewetting occurring during the blowdown phase of a large LOCA, and the remarkable effectivenss of the pumps to cool' the core with two-phase steam of very high quality--as demonstrated in the recent L3-6 test.
These insights into new phenomena together with the influence of recent unanticipated incidents in large coerational power reactors have biased the WFT program towaro conduct.lig tests related to phenomena which were largely unrecognized more th.in two years ago.
The LCFT results have shown the need for improvements in elements of the large nuclear reactor safety analysis codes and. outstripped the current code development, assessment, and evaluation efforts.
Since the current direction of the - IDFT program is heavily influenced by the events of the past two years, it seems incongruous to believe that the LSRG has sufficient insight into the future to suggest that the program will have outlised its usefulness within the next two years.
I have cc,me to the conclusion that the IDFT f acility and a nucleus of its caare of personnel ao indeed represent a unique national resource.
A ' resource which has not only contributed much useful data with respect to reactor
- safety, but has also proviceo equally significant psychological benetits to the public.
IDFT has a unique An Equal Oppmmuity Employer GENERAL OFFICES LOC ATfD AT: 3390 Eas? EL stGuhoo souLtvano. EL stouMDO. CALIFORNI A A-2
capacity for examination of the resulu, of imposed reactor upsets, transients, and postulated accidents--as acknowledged by the LSRG.
The results of the IDFT experiments have oemonstrated to the public that reactors can tolerate the imposition of substantial insults ano abuse ano still recover without damage.
The psychological benefits of these demonstrations may be immeasurable but they are certainly substantial.
It seems to me that the principal issue facing the LSRG and the ACRS with regard to continuation of the LOFT program has been the t
cost-cifectiveness of its results.
It must be acknowleogea that the weaknesses displayed in presentations. of the LOFT budget by EG6G and Nhc program management have contributed to feelings of frustration when assessing the program's cost-effectiveness.
The management tactic of presenting the budget as a single,. unitary line item of about $50 i
million per year is largely indefensible.
However, with the limitec insight which the LSRG recieved into subdivision of the total LOFT budget, it appears that the program could be conducted successfully at a lower annual cost.
It would be presumptive of me to give a precise estimate of what a more reasonable annual budget might be.
However, I i
believe that the operational me.npower for the program could be red uced.
I do not believe it is necessary to staff the facility for full-time, 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, around-the-clock operation.
I also have the impression that EG&G management could apply a
matrix type of organization to the supporting engineering and numerical methods personnel which woulo permit greater tiexibility in application of their talents to other related INEL programs as well as LOFT.
Such a change might substantially reduce direct charges to LOFT in comparison with the current practice ot assigning the supporting technical personnel tull-time to the project.
J I also agree with the LSRG position that many of the peripheral LOFT programs associated with areas such as human factors and the man-machine interface could be conducted more effectively outsiac of LOFT.
Some nominal reduction of the LOFT budget could be achieved by performing these tasks under another aegis.
I find myself in funoamental oinagreement, however, with the LSRG position that tests in the areas of cegraded core cooling and severe transi'.ds (those included under the LOFT L8 and L9 test series) are too hatardous for consideration in the program.
These tests represent the primary area in which IDFT data might result in substantial reduction of the potential risks of nuclear power reactors.
Uhe
" edge-of-the-safety envelope" tests provide the principal opportunities (as well as challenges) for demonstrating the effectiveness of reactor safety systems.
LOFT provides a unique facility for investigation of edge-of-envelope phenomena--a system designed for testing at or near the limits of operational extreme conditions.
If the safety analysis cevelopment and psychological benefits ci the IDFT program are to be fully utilized, I do not believe this kind of testing can be bypassed.
. A-3 x -_
i
.It is recognizeo that a major barrier to conducting such tests in IDFT is the uncompleted tission product cleanup system for the facility.
LOFT personnel have not been eager to discuss this problem--probably because it represents another and somewhat embarrasing major moditional expense for the system.
Never-the-less,
,it will be necessary to complete the fission product cleanup system before serious consideration can 'be given to tests which might result in fuel damage and fission product releases.
Therefore, I recommend that the money needed to complete the fission product cleanup system be budgetted and allocated as soon as possible.
Following completion of the cleanup system, an orderly test program should be ~ conducted--designed about tests of increasing potential severity as the program progresses--until tests demonstrating the effects of ~ inadequate core cooling and the systems characteristics f
of fission product releases have been performed.
Sufficient numbers of such tests should be conducted to provide the background data needed to support further refinement of risk analysis codes.
The' test program should include an increased effort at communication with the utilities, the general technical community, and the public at large.
Improved communications are essential in order to ensure that the data developed with regard to the ability of reactors to take insults and Sbuse and retain their inherent safety is adequately disseminated, reviewed, and assimilated.
With regard to the cost-effectiveness of this recommendation for an extended IDFT program, several aspects should be noted.
First, the U.S. annually produces over 2 trillion kwh of electricity, with a value of the order of $100 billion per year.
Moreover, the ' value of the annual electrical energy production of a single nuclear power plant is of the order of $250 million.
- Thus, if the LOFT results make it possible to licence one more nuclear power plant, or it they reduce the licensing time of a single plant by one or more years (as a result of increased public confidence), or if they reduce the probability of l
another accident like TMI by a notable amount, the. f ull costs of the IDFT program (f rom its inception) could easily be covered.
Similar arguments apply with respect to the magnitude of annual capital investments associated with nuclear power plants to be constructed in the near future.
It is probable that the tangible and intangible benefits of the IDFT program have already been great enough to cover past costs.
If the program recommended above is followed, future i
benefits should be even greater.
l If LOFT is believed to be out of balance with other aspects of the-nuclear reactor. safety research budget, perhaps a new source of revenue I
should be sought to finance it.
A general surtax of.02 mil /kWh on the U.S. electrical energy proouction would cover 'he entire current LOFT annual budget.
The incremental costs to consumers for conducting the l
l A-4
WFY program are insignilicant, particuP-ly when comparea' to the atrect benefits obtaineo from the program o3
.:h member of the public.
