ML20058A825
| ML20058A825 | |
| Person / Time | |
|---|---|
| Site: | 05200004 |
| Issue date: | 11/04/1993 |
| From: | Mcintyre R, Potapovs U Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20058A814 | List: |
| References | |
| REF-QA-99900403 NUDOCS 9312010256 | |
| Download: ML20058A825 (12) | |
Text
^
ORGANIZATION:
GE Nuclear Energy San Jose, California REPORT NO.:
99900403/93-01 CORRESPONDENCE Mr. Patrick W. Marriott, Manager ADDRESS:
SBWR Project GE Nuclear Energy 175 Curtner Avenue San Jose, California 95125 ORGANIZATIONAL Mr. Kenneth W. Brayman, Manager CONTACT:
Quality Assurance Systems (408) 925-6587 NUCLEAR INDUSTRY GE Nuclear Energy (GE-NE) is engaged in the ACTIVITY:
supply of advanced boiling water reactor designs to utilities.
GE-NE also furnishes engineering services, nuclear replacement parts, and dedication services for commercial grade electrical and mechanical equipment.
INSPECTION CONDUCTED:
August 9 through 13, 1993 lO 2't h3 SIGNED:
4 Richard P. McIntyr'e, Team Ledher Date Reactive Inspection Section No. 1 Vendor Inspection Branch (VIB)
APPROVED:
1)_A2
(
rh e ll-4 'T3 Vidis Potapovs,"Chie f Date Reactive Inspection Section No. I Vendor Inspection Branch (VIB)
INSPECTION BASES:
10 CFR Part 50, Appendix B and 10 CFR Part 21 INSPECTION SCOPE:
To determine if activities performed as part of l
the Gravity-Driven Cooling System (GDCS)
-Integrated Systems Test (GIST) program were conducted under the appropriate provisions of the GE-NE 10 CFR Part 50, Appendix B, QA program, as implemented by the " Advanced Light Water Reactor Quality Assurance Program Plan",
prepared for Department of Energy Contract Number DE-AC03-86SF16563 and also to review the input modeling of the TRACG computer code for the GIST facility.
PLANT SITE APPLICABILITY:
None 93:201o256 931118 PDR ADOCK 05200004 A
1 INSPECTION
SUMMARY
1.1 Nonconformances 1.1.1 Contrary to Criterion III of Appendix B to 10 CFR Part 50 and Section 4.4.1 of Engineering Operating Procedure (E0P) 40-3.00, " Engineering Computer Programs" (ECPs), (1) the TRACG input decks used to model the gravity-driven cooling sysic= integrated systems test (GIST) facility were not independently verified to be correct, and (2) the GE-NE Code Qualification Document (CQD),
Licensing Topical Report NEDE-32177P, "TRACG Qualification," dated February 1993, which provides a description of the qualification of TRACG against various activities including the gravity-driven cooling system (GDCS) integrated systems test, was submitted to the NRC for review and approval for referencing in licensing actions for the Simplified Boiling Water Reactor (SBWR) without receiving independent design verification or design review as required for a level I code used to support design basis analyses.
(93 01) 1.1.2 Contrary to Criterion XVII of Appendix B to 10 CFR Part 50, E0P 42-10.00, " Design Record Files" (DRF), E0P 35-3.00, " Engineering Tests,"
and the GIST Program Test Plan and Procedure (TP&P) 521.1322, Revision 2, dated November 29, 1988, certain documentation required to be contained or referenced in the DRF was rot included therein. Specific documents that should have been part of the DRF were: the Final Test Report (NED0-31680) for the GIST Program; instrument calibration records, which were located in a desk drawer in another building; and final design drawings for the facility.
Some drawings were found in a cabinet at the facility itself.
This set of drawings did not include final numbered, approved, as-built design drawings, which are required by the QA Plan to be retained for the lifetime of the item. Also, data tapes for the GIST tests, which are part of the test records specified for inclusion in the DRF by TP&P 521.1322, were not referenced therein.
(93-01-02) 1.1.3 Contrary to Criterion III of Appendix B to 10 CFR Part 50, and Section 3.11 of the GE-NE QA Program Description topical report, NED0-ll209-04A,
" Design Change Control," there was no documentation or verification of changes made to the TRACG code as a result of the GIST program.