In summary, I believe that WFT oces indeed represent a unique national resource [or investigating the eticcts of potentially severe reactor transients una accidents.
As such, I believe it is premature to specify a termination uate tot the program as early as FY 1983.
Since trosh insight into now ano unexpecteo phenomena are being obtainco with each LOFT test, its mission does not appear to be uutticiently satistled to justify suele an early termination date.
Thus 1 recommend continuation of the program at least through the period currently recommended by the N RC/BES LOFT management (FY 1986).
The emphanin of the program should be placed upon development of the facility's capability to test "eoge-of-the-envelope" saf ety conoitions associated with inadequate core cooling.
An orderly series of tests leading to such investigations shoulo be conouctea with results aimeo at coveloping data for code refinements to support risk analysis calculations so that the phenomena associated with incipient core damage and fission product releases are adequatly determined.
Sincerely yours,
/
- n &!.4 /
Alcl9n Fica C. F'nlaysdn, Manager
'Juclear and Geothermal Systems FCF:jmf
-A A-5
APPENDIX B As part of its review, the LSRG solicited comments from the nuclear steam supply system vendors regarding the continued operation of the LOFT facility. This appendix contains a copy of the request and copies of the responles received.
REQUEST Letter from D. F. Ross (Chairman, LSRG) to T. Anderson (Westinghouse) dated December 8, 1980. (Identical letters sent to J. Taylor (B&W), E. 0:herer (CE) and G. Sherwood (GE).) NOTE: Options A, B, f. C discussed in these doc-uments are different that the options presented in Chapter 1 of the LSRG Report............................................... B-2 RESPONSES Letter from T. M. Anderson (Westinghouse) dated January 16, 1981... B-14 Letter from J. H. Taylor (B&W) dated January 20, 1981..............
B-21 Letter f rom A. E. Scherer (CE) da ted January 9,1981............... B-23 Letter from G. G. Sherwood (GE) dated January 15, 1981.............
B-25 B-1
su ask UNITED STATES o
l 8'
NUCLEAR REGULATORY COMMISSION f
T,;
.E W ASHINGTON. D. C. 20555
%,*...*/
DEC 8 1980 Mr. Thomas Anderson, Manager Nuclear Safety Department Westinghouse Electric Corporation P. O. Box 355 Pittsburgh, Pa.15230
Dear Mr. Anderson:
REQUEST FOR COMMENTS REGARDING THE LOFT SPECIAL REVIEW GRO
SUBJECT:
The purpose of this letter is to solicit your coments regarding the continued In particular, we would appmcf ate your thoughts cperation of the LOFT facility.
concerning operation through FY 1985 as contrasted with the recomendations of the ACRS. The ACRS has recomended to the Comission (letter from M. S. Plesset to J. F. Ahearne, dated July 17,1980, "Ccanents on the NRC Safety Research Program Budget for Fiscal Year 1982." NUREG-0699) that following the FY 1982 tests scheduled to be conducted at LOFT, the facility should be decomissioned (or perhaps taken cver by industry).
(A copy of this recomendation, as extracted from NUREG-0699, is included as Enclosure 1 to this letter.)
As a result of this recommendation, the Comission established, on October 28, 1980, a LOFT Special Review Group (LSRG) for the purpose of reviewing the LOFT The report is intended to program and reporting its findings to the Comission.
atd the Comission in their decision whether to continue NRC support of the LOFT I have been appointed as the Chaiman of the LSRG.
project beyond FY 1982.
Enclosum 2 shows a listing of the membership of the group and Enclosure 3 is a copy of our charter.
As noted in our charter, our efforts will include addressing specific regulatory needs and how the results of the LOFT program are expected to meet those needs.
In addition, we will report, based upon the perfomance of the program to date, on the likelihood that the planned program will provide the expected infomation and that the program can maintain reasonable flexibility to address changing regulatory issues.
Our group is divided into four subcomittees, each of which will evaluate the technical adequacy and usefulness of LOFT as it relates to the following areas:
1.
Operational Aspects (Man-Machine Interface);
2.
Code Verification and Scaling; 3.
Transients, ATWS, Risk Reduction; and, 4.
Degraded Cooling.
B-2
Mr. Thomas Anderson DEC 8 1980 2-Based upon our review in these areas, we will evaluate the following three options i
for the LOFT pro' gram: Option A-run through FY 1982; Option B-run through FY 1953; l
and Option C-run through FY 1985. provides a sumary of the planned l
LOFT test sequence for the th ne options.
The review group would very much appreciate your views and comments regarding the matters under consideration by the LSRG. Specifically, any comments with respect to the usefulness of the planned tests for providing needed safety-related msearch in the areas being addressed by the four subcommittees, would be most helpful.
Due to the short amount of time allotted to the LSRG to perform this review. I would like to have any comments you may have by January 1,1981. This will assure sufficient time for the LSRG subcommittees to integrate your coments into the review process.
If you would like additional information associated with the LOFT program and/or our review group effort, please feel free to contact me at (301.)492-7373.
Sincerely.
mmh YMG Denwood. Ross, Jr., Director Division of Systems Integration Office of Nuclear Reactor Regulation
Enclosures:
As stated cc: LSRG Members & Consultants IDENTICAL LETTERS SENT TO THE FOLLOWING:
James Taylor, B&W Edward Scherer, CE Glen Sherwood, GE B-3
ENCLOSURE 1 n Ma,q'o*%
a c
UNITED STATES NUCLEAR REGULATORY COMMISSION 8
- E ADVISORY COMMITTEE ON REACTOR SAFEGUARDS WASHINGTON, D. C. 20555 July 17, 1960 o
i l
Honorable John F. Ahearne Chairman U.S. Nuclear Regulatory Commission Washington, DC 20555
Dear Dr. Ahearne:
The Advisory Committee on Reactor Safeguards submits herewith its comments on the budget for FY 1982 of the Office of Nuclear Regulatory Research.
Only that portion of the budget relating to Program Support has been con-sidered.
The funding levels considered are those allocated by the E00 Staff in its preliminary markup of 2 July 1980 and those requested by RES in its reclama of 9 July 1980.
Coninents on personnel requirements and allocations are included in a few instances where particularly appropriate.
Sincerely, Milton S. Plesset Chairman
Attachment:
NUREG-0699 B-4
ENCLOSURE 1 (continued) 2.
LOFT 2.1 Introduction The LOFT facility is the only integral facility which models a PWR.
The shortcomings of the facility are well known and relate for the most part to deficiencies in vertical dimensions.
The nuclear core is slightly less than half the height of a PWR core.
This reduced height introduces some uncertainty in translating the early quench observed in the large LOCA test in LOFT to a full-size system.
Further, the height relation-ship between the core and the steam generators affects the interpretation of measurements of natural circulation heat transfer.
2.2 The LOFT Test Program LOFT tests were for some time directed toward a design basis accident involving the instantaneous double-ended cold leg break (DECLB).
Tests of this type have contributed to the understanding of this kind of acci-dent and also have contributed to code assessment.
In response to a strongly modified view of more immediate needs, the LOFT program was redirected in FY 1980 to the study of reactor transients which were the result of small breaks.
The current plans call for further tests of this kind in FY 1981.
Both the FY 1980 and the FY 1981 programs as now planned include other types of t ransients, including, particularly in FY 1981 tests concerned with anticipated transients without scram.
The signifi-cant test proposed for FY 1982 is a DECLB at the higher core power of 16 kw/ft.
No further small break tests are scheduled for FY 1982.
A test has been proposed for FY 1983 with pressurized fuel.
Although we believe that LOFT will essentially complete its NRC mission in FY 1982 with NRC funding phased out at the end of FY 1982, the LOFT System could still be a valuable tool for the nuclear power indu st ry.
The LOFT installation could be of fered to the nuclear industry to be operated with industry financial support as a facility which would en-hance operational capabilities of the nuclear industry.
2.3 Recommendations LOFT represents the largest single expenditure in the safety research budget so that its program must be considered with special care.
We recommend that the tests through FY 1982 be adequately funded and that following the 1982 tests the f acility be decommissioned unless it is taken over by the nuclear industry.
The final tests to be run to the completion of the program should be carefully scrutinized and evaluated by RES to obtain the most useful final series.
We would also wish to contribute to the choice of these tests.
Ef ficient operation of the facility appears to require the requested level of support and therefore we endorse that level.
19 B-5
m---
l ENCLOSURE 2 MEMBERS AND CONSULTANTS
{
OF THE LOFT SPECIAL REVIEW GROUP AFFILIATION MEMBERS D. Ross (Chairman)
Office of Nuclear Reactor Regulation, NRC Office of Nuclear Reactor Regulation, NRC B. Sheron Office of Nuclear Regulatory Research, NRC R. Bernero R. Woodruff Office of Inspection & Enforcement, NRC Office of Nuclear Reactor Regulation, NRC R. Capra L. Jaffee National Aeronautics & Space Administration A. Pressesky Department of Energy CONSULTANTS H. Kouts Brookhaven National Laboratory A. William Snyder Sandia Laboratories H. Isbin University of Minnesota R. Pack Institute of Nuclear Power Operations B. Sun Electric Power Research Institute F. Finlayson Aerospace Corporation P. Griffith Massachusetts Institute of Technology I. Catton University of California at Lot. Angeles /
ACRS Consultant 1
B-6
I ENCLOSURE 3 CHARTER OF THE LOFT SPECIAL REVIEW GROUP "This group is established for the purpose of reviewing the LOFT program and reporting on their findings to the NRC Commissioners.
The review shall be technical in nature, focusing on, but not limited to, the I
benefits expected from the program planned for the FY 1981 to FY 1983 period.
The primary purpose of the group is to consider whether LOFT should be deccamis-sioned in FY 1983, as recommended by the ACRS.
The group would be expected to consider the LOFT program from the perspective of NRC's overall research program and in terms of the needs of reactor regulation.
To facilitate this work, NRC and the INEL, where LOFT is located, would provide presentations, reports, and tours and interviews.
Also, the group would be welcome to attend any tests performed in the LOFT reactor or related facilities.
The report would be intended to aid the Commissioners in their decision whether to continue NRC support of the LOFT project beyond FY 1982.
The report should address specific regulatory needs and describe how the results of the LOFT pro-gram are expected to meet those needs.
Furthermore, based on the performance and responsiveness of the program to date, the report should indicate the likeli-hood that the planned program will provide the expected information and that it maintains reasonable flexibility to address changing regulatory issues.
A final report would be issued by February 3, 1981 and after follow-up discus-sions with the Commissioners, the group would be dissolved."
8-7
ENCLOSURE 4 PLANNED LOFT TEST SEQUENCE FOR DIFFERENT OPTIONS OPTION A: Run through FY-1982/Decomission FY-1983 OPTION B: Run through FY-1983/Decomission FY-1984 OPTION C: Run through FY-1985/Decomission FY-1986 The attached sheets show the planned tests and sequence should either Option A, B or C be selected by the Comission.
3
'I d
B-8
OPTI21APf0GPM l
IEC S LM, S?ALL BEAK WITH PLPPS Gi /LS-L COE tricom, NO IW%E JK481 EB mR01 APRIL 23, St'ALL BEAK WIE L6S OF S.G. 2 arf /L9-L LOSS OF EEIMAER WITH t%Y C31TAINEIT NESSEL LEAK JIIE JULY 17, ARK #!SAS NUCEAR-1 STARTlP ACCIIENT /3L COLIMAER ACCIIENT AUG 15-1, ItTEPEDIAE BEAK ACClFULATOR LI!E E COE !?!COM S3'T COE CLTIE, ENTPAL BltDE AT 600 oSI OCT
-NOV IEC 81 12-6, lARI BEAK WITH LOSS-OF-0FP3ITE PQER, EXPECT CLAD BURST #ID J#1E2 EB ENER BLIDLE GiA'iE, PEPESSURIZED TO 353 PSI
".AR AoRIL MY
@-3, TPANSISIT WITH0 LIT SCPM J!?E JULY AUG M, COE lf!C0\\ER( WITH SEW.E FLEL IW%E SEPT E2 BEGIN CENitP, POST IRPADIATION E7MINATI0'l #D FINAL #W_YSES, IECDmISIS21 A'S DISPGE OF REL DURING 1983.