The changes include changes to the interfacial shear and heat transfer when a two-phase level is present, changes to the model for condensation on cold walls when air is present, and the implementation of a horizontally stratified flow map.
(93-01-03) 1.1.4 Contrary to Criterion XII of Appendix B to 10 CFR Part 50, Section 2.2 of E0P 35-3.20, " Calibration Control," and Section 4.1.2 of the GIST Program TP&P 521.1322, GE-NE purchased flow meters used in the GIST tests from a commercial grade supplier, not on GE-NE's approved supplier list, and accepted and used the instruments as calibrated by the supplier without further verification of the quality or traceability of those calibrations.
(93-01-04) 1.1.5 Contrary to Criterion XI of Appendix B to 10 CFR Part 50, E0P 35-3.00,
" Engineering Tests," E0P 42-10.00, " Design Record Files," and Section 4.2.4 of.
the GIST Program TP&P 521.1322, GE-NE failed to document in the DRF the review and disposition of anomalies in three tests, C01, 001, and D03.
These tests were considered to be " invalid" as a result of incorrect valve alignment (C01) or incorrect power input to the test section (D01 and D03).
For one of the tests (C01), a note was found on the folder in the DRF in which hard-copy data plots were stored, indicating that a problem existed for the test; however, the problem indicated on the folder (incorrect power input) was not consistent with the actual reason given in NED0-31680 for the test's invalidation (incorrect valve alignment).
(93-01-05) 1.2 Unresolved Item The lack of independent design verification for the TRACG computer code raises questions concerning the validity of other calculations included in Chapters 6 and 15 of the SSAR for SBWR. This issue is considered an unresolved item and will be discussed with GE-NE in future meetings.
(93-01-06) 2 STATUS OF PREVIOUS INSPECTION FINDINGS No previous inspections have been conducted in this area.
3 INSPECTION FINDINGS AND OTHER COMMENTS 3.1 Ouality Assurance Proaram The quality assurance program implemented for the GIST program is described in the Advanced Light Water Reactor Program Quality Assurance Program Plan (QA Plan) that was prepared for the Department of Energy, Contract No.
DE-AC03-86SF16563.
This DOE QA Plan mandates the application, as appropriate to the contract scope, of the QA program described in Revision 6, dated June 30, 1986, of the NEDO-ll209-04A, " Quality Assurance Program Description."
This is the GE-NE topical report that has been reviewed and approved by the NRC and meets Appendix B of 10 CFR Part 50.
The DOE QA Plan contains a work element / implementing procedure matrix that contains 18 major subdivisions which correlate with the 18 criteria of Appendix B.
The 18 subdivisions are further broken down into 78 work elements committed to the QA Plan.
Four types of GE-NE procedures are described that implement the worl. elements of the QA Plan. These are Nuclear Energy Business Operations (NEBO) Policies and Procedures (P&Ps), BWR Engineering Operating Procedures (EOPs), Nuclear Systems and Technology Operation (NSTO) Policies and Instructions, and Nuclear Service Procedures (NSPs).
During the time frame of the GIST program, the NSPs were phased out and folded into the E0Ps.
The NEB 0 P&Ps are high level GE-NE policies that establish overall policies and responsibilities for NEB 0.
As a result of a reorganization, GE-NE nuclear activities are currently under the Vice President of GE Nuclear Energy and NEB 0 no longer exists.
The E0Ps are a series of procedures that implement GE-NE policies and the QA program. NST0 Policies and Instructions deal with subjects such as cost schedules, budgeting, contract award, and business management and, as such, have no importance in implementing the QA plan. The NSPs have been subsumed by the E0Ps. This process was ongoing during the GIST test period.
The GIST test program and QA program both fell within the NEB 0.
The Quality Assurance Operations within NEB 0 was a staff organization assigned responsibility for establishing the NEB 0-level quality related P&Ps, and i
auditing the functional organizations involved in the activity.
The QA program covered the design, procurement of parts and testing equipment, instructions and drawings, document control, inspection, test control, control of measuring and test equipment, corrective actions, quality assurance records, and audits.