B-9
OPTION B PfDGRAM SEPT S1 C0fE 01ANE, EffTRAL Bi1EE AT 350 PSI OCT NOV EC S1 12-5, LARE BEAK WITH LOSS-OF-OFFSITE P3ER, NO DAl100NING EXPECTED J#182 FEB 02 ITER BlICE CHAiE, PEPESSURI2ED TO 500 PSI
?%R APR E LARE BE#'i!ITd LOSS-0F-0FFSITE POER, EXPECT CLAD BURST
- 6 CEANlP V#{
JifE JULY CENTER Bif0E Gi#lE AUG SEPT L5-2, INTEPfEDIATE M_AK (PESSURIZER SURELIfB OCT LE-4, CONTROL ROD WITIERR!AL NOV EC S? g-3 TRN1 SIB 4T WITHOUT SCRM JR183 RB L9-4 TRN!SISff WIlliOUT SCRAM f%R
- )RIL CENTER BlIEE CHANE, PEPESSURIZED "3 350 PSI "sf TDDIF/ STEAM ENPATOR R)R TLE RL9JE TEST JLIE L7-1 STEAM ENERATOR TLEE RUPTLPE TEST l
JULY AUG LQ-9, COE [fiC0VER/ Willi SEVEE REL WAE SEPT 83 BEGIN CLERRP, POST-IRPADIATION EXN'JNATION AND FINAL #W.YSES, l
EC0fMISSIUi #,'D DIS OSE OF FIL DURING P/1984, l
B-10 l
PLANNED LOFT TEST SEQUENCE
-0PTION C **
AND TARGET DATES
- for each year, committment dates are roughly 2 months later INITIAL INITIAL TEST TARGET POWER CORE ID DATE LEVEL (MW)
AT F COMMENTS L3-6 12-1-80 50 35 Small break (2.5%) intact loop cold leg --- pumps on.
Pumps tripped at end of experiment to measure water remaining.
L8-1 12-1-80 Add on to L3-6 Core uncovery without ECC at low decay heat level.
L9-1 3-4-81 50 35 Loss of all feedwater (multiple failures) with scram on high pressure; PPS setpoints representative of LPWR (PORVchallenged.) Mild ATWS.
=
i' L3-3 4-8-81 53 *"' 50 35 Small cold leg break (0.16%) HPIS flow approximately.
r-equal to break flow.
Dry steam generator secondary.
'C
@i Determine the boundary between break heat removal and PORV heat reraoval.
Needs further justification.
CV Leak 6/81 Required test of containment leak integrity.
Test L6-7 7/81 50 65 LOFT typicality to Arkansas Nuclear One startup test.
L9-2 7/81 Add on to L6-7 Rapid cold water accident, upper pienum voiding.
LS-1 8/81 50 65 Intermediate size break (accumulator line).
Determine if large break and small break models continue to predict intermediate break results. Also check out liquid level device.
L8-2 8/81 Add on to L5-1 Core uncovery at high decay heat level.
Reflood with degraded ECC capability.
May be the same as L5-1.
- -NOTE: Option C ' includes all items listed, including those under Option A and B 1
PLAN'8ED LOFT TEST SEQUENCE AND TARGET DATES AS OF SEPTEMBER 1980 (ccntinued)
INITIAL INITIAL TEST TARGET POWER CORE ID DATE LEVEL (MW)
AT F CO MENTS Whole core 10/81 F1 center bundle at 350 psi (80L).
Large peaking Changeout factor if only CB changed.
L2-5 1/82 B
16 kw/ft 65 Worst prototypic hydraulic conditions in core.
Investigate fuel behavior at BOL fuel pressure (no fuel damage expected).
Replaces CB 3/82
.F2 will be pressurized to 700 psi.
F1 with F2 L2-6 5/82 A,B 16 kw/ft 65 Same as L2-5 with 700 psi fuel pressure (E0' L Fuel damage and fission product release expected.
Replaces ;2 7/82 Only minimal fuel damage experiments can be done until with unpress F1 is examined for damages.
Al 5
LS-2 9/82 3
16 kw/ft 65 Intermediate size break on hot leg.
Pressurizer surge line.
Needs further justification based on LS-1.
L6-4 9/82 B
16 kw/ f t 65 Uncontrolled rod withdrawal at power.
Investigate worst case moderate frequency accident.
L9-3 12/82 A,B 16 kw/ft 65 ATWS. Loss-of-Feedwater is initiating event.
(Multiple failures.)
L9-4 3/83 B
16 kw/ft 65 ATWS.
Loss of offsite power is initiating event.
(Multiple failures.)
Put Fl Bundle F1 inspection completed and fuel is assumed not damaged.
back in L8-3 8/83 16 kw/ft 65 Small break with slow core heat up (l*F/ min).
Uniform clad swelling and blockage of flow channel.
Investigate potential initiating events.
(Candidate: Loss-of-Feedwater.)
Replace F1 with A3
PLANNED LOFT TEST SEQUENCE AND TARGET DATES AS OF SEPTEMBER 1980.(Centinued)
INITIAL INITIAL TEST TARGET POWER CORE ID DATE
-LEVEL (MW) uT'*F COPNENTS L7-1 12/83
- B 16 kw/ft 65 Large break with S.G. tube ruptures at start ofi reflood/ refill (>25 tubes ruptures).
Provides upper bound of envelope on effect of ruptures). Critical number of tube ruptures resulting in extreme core
. temperatures expected to be between 10 and 25 based' on Semiscale results.
L7-2 2/84 16 kw/ft 65 Large. break with S.G. tube rt:ptures at start of reflood/
refill (<10 tubes ruptu: ed).
Provides a lower bound of envelope on effect of ruptures.
L7-3 should be inserted if possible which has critical number of ruptures.
L4-1 5/84 16 kw/ft 65 200% cold leg break. Accumulator injection into U.P.
Investigate topdowre core quench. Applicability to I
UHI plants.