4 3.2 Jnstruction. Procedures and Drawinas The quality requirements utilized for GIST are contained in the following GE-NE documents reviewed by the inspection team:
NEB 0 P&P 70-11, " Quality Systems Requirements"
=
E0P Nos. 15-2.00, "EOP Application"
=
30-5.00,
" Engineering Records Documentation Supplied by External
=
Sources" 30-7.00,
" Technical Design Procedures" 35-3,00,
" Engineering Tests" 40-3.00,
" Engineering Computer Programs"
=
40-7.00,
" Design Reviews" 42-5.00,
" Engineering Requirements Document Release" 42-6.00,
" Independent Design Verification" 42-8.00,
" Document Issue and Application by ERM" 42-10.00, " Design Record Files"
=
55-2.00,
" Engineering Change Control" 60-3.10,
" Engineering Records Retention" i
60-6.00,
" Drafting Manual Control" 65-2.00,
" Product Safety Requirements" 3.3 Document Control Document control is prescribed by NEB 0 P&P 70-11, " Quality Systems Requirements" and numerous E0P's such as:
15-2.00, "E0P Application;"
30-5.00, " Engineering Records Documentation Supplied by External Sources;"
i 40-7.00, " Design Reviews;" 42-5.00, " Engineering Requirements Document Release;" 42-6.00, " Independent Design Verification;" 42-8.00, " Document Issue and Application by ERM;" 55-2.00, " Engineering Change Control;" and 60-6.00,
" Drafting Manual Control."
The team found poor control of some design and calculation documents.
Some calculations documents were kept in an individual's desk, and original design drawings were kept in an unlabelled shop drawing file.
Considerable effort was needed to locate and determine what drawings contained the as-built j
elevations of key components of the test facility. Additionally,. there were no final approved revisions for the drawings reviewed.
See Section 3.5 below on Design Control. -
3.4 ILuality Assurance Records Quality Assurance records are prescribed by NEB 0 P&P 70-11, " Quality Systems Requirements," and numerous E0P's such as:
35-3.00, " Engineering Tests;"
40-7.00, " Design Reviews;" 40-9.00, "ASME Code Design Verification;" 42-6.00,
" Independent Design Verification;" 42-10.00, " Design Record Files;" and 60-3.10. " Engineering Records Retention."
E0P 42-10.00 describes Design Record Files (DRFs) as formal, organized accumulations of information, which provide a controlled system for retention of documented engineering activities, necessary to substantiate significant design decisions.
The DRF provides a mechanism for controlling and archiving important design records, such as design verification, studies and analyses.
It does not include documents, such as drawings and specifications, which are maintained under separate corporate design controls.
However, the DRF is to include documentation and pertinent references that support the design.
The procedure also says that the DRF should provide for design notes, calculations, records and other supporting information, and cross-reference to related or supporting DRFs.
The team found that record retrievability was lacking in that the DRF did not contain all relevant records or a reference to important test records such as design drawings (or reference to their location, since the DRF is not required to contain drawings not easily reproducible into such media as a microfiche),
the final test report, instrument calibrations, and reason (s) for not repeating a test.
The location of several records was determined by calling in the retired responsible test engineer. Some calculation records were in another DRF and some were located in an individual's desk drawer. - This resulted in the team spending considerable effort in finding if records existed and being able to obtain the record that did exist. The DRF did not provide traceable and retrievable evidence to support the GIST test and results.
3.5 Desian Control The NRC inspectors examined the DRF for the GIST test program, DRF A00-02917, to determine if the documentation in the DRF complied with the QA Plan requirements in effect during the GIST program, and associated E0Ps 42-10.00,
" Design Record File;" 42-6.00, " Independent Design Verification;" 40-7.00,
" Design Reviews;" and 35-3.00, " Engineering Tests."
E0P 42-10.00 requires, in part, that the DRF contain or reference (as applicable) design and evaluation records, test reports, and controlled documents.
Evaluation records are further defined in E0P 35-3.00 as including instrument calibration records.
Design records include drawings for the facility, specifically showing nominal and as-built dimensions for components and the facility as a whole.
These drawings should be numbered documents that are retained and retrievable, and that can be referenced as part of the DRF.