L4-2 8/84 16 kw/ft 65 200% cold leg break.
U.P. LPIS injection.
Investigate }[ two loop plant phenomena.
w Replace A3 12/d4 with press F3 1
L8-4 3/85 A,B 16 kw/ft 65 Severe core damage.
Investigate potential initiating events.
(Candidate: Loss of offsite power.)
Whole core 4/85 F4 Center bundle.
changeout L10-1 7/85 16 kw/ft 65 Override test. Override of L8-3. transient.
L10-2 9/85 16 kw/ft 65 Override test.
0verride of L8-4 transient.
L8-5 11/85 16 kw/ft 65 Severe core damage.
Investigate potenthi initiatingevents(Candidate: Steam line rupture).
Decommission 12/86
l Nuclear rectvdogyomson Westinghouse Water Reactor Electric Corporation Divisions Box 335 PlmDurghPemsylvania 1523)
January 16,1981 Denwood F. Ross, Jr., Director NS-TMA-2369 Division of SystemsIntegration Office of Nuclear Reactor Regulation U.S. of Nuclear Regulatory Commission Washington, D.C. 20555
Dear Dr. Ross:
Subject:
COMMENTS REGARDING THE LOFT SPECIAL REVIEW GROUP'S STUDIES References L T. M. Anderson (Westinghouse) to G. D. MePherson (NRC) letter NS-TMA-2231 dated 5/6/80.
2.
S. Kellmen (Westinghouse) to L. Leach (EG&G) letter SE-SRA-580 dated 12/6/79.
3.
V. J. Esposito (Westinghouse) to L. S. Tong (NRC),
letter VJE-NS-736 dated 5/21/79 The purpose of this letter is to provide the LOFT Special Review Group with Westinghome's comments pertaining to future LOFT Test plans. The general framework of the potential future direction for LOFT was obtained through your letter of December 8,1980, which outlined three options extending operation as long as FY 1985 and requested comments.
Westinghouse has provided extensive input in the past expressing our needs, concerns and desires for information from LOFT, as well as specific test definition and instrumentation recommendations to improve the quality and usefulness of test data. This information has been communicated through Westinghouse attendance at LOFT Review Group meetings and by numerous letters sent to EG&G anel the NRC Division of Reactor Safety Research. A partiallist of recent letters is specified by the References.
In assessing future LOFT test plans, as proposed in your December 8,1980 letter, we first asked ourselves whether any cf these tests are needed to provide data which can be used to further assure ttat safety of Westinghome designed PWRs. Our conclusion is that none of the proposed tests are needed from a fu idamental safety viewpoint. We next considered whether any of the proposed tests could peovide additionalinformation useful for code essessment. The following table summarizes those tests that, in our judgement, could yield potentially useful additional information. Some summarizing commerts on each test scenario are provided in the encimed attachment.
L3-6 Small Break with RCPs Running LS-1 Intermediate Break In Cold Leg L9-1 Loss of All Feedwater Transient IA-1 Large LOCA with Upper Head Injection IA-2 Large LOCA with Upper Plenum Injection B-14
Westinghouse believes that the remaining tests presented in the table attached to your December 8,1980 letter can be eliminated from future consideration fw the s
reasons stated in the attachment. We believe that the Westinghouse recommended test matrix could be completed by the end of FY 1982. This would be similar in time to Option A of your letter.
We also would like to take this opportunity to again express our belief, as we have in the past, that LOFT tests are not representative of expected PWR behaviw nw are the transient thermal hydraulic phenomena necessarily applicable to PWR transients due to scaling and other differences.
Because LOFT is not a demonstration test of a commercial PWR, the test results are only useful as a tool ta provide data to ames and verify models contained in computer codes.
We thank you fw giving us this opportunity to express our views on LOFT. If we can be of furthe assistance w you request clarification on any of our comments, please do not hestitate to call Dr. V. J. Esposito of my staff.
Sincerely yours, T. M. Anderson, Manager Nuclear Safety Department TMA/wpc a
B-15
_J_@2L4L___________ ___ __ _
ATTACHMENT WESTINGHOUSE RECOP9ENDATIONS REGARDING THE LOFT SPECIAL' REVIEW GROUP'S OPTIONS FOR FUTURE LOFT TESTS This attachment-provides Westinghouse coments on' the potential options for future LOFT tests, as specified in the December 8, 1980-letter from
- 0.. F. Ross -(IRC) to_ T. M. Anderson (W). All tests are discussed and
~
organized-in this report by their test series group.
.\\
u s,
'l L3 Test Series i
The first test series discussed is the L3 small break series. Test L3-6 is a small break test with RCPs operating, and is the companion to j-L3-5.
Both of these tests have been performed at the time of this writ-ing. Westinghouse has agreed with the usefulness of these tests and has previously provided comments regarding them (Reference 1). These two 5
j tests will provide information to verify computer models used to assess the effect of RCP operation on the total vessel inventory during small break transients. L3-6 represents the expected worst case in_ terms of vessel inventory. Westinghouse analyses have predicted the worst case, in terms of PCT, to occur when the RCPs are tripped at some' intermediate l
time. Two phase separation phenomenon imediately af ter RCP trip for i
this case and core recovery by accumulators are important determinants i-of the extent of clad heatup. An additional test with RCPs tripped at an intemediate time may also be shown to be desirable. However, this cannot be determined until full evaluation of L3-6 is completed. We are not recomending an additional test at this time. The L8-1 add-on test to L3-6 could provide steam cooling data at low decay heat. However, out of pile tests. planned or already perfomed elsewhere are better designed to provide this type of data because of the LOFT neernal thermocouple attachment method. Therefore, the L8-1 test is not a wr.rthwhile test to perform, in our opinion.
I I
7932A B-16
The proposed L3-3 small break test is defined to determine the boundary between capabilities of heat removal through the break and secondary side. The critical break size for this situation is highly dependent on secondary heat removal characteristics and inventory, 'as well as the scenario which scrams the reactor and interrupts feedwater. Therefore, the characteristics are highly plant and scenario dependent. We believe that the L3-3 test will not provide meaningful and useful data to increase the understanding of small break behavior, and identification of the energy removal boundary for LOFT is not of great significance.