The inspectors found that documentation required to be contained or referenced in the DRF was not included therein.
Specific documents that should have been part of the DRF were:
the Final Test Report for the GIST Program, NED0-31680;
_4_
.N instrument calibration records, which were located in a desk drawer in another i
building; and design drawings for the facility, some of which were found in a cabinet at the facility itself.
This set of drawings, however, did not include final numbered and approved, as-built design drawings, which the QA Plan requires to be retained for the lifetime of the item.
Sketches found in the DRF, apparently made by the responsible test engineer and containing some as-built dimensions, are not acceptable substitutes for numbered, retrievable, as-built drawings.
In addition, the drawings that were located did not in all cases represent a final as-built configuration, and determining which drawings had been superseded and which were representative of the final facility was difficult.
Other deficiencies in the DRF include:
(1) failure to reference the original data tapes from the GIST tests, which should be appropriately referenced as retrievable information; (2) failure to include verified input from TRACG analyses related to GIST; and (3) failure to include documentation or experimental verif L. Mn of engineering calculations.
For instance, the heat loss to the envir
'ent ' rom the facility was a necessary datum for accurate analysis of the test results. A single page with an estimate of the heat loss was located in the DRF, but no documentacion was included to show the basis for the estimate of effective heat transfer coefficient (e.g., thermophysical properties of insulation, film coefficient on outside of insulation, etc.),
nor was there any indication that the calculated estimate was verified experimentally.
As a result, Noi.conformance 93-01-02 was identified during this part of the inspection.
GE-NE acknowledged the shortcomings in the DRF, and committed to upgrade the DRF to include, to the extent possible, the missing information and documents i
described above or references thereto, where appropriate.
3.6 Test Control E0P 35-3.00, " Engineering Tests," describes the classification system of test program types.
Appendix B to E0P 35-3.00 describes five test program types.
Type B1 is a development type, B2 is a design basis type, B3 is a design qualification type, B4 is a manufacturing type, and B5 is a special test.
Type B5 is classified as not safety related.
The GIST test was considered a type B1 and was not classified as a safety-related test.
The inspectors reviewed the records relating to control of the GIST test program, including the Test Requirements and Test Specifications (TR&TS) and the Test Plan and Procedures (TP&P) dncuments. Conformance with the requirements of E0P 35-3.00 was also checked.
The inspectors determined that the classification of the GIST test program as a non-safety-related type B1 (developmental) test, per Appendix B of E0P 35-3.00, was not conaistent with the use of the results of the test program.
This inconsistency had a significant impact un several of the activities associated with the test program, including GIST-related computer analyses using the TRACG code, QA and control of equipment and instrumentation purchased for use in the facility, independent design verification of the facility, and treatment of the data.
The NRC asserts that the GIST test program comprised a safety-related activity, and that the program's purpose was, in part, to obtain design basis data for the proposed SBWR gravity-driven cooling system.
The SBWR design is being considered for certification as a " passive" reactor design under the requirements of 10 CFR Part 52.
Included in Part 52 are specific requirements for the data and testing requirements; these are found in l
10 CFR 52.47(b)(2)(i)(A), and state, in part, that certification of a passive standard design will be granted only if:
(1)
The performance of each safety feature of the design has been demonstrated through either analysis, appropriate test programs, experience, or a combination thereof; (2)
Interdependent effects among the safety features of the design have been found acceptable by analysis, appropriate test programs, experience, or a combination thereof; and 1
(3)
Sufficient data exist on the safety features of the design to assess the analytical tools used for safety analyses over a sufficient range of l
normal operating conditions, transient conditions, and specified accident sequences, including equilibrium core conditions.
]
Documentation in the DRF clearly shows that the GIST test program was considered, at the time it was being developed, to be the only program of its i
kind needed to demonstrate the principle and test the performance of a gravity-driven cooling system.
The intent of the test program, as stated in various memoranda in the DRF and in the GIST program Final Report, EEFR-00850, has been to use the test data to demonstrate the performance of the GDCS, as required in (1) above, and to establish a database for code qualification to perform safety analyses, as required in (3) above.
In the view of the NRC
{
staff, it appears that there is a clear and direct connection between the
{
performance of the GIST tests and the safety analyses contained in the SBWR
)
Standard Safety Analysis Report. Therefore, the use of the data constitutes a type B2 test classificate;n of E0P 35-3.00 and should have been conducted as a safety-related activity with appropriate QA controls in place commensurate with the requirements for safety-related tests.
Furthermore, since GIST testing was a safety-related activity, the use of the data in the TRACG Code Qualification Document (CQD) that was submitted to the NRC for review and approval for referencing in licensing actions for the SBWR, constitutes a design application of the TRACG02 (Level 1) computer code, and as such requires independent verification or design team review of all calculations.
Failure to properly classify the test program led to a failure to recognize the design application of TRACG02, which in turn led to a failure to independently verify GIST-related code calculations as well as the CQD for TRACG. As a result, part of Nonconformance 93-01-01 was identified during this part of the inspection.
Insofar as the test data themselves are concerned, it appears that GE-NE was not consistent with E0P 35-3.00 requirements with regard to disposition of the data in a Final Test Report (FTR).
Item 4.3.16 in E0P 35-3.00 instructs the responsible test engineer to " prepare FTR providing for the complete description of the components tested and reduction, interpretation, and _ - - _ _ _ - - _ _ _ - _ - _ _ _ _ _ _ - _ _ - - _ _ - _ _ _ - _ _ _ _.
correlation of the data as specified in the work authorization. Obtain required evaluations of technical validity by the Test Requestor and Responsible Test Manager and approval to issue the FTR as a traceable document consistent with E0P 42-10.00."
As noted in section 3.5 of this report on Design Control, the F;r was not included or referenced in the DRF.
Furthermore, there is i.o evidence in the FTR that the data were " interpreted and ccrrelated." Specifically, no attempt appears to have been made to estimate the errors associated with the data, either from instrument uncertainties or from basic experimental uncertainties, such as heat losses from the facility.
No evidence of verification of facility operating characteristics was available, such as steady-state heat balances to verify calculations of faci 1My heat loss or loss coefficients for flows through facility components and piping.
Failure to provide such information renders the data of questionable validity.
GE-NE personnel also raised two issues regarding classification of the test program in discussions during the inspection.
First, the point was made that 10 CFR 52 was incorporated into the Cod. of Federal Regulations in April 1989, after the GIST program was completed. The implication by GE-NE was clearly that, since the regulation did not exist when the program was performed, its requirements related to passive reactor testing programs were inapplicable to the GIST tests.
The NRC staff disagrees with this position. Memoranda in the DRF state clearly that the GIST program was to be a one-time-only test program; thus, there was never any intention to use the GIST tests as a developmental step in an integrated design and test effort.
The intent to use i
these data to qualify TRACG for SBWR safety analyses was also clear, since such analyses of the early stages of design basis accidents in the SBWR were used to help establish the required initial conditions for the GIST tests.
Comparisons of GIST results to expected SBWR behavior are made in GEFR-00850.
The safety-related nature of the tests should therefore have been recognized by GE-NE, whether or not Part 52 existed at the time of the tests.
Additionally, the submission of both the SBWR SSAR and the TRACG CQD, and the inclusion of GIST information as supporting documentation, occurred after issuance of Part 52. Any material submitted to the NRC must therefore conform to the requirements of the regulations.
It was also asserted initially by GE-NE that the requirements for independent design verification in E0P 35-3.00 did not apply to type B1 tests.
This claim was made on the basis of item 4.1.3(b) in the E0P, which gives as examples i
type B2 and B3 tests, but does not mention type B1 tests.
In subsequent discussions between the NRC and GE-NE QA personnel, GE-NE admitted that this was an erroneous interpretation of the E0P requirement, and that type B1 tests were also covered.
In fact, the DRF contains a memorandum requesting an exemption from independent design verification of the GIST test-facility, with the justification that no controlled drawings of the SBWR design existed at that time (ca. 1986) against which to verify the design. The independent verification process per E0P 42-6.00 was therefore waived.
However, an internal design review process was established through use of Engineering Review Memoranda (ERMs) and a Design Review Team (DRT). Appropriate records of ERMs and their resolution, and of the review by the DRT, are contained in the DRF. The NRC inspectors consider these processes to constitute an adequate design review, and regard the documentation of this process in the 1
DRF as acceptable.