Much of the transient phenomenon is overlapped by other tests in the series. Therefore, we do not recommend further consideration of test L3-3.
L5 Test Series The intermediate break test series L5 represents a region in break size between the small break L3 Series and the large break L2 Series. These tests provide unique data for LOFT to identify the transition during the transient from predominantely inertial behavior, typical of large break, to predominantely hydrostatic dominated, typical of a small break. This test will demonstrate a smooth transition of variables such as cladding temperature from large to small break. We believe that the cold leg intermediate break test L5-1 is the most interesting in the L5 series.
We do not believe that L5-2, the hot leg interriediate break, will pro-vide a significant amount of additional new information given that L5-1 is run. Either intermediate break location could provide informa^ica regarding the flow transition throughout the phases of the transient.
The cold leg break location will be more severe, with greater primary mass depletion and core uncovery, and LOFT data at the two extreme break sizes at this location is available. Therefore, we do not think it is necessary to run both tests, and reconrnend that LS-2 be excluded.
L2 Test Series The L2 large break test series has two tests included in the options presented. Test L2-5 appears to provide base line data for the L2 6 7932A P-17
test where fuel damage is expected. Difficulties may arise for this test in terms of instrisnentation and measurement of necessary variables as well as interpretation of the results. Other test facilities are more appropriate than LOFT to evaluate fuel behavior during worst hydraulic assumptions. It is our opinion that fuel damage tests are not However, if fuel dnage tests must be considered, it appears necessary.
to be more prudent to consider separate effects tests to look at pheno-menological behavior. Therefere, it is our opinion that these LOFT tests should eliminated.
L6 and L9 Test Series j
The L6 and L9 test series represent, in general, the operational transients and proposed ATWS scenarios. The L6 series tests for LOFT represent mild transient frun a system view point that are of very limited value for code assessment. The reactivity feedback in LOFT due to the high enrichment will result in significantly different responses for transients such as rod withdrawal (L6-4) and cooldown transients (L6-7) than found in a connercial PWR. Furthermore, these types of transients in the L6 Series are typically influenced by the action of the various control systems. The control systems in LOFT and a conner-cial PR are different enough to yield different transient resporise.
Our connents pertaining to these tests have been passed along via Refer-ence 2, as well as through informal discussions with EG&G personnel.
The L9 test series represents the classical ATWS transients. Westing-house has concerns regarding the capability of the LOFT facility to model an ATWS event. The ATWS tests proposed for the L9 Series are for those initiating events that result in a pressure increase. From the numerous studies that Westinghouse has done as well as studies done by l
your_ own staff, the resulting overpressure is a very strong function of l
at least three parameters. The first parameter is the ratio of the l
moderator temperature coefficient to the fuel doppler temperature coef-l ficient. The LOFT core necessarily has no doppler coefficient which would result in significantly different phenomena with regard to pos-sible code assessment being modelled in the performed test. The second 7932A B-18
{
parameter is the heat transfer in the steam generator during tube uncovery.
If this paraneter is not modelled correctly the time response will be significantly different than in a commercial PWR.
Finally, the amount of overpressure that will result is dependent on the throat area of the safety valves. These parameters are not modelled in LOFT as they would be in a cannercial PWR. These represent areas that are signifi-cantly different than in the previous LOCA review.
We would recommend that no tests in the L6 and L9 Series be performed because of the major differences found in LOFT and a cannercial PWR with regard to the transients and ATWS modelling.
One exceptisi to the above is test L9-1, which represents a loss of all feedwater. We believe that this test could be useful, not necessarily as an ATWS simulation, but as a multiple failure scenario. The most important phase of the transient occurs as the steam generator drys out, RCS repressurization occurs, and PORV discharge results in primary mass depletion. Westinghouse has analyzed this accident to assess the effectiveness of recovery techniques such as initiation of auxiliary feedwater and holding open PORVs. This type of experiment on LOFT including recovery scenarios would be useful for code assessment.
L8 Test Series The L8 series tests represent the severe core danage scenarios. We believe that this test series should not be performed.
The proposed tests are complicated integral system scenario tests that will yield a complex and potentially meaningless set of results.
It is our opinion that separate effects tests are more useful to understand the severe core damage phenanena and provide infonnation to further verify the conservatisms in existing models. Without these more basic tests, infonned decisions on integral systems test definition and instrumenta-tion requirements cannot be made. Tests performed to date on simpler facilities with less variables have previously looked at severe fuel damage. Two such tests are P8F and Kralsrule melt tests. These tests were better defined core melt tests, and will provic% better data than 7932A B-19
the LOFT facility. We also believe that L0f T core damage tests are not justified in light of the planned study of the TMI core. This program is already in place, and therefore, additional spending to obtain similar data from the LOFT facility is not warranted.
It is our opinion that the L8 and L10 tests should be eliminated.
L7 Test Series, The L7 series tests represents the large LOCA plus stean generator tube rupture, with an objective to produce the case with the critical neber of tubes ruptured. Westinghouse believes that these tests provide mini-mal new useful data and question the cost / benefit relationship for the following reason. The proposed scenario is an extremely low probability event. The large LOCA itself is of low probability and the steau gener-ator tubes are designed to withstand the LOCA loads. The combination 1
event postulating the critical tube f ailure is highly unlikely. These tests will not provide significant additional information regarding the system behavior characteristics or the effects of the multiple f ailure, since a neber of similar tests have been run on the Smiscale Facility. Due to the design of the LOFT Facility, the test will not provide any information on recovery techniques. Given the extensive cost of these tests, the low probability of the event, and the limited data expected, we recommend that this test series be eliminated.
L4 Test Series The final test series proposed are the L4 tests which include upper head injection and upper plenun injection large break tests. These systems, in comercial PWs, represent major improvements to ECCS design and PW s af ety. These tests coult' + ovide additional infonnation fcr code assessment, such as the quan, 'ication of the large degree of conserva-tism imposed by present regulatory requirenents. However we remain cautious of scaling and other LOFT atypicalities such as upper head and upper plenun volune as compared to a PW that may result in behaviors that would not be expected in a PW. From this standpoint, these tests similarly to all proposed tests, are not PW demonstration tests. For example, scaling and other differences in the Seniscale Facility for similar tests resulted ir. behavior we believe is not typical for a PWR.