The NRC also notes that an external design review was performed by the Electric Power Research Institute, documentation of which is contained in the DRF.
i 3.7 Instruments. Calibration. and Procurement The QA Plan stated t At the control of measuring and test equipment was covered by E0P 35-3.20, " Calibration Control." Section 2.2 of E0P 35-3.20 stated that maintenance and test equipment calibrations were to be performed using controls which assured traceability to certified equipment having known valid relationships to nationally recognized standards.
In aadition, E0P 35-3.20 stated that calibration services should be classified as safety-related services unless justified and documented otherwise.
GE-NE TP&P 521.1322, Revision 2, dated November 29, 1988, stated in section 4.1.2 that test equipment should be calibrated against auditable standards traceable to the National Bureau of Standards.
The GIST test facility used approximately 120 instruments to monitor process parameters such as temperature, pressure, conductivity (level), and flow.
The NRC inspectors asked to see the calibration records for the instruments used during performance of the GIST tests It was determined that the calibration records were not part of the DRF and that they were not on file in the Instrument Laboratory or the Calibration Laboratory.
GE-NE located the calibration records in the GIST test facility.
Review of the records and discussion with GE-NE indicated that the thermocouples and pressure transducers had been purchased as commercial grade items, and the conductivity probes manufactured by GE-NE. All of the thermocouples, pressure transducers and conductivity probes had been calibrated by the GE-NE Instrumentation Laboratory in accordance with established procedures and the requirements of the GE-NE Appendix B quality assurance program and that the calibrations were against auditable standards as required by the TP&P.
Although verification of the quality of the performance of test instruments purchased as commercial grade could be accomplished by calibration under an Appendix B quality assurance program, this would not apply to construction material.
However, review of purchase orders associated with the GIST tests indicated that GE-NE had also purchased the material used to construct the test facility as commercial grade items without any further dedication activities.
GE-NE did indicate that consideration of quality was indicated for the piping used in the GIST facility which was purchased to ASME specifications.
GE-NE had purchased the flow meters as commercial grade items from Flow Technology Inc., of Santa Clara, California, by purchase order number i
190-RP666, dated September 8, 1987. The purchase order specified that the instruments were to be calibrated by the supplier, although-it did not request a certificate of conformance. The purchase order did not specify that the calibrations performed by the supplier were required to meet any specified quality assurance program, 10 CFR Part 50 Appendix B, or 10 CFR Part 21.
Flow Technology Inc., was not on GE-NE's approved suppliers list as qualified to provide safety-related instruments or to perform safety-related calibrations of instruments and GE-NE had not performed any audits or surveys of Flow
e Technology Inc., to support acceptance of the calibrations.
GE-NE had received calibration records from the supplier and used this information as computer input for performance of the GIST tests although the quality of this information, from an unapproved commercial grade supplier, was not verified.
GE-NE had performed a single point check of the instruments at 1
zero flow but this did not meet the criteria of a calibration of the
)
instruments or verify the required operation over the specified range.
GE-NE had not classified the calibration of the flow meters as a safety-i related service as required by E0P 35-3.20 and did not procure the service from a qualified supplier, take actions to qualify the supplier, or perform the service under the GE-NE quality assurance program.
Since GE-NE had not audited Flow Technology Inc. for their approved suppliers list or performed surveys to support the calibration activities, GE-NE did not have a basis for accepting the calibration of the flow meters by Flow Technology Inc., and 1
therefore had not ensured that the instruments had been adequately or correctly calibrated against auditable standards as required by the Program Plan and the TP&P.
In addition, GE-NE had not verified that the flow meters, instruments used in an activity affecting quality, were properly controlled, calibrated, and adjusted to maintain accuracy with necessary limits as required by Criterion XII, " Control of Measuring and Test Equipment," of Appendix B to 10 CFR Part 50. As a result, Nonconformance 93-01-04 was identified during this part of the inspection.
3.8 Nonconformances and Corrective Actions TP&P 521.1322 stated in section 4.2.3 that deviations and out of specifications should be documented on the Deviation Log, figure 1-7 of the TP&P, and reported to the Responsible Test Engineer.
Section 4.2.4 of the TP&P stated that a nonconformance report should be generated for test results which did not meet the acceptance criteria, expected results, or contained anomalies.
The TP&P required that the Test Requestor establish a test disposition and approve all dispositions prior to resuming the test and include a copy of the completed, approved nonconformance reports in the DRF.
The inspectors reviewed the Final Test Report - Testing of the Gravity-Driven Cooling System for the Simplified Boiling Water Reactor (Final Test Report),
NEDD-31680, dated July 1989, and the SBWR Program GDCS Integrated Systems Test Final Report (Final Report), GEFR - 00850, dated October 1989, and determined that at least three anomalies occurred during performance of the matrix texts which required a change to an input parameter or the test configuration and met the definition of nonconformance provided in the TP&P. The three matrix tests (0 1, D01, and D03) were indicated in test reports as being " invalid,"
for various reasons,. These tests were rerun successfully as tests C01A, DOIA, and D03A and are described below.
GDCS Line Break - Base Case (Tests C01 and C01A)
The Final Report stated the following: Test C01 was repeated because the vacuum breakers between the wetwell and the upper drywell were not functioning. The broken GDCS line injected not only hot water and steam
-9 i
from the vessel but also cold water from the suppression pool.
This cold water condensed the steam in the upper drywell, lowering its pressure below that of the wetwell.
Since the vacuum breakers were out of service in Test Col, there was a danger of back-flow from the suppression pool through the main vents. This could have caused such a rapid (and non-SBWR typical) depressurization in the upper drywell that the GIST facility could have been damaged.
To prevent this, the operator continually injected steam into the upper drywell during the test.
Since this injection was not measured and the containment behavior was not typical of the SBWR (as all base cases were required to be), the test was invalid. The vacuum breakers were functional in Test C01A and all other GDLB tests.
In addition, a note was found on the folder in the DRF in which hard-copy data plots were stored which indicated that a problem had existed concerning the power during the test. This was not consistent with the actual reason given in NED0-31680 for the test's invalidation (incorrect valve alignment).
No Break - Base Case (Tests D01 and D01A)
The Final Report stated that during performance of test D01 the power (in the reactor pressure vessel) did not decay as expected such that a repeat test was necessary.
No Break - Appendix K Decay Heat (Tests D03 and D03A)
The Final Report stated that during performance of Test D03 the power provided to the heater coils was too low.
In Test D03A the correct power was provided to the heater coils.
GE-NE did not document the reasons for the failure-of the tests to meet acceptance criteria as either deviations or nonconformances in accordance with the TP&P, although in each case a change to an input parameter or the test configuration was required.
Ultimately, these anomalies were described in the Final Report and the Final Test Report (but not documented as deviations or nonconformances and thus not available for inclusion in the DRF).
During the period of time between performance of the test -(fall 1988) and issuance of the Fi al Test Report (July 1989) and the Final Report (October 1989) there was no documentation of the anomalies or actions taken. GE-NE had not ensured that test results were documented and evaluated to assure that test requirements had been satisfied as required by Criterion XI, " Test Control," of Appendix B to 10 CFR Part 50. As a result, Nonconformance 93-01-05 was identified during this part of the inspection.
3.9 TRACG Computer Modelina of GIST Several concerns were identified in the TRACG computer modeling of GIST and in TRACG configuration control.
E0P 42-10.00, " Design Record File," requires that supporting information on calculations be included or referenced in the l
DRF. The computer model calculational notebook was not part of the GIST DRF and the DRF did not contain enough information to generate a TRACG computer i
r.
s model. However, a calculation notebook for the GIST TRACG input deck used for the calculations included in the GIST final report was maintained by the modeler.
It contained enough information to generate a TRACG input model for GIST but it did not adequately document where the information came from.
The calculation notebook was not independently checked for accuracy. The input models contained an error in the elevation of the GDCS tank.
Different input decks contained different elevations for the GDCS tank, which resulted in errors in the input to TRACG02. As a result, part of Nonconformance 93-01-01 f
was identified during this part of the inspection.