7932A B-20
Babcock &-Wilcox-noci..t Power Generation Division a McDermott company 3315 Old Forest Road P.O. Box 1260 Lynchburg, Virginia 24505 (804) 3M 5111 January 20, 1981 Dr. D. F. Ross U. S. Nuclear Regulatory Conmission Washington, D.C.
20555
Dear Dr. Ross:
In response to your request (letter, D. F. Ross to J. H. Taylor dated December 8,1980 - subject: Request for Comments Regarding the LOFT Special Review Group's Studies), the Babcock & Wilcox Company submits
{
the following comments regarding the continued operation of the LOFT l
facility.
It is our belief that the LOFT facility has served a useful purpose in establishing a sufficient data base to assess ECCS codes. B&W's involve-ment in the NRC's Standard Problem and Required Licensing Problem Programs has given us added confidence in our ECCS code's abilitv to sufficiently predict conservative results.
This confidence stems from the pretest predictions and the post-test evaluations performed to understand differences between the test results and the prediction due to test anomalies.
Combining this with the fact that vendors are required to perform ECCS design calculations based upon evaluation models and ECCS rules which contain conservatisms, we believe the future benefits to be gained from the LOFT facility are limited.
In direct response to the four areas of review on which the special review group is concentrating, we offer the following comments:
1.
Operational Aspects (Man-Machine Interface) -
It is our belief that operational aspects can best be addressed by studying actual environ-ment of the operating plant. As a second choice for studying MMI, plant simulators can also ue used. The use of the LOFT facility would not only be atypical to an actual PWR but would also be an excessively costly means of investigating man-machine interfaces.
2.
Code Verification and Scaling - Because of the atypicalities of the LOFT facility as compared to a PWR (smaller core height, larger surface to volume ratio, one active loop, the height relationship between the core and the steam generator, etc.), the results obtained have always yielded a degree of uncertainty from both an industry and public perspective.
In addition, a number of uncertainties related either to the analytical code predictions or to the test facility B-21
Babcock &Wilcox - itself have hindered the use of LOFT as a viable computer code verification facility. - We believe that the establishment of various separate effects test programs directed towards investigating the uncertainties identified.during the LOFT program would be more cost effective, eliminate LOFT system interaction effects and provide qualitative information which could potentially be used by the NRC for code verification purposes.
3.
Transients, ATWS and Risk Reduction - We believe that the atypicalities between the LOFT facility and actual PWR's will be even more predominant for these tests.
This is particularly true for those non-LOCA transients which are affected by asymetric steam generator responses.
Since the LOFT facility has only one loop, the usefulness of these tests is also limited.
In addition, the tests scheduled to assess either ATWS or risk reduction have not been sufficiently defined to warrant additional coments at this time.
4.
Degraded Cooling - We believe there is reason to question the validity of tests in this area which may be run on the LOFT facility due to the l
extreme complexity of the event especially when the atypicalities of l
the facility are considered.
In view of the questionable test benefits, it would seem inappropriate to conduct this very costly series of tests which involve severe core damage and will require expensive subsequent decontamination operations.
In sumary, it is B&W's conviction that the monies and other resources being invested in the LOFT program would have a greater return in terms of added safety if they were applied more directly to understanding and minimizing actual plant operating problems rather than being applied to a facility which is as atypical as LOFT. Therefore, based upon the options presented in your letter, B&W would recomend the option "A" program. We would also recomend that the last three tests L2-6, L9-3, and L8-4 be carefully evaluated from a cost-benefit standpoint.
We appreciate the opportunity to provide you these coments.
If you have any questions, please contact me or George Geissler (Ext. 2536) of my staff.
Ver truly yours, James H. Taylor Manager, Licensing JHT/dsv cc: R.B. Borsum D.H. Roy B-22 e
C-E Power Systems TEL 203/688-1911 Combustion Engineenng. inc.
Tr.in 99297 1000 Prospect Hill Road Windsor. Connecticut 06095 H SYSTEMS POWER l
January 9,1981 LD-81-003 Dr. Denwood F. Ross, Jr.
Director Division of Systems Integration c
Office of Nuclear Reactor Regulation U. S. Nuclear Regulatory Comission Washington, D. C.
20555
Subject:
Request for Coments Regarding the LOFT Special Review Group's Studies 1
Dear Dr. Ross:
In response to your letter of December 8,1980 Combustion Engineering offers the following coments regarding studies being conducted by the LOFT Special Review Group.
J As you are aware, Combustion Engineering has consistently been of the opinion that separate effects testing is the most direct and cost effective method for conduct-ing research into phenomena associated with loss of coolant accidents.
Nonetheless, Combustion Engineering believes that integral te:; ting done to date at the LOFT facility has been of value in confirming separate effects results and, perhaps as importantly, has been perceived by a significant segment of the public to be j
more definitive and convincing.
With regard to the range of tests which might be conducted in the future at the LOFT facility, we believe that some of the proposed tests may produce results of value, while others are of less technical merit and could be dispensed with.
In any event, it is our opinion that performance of the most significant tests should not be unduly delayed.
Of the three options identified in the charter of the LOFT Special Review Group, we tend to support option A.
While we might have chosen a somewhat different set of tests, option A appears to do a reasonable job of selecting the most meaningful tests and accomplishes them with expediency (by the end of 1982).
Regardless of the test program selected, we would suggest that the facility not be decommissioned in such a way as to preclude its potential use in other testing in-cluding possible support for the degraded core rulemaking proceedings. An evaluation of the potential usefulness of the LOFT facility in this regard can be made at the conclusion of the current test programs.
B-23
Dr. D. F. Ress, Jr. We would be happy to share with you some of our thoughts on the merits of the proposed test programs if you think such a dialogue would be beneficial.
Please call me or Mr. E. H. Kennedy of my staff, if we can be of any further assistance.
Very truly yours, COMBUSTION El INEERING, INC.
>> m T. E cherer- '
Director Nuclear Licensing AES:cw 1
B-24
E N E R A b((h E LE CTR I C-ct ngows GYSTEMS DIVISIOtl OF.N9UL ELEC1Ric COMPANY, 05 cuRINER AV6.. $AN JOSti CAUFORPA 95123 liC G82, (403) 925 5040-January 15~, 1981 U.S. Nuclear Regulatory Com:nission
- Office of Huclear Reactor Reguletion Washington, D.C.
205b5 Attention:
Mr. Denwood F. Ross, Jr., Director Olvision of Systems Integration Gentlemen:
SUBJECT:
COMMENTS ON FtlRTilER TESTING IN LOFT
Reference:
Letter to G. Sherscod from D. F. Ross, Jr., " Request for Carments Regardi,ig the LOFT Special Review Grcup's-Studies," dated Decernher 8,1930 This letter responds to your request (Reference) to coment on the continued operation of the LOFT facility.
The LOFT t':r,t progr:p i,
oriented toward resolution of PWR safety issuas and nakes na sututantial contribution to BWR safety technology. Therefore, General Elect' ele will only provide General comments on the propoud tuting program.
We note that in the area of BWR ECCS technology, both General Electric and-EPRI are directly involved in technical direction of NHC funded programs (BU/ECC and Reff11/Refiend).
This involve;nent insures the value of these programs in resolving real technical issues.
Cor-respondingly, part'icipation by the PWR vendors and EPRI would help 4
insure the value of programs conducted in the LOFT facility.
The LOFT Special Review Group identified four arqas for evaluation.
Coments on each follow:
1.
Operational Aspects Ulan-Hachine Interface)
This area of research is in a relatively early state of development, inerefore, inclusion of tosts specifically B-25
ti.S.110 clear Regulatory Com.lscion Page 2 directed to it in a facility such as LOFT is premature.
This effort is core suited to evaluatton in non-nuclear facilities, with any necessary nuclear evaluations conducted by perfoming less severd transient tests in operating reactors.
2.
Code Verification and Scaling Any effort in this area.will primarily apply to PWR technology.
Therefore, General Electric will not cowst.
3.
Transients, M'dS, Risk _ Reduction Any ef fort in this area will primarily apply to PWR technology.
Therefora, General Electric will not coment, 4,
Degraded Cooling Detemination of any necessary testing will not be possible until completion of the degraded core rulemakIng. Thus, consideration of this area for LOFr is premature, The -
appropriate course of action in this area is first to perfom a probabilisi.ic risk asses: ment to established safety Goals; and then, to identify what tests, if any, are needed.
I In surrinary, since the LOFT facility is primarily oriented towards testing for resolution of PWR safety issues, General Electric can not coment on its value. For areas that could possibly benefit the BWR, courses of action other than testing in LOFT are considered more appropriate at this time.
If you would like additional infomation or clarification of these connents, please contact Mr. R.H. Buchholz (408)925-5722, Very truly yours,
[M 1 n i
Safety & Licensing Operatton bcc: it.ll. Klepfer J.E. Wood GGS:rf/695-7, 4C G.E. Ofx R.H. Buchholz W.H. 0'Ardenne E. Smith B-26
U.S. NUCLEA*.') f EEULATORY COMMIS$lON 7
B12LIOGRAPHIC DATA SHEET NUREG-0758
& TfTLE AND SUBTITLE (Add VNume No., of appropriaart
- 2. (Leave bimk)
REPORT OF THE LOFT SPECIAL REVIEW GROUP
- 3. RECIPIENT'S ACCESSION NO.
- 7. AUTHOR $1
- 5. DATE hEPORT COMPLETED LOFT Special Review Group MON TH l YEAR D. F. Ross, Jr.. Chairman February 1981
- 9. PERFORMING ORGANIZATION N AME AND MAILING ADDRESS (include Esp Codel DATE REPORY ISSUED MONTH l YE AR U.S. Nuclear Regulatory Commission February 19P.I Washington, D.C. 20555
- s. (t,a,, Nanks
- 8. (Leave Nank)
- 12. SPONSORING ORGANIZATION N AME AND MAILING ADDRESS (Include Zip Codri PR E W S M M M M Same as 9 above
- 11. CONTRACT NO.
- 13. TYPE OF REPORT PE RIOD COV E RE D (Inclusive daars)
Technical
- 15. SUPPLEMENTARY NOTES
- 14. (Leave um&f
- 16. ABSTR ACT (200 words or. *ssl In July 1980 the Advisory Committee on Reactor Safeguards (ACRS) published NUREG-0699,
" Comments on the NRC Safety Research Program Budget for Fiscal Year 1982." On page 19 of NUREG-0699, the ACRS made tne following recomendation to the Commission:
LOFT represents the largest single expenditure in the safety research budget so that its program must be considered with special care. We recomend that the tests through FY 1982 be adequately funded and that following the 1982 tests the facility be decomissioned unless it is taken over by the nuclear industry.
As a result of this recommendation and pursuant to the Commission's decision on SECY 398, " Creation of a Panel of Experts to Review the LOFT Program and Report to the Com-mission," the NRC Executive Director for Operations established the LOFT Special Review Group (LSRG) on November 14, 1980. The LSRG was established to review the LOFT program and report its findings to the Commission. The primary purpose of the group was to consider wh2ther LOFT should be decommissioned in FY 1983, as recomended by the ACRS. This report represents the results of the LSRG evaluation of the LOFT program and is submitted to the Comission as an aid in its decision whether to continue NRC support of the LOFT project beyond FY 1982. The principal consensus reached by the LSRG recommends continued NRC 17.
WORDS AND bOC J ENT AN YS S 17a DESCRIPTORS i
17b. IDENTIFIERS /OPEN ENDED TERMS
- 18. AV AILABILITY STATEMENT
- 19. SE CURITY CLASS (This report /
21 NO OF PAGES Unlimited Availability Unclassified 20 SECURITY CLASS /This page)
- 22. PRICE Unclassified S
NEC FORM 335 (7 777