Since the errors involved, in part, elevations of components in the simulated GDCS in GIST, which determine the driving head available for gravity-driven injection, they have a direct impact on the results of GIST analyses, and call into question all i
safety calculations related to the GIST program.
The subject of other SBWR SSAR Chapter 6 and Chapter 15 safety calculations performed with TRACG was discussed with GE-NE. The calculations were performed with a level I version of TRACG, which means that the input, output and code range of applicability must be determined for each calculation performed.
GE-NE stated that the calculations were not yet independently verified and that the NRC should be aware of this fact.
It was mutually agreed that this issue will be discussed with GE-NE in a future meeting.
Therefore, this issue is considered an unresolved item (93-01-06).
A separate calculation notebook was kept for the TRACG GIST calculations included in the TRACG Code Qualification Document. This calculation notebook did not contain enough information to construct the TRACG model and only documented the differences between the input model used for the GIST final report and the input model used for code qualification. This calculation notebook also was not independently checked.
These input decks also had the same elevation errors for the GDCS tank until the errors were discovered by an Idaho National Engineering Laboratory (INEL) staff member doing an independent analysis of GIST for the NRC Office of Nuclear Regulatory Research. A recent i
calculation of one GIST experiment using correct elevation information showed 20 percent less GDCS flow than the calculation presented in the code qualification document.
These elevation errors could have been prevented by checking the code messages which provide a warning when there are improperly closed loops.
GE-NE may consider identifying this as a fatal input error instead of just a warning, considering the importance of natural circulation in BWR systems.
Heat losses to the environment for the computer modeling were taken from a hand calculation and not from experimentally measured data.
The memo with the hand calculation indicated that the engineer in charge of the experiment remembered that the heat losses may have been twice the amount of the hand calculation.
Even with the uncertainties and heat losses that may have been as high as 15 to 20 percent of the total power input, no sensitivity studies were performed in the TRACG computer studies of the GIST experiments.
The inspection also examined computer code changes made for the GIST program.
The specific changes include changes to the interfacial shear and heat transfer when a two-phase level is present, changing the model for condensation on cold walls when air is present, and the implementation of a horizontally stratified flow model.
In GE-NE's code classification system a level I computer code is a developmental code.
It may be used for design,
l i
calculations provided that the input, output, and range of applicability are independently verified.
A level 2 computer code is a design code.
When using a design code, only the input and the range of validity have to be verified.
The inspection team was told that documentation on the implementation and l
testing of these new code models is not required for a level I computer code and does not exist in any official form.
GE-NE seems to have not performed i
the more extensive documentation and testing required to change a level 2 code by declaring the changed code a new level I code.
This also circumvents the responsibility of assessing how the changes to the code affects previous calculations done with the old level 2 code.
The inspection team was also told that GE-NE does not use a code librarian such as HISTORIAN or UPDATE for configuration control of TRACG. Code changes and configuration control are not maintained manually by the code responsible engineer. The lack of documentation for testing of new models, the method of implementation into the computer code and the lack of independent verification may allow programming errors to slip by undetected.
As in the case of the input deck errors, code changes could have significant effects on calculated results and still not be discovered through the design review process.
Independent of the lack of verification at the present time are the issues of the minimum standards that must be met in order to independently verify a calculation and what kind of checks must be made of the calculationi.
notebooks and input decks.
GE-NE uses a process called design revie.:... order to meet the independent verification requirements of Appendix B.
Design review is covered by E0P 40-7.00.
While this procedure has some suggested means of independent verification, it does not have any minimum requirements.
The level of independent verification performed is decided upon by the design review team.
In the case of the TRACG design review, the committee was only reviewing the models that were stated to be in the code and the code output for test cases.
l The review committee seemed to be assuming that the inputs for the test cases were already verified to be correct.
This assumption was found not to be correct. None of the individual test cases had been treated as a design calculation and therefore none were independently verified. Comments and questions from the design review team indicated that no review team member had any questions about the incorrect GIST calculations since the output of the calculations looked " reasonable." The NRC staff considered this an example of how the GE-NE design review process could fail to detect errors when no minimum level of independent verification is required.
GE-NE's methods of independent verification and record keeping do not appear to meet the requirements of its QA Topical Report.
As a result, Nonconformance 93-01-03 was identified during this part of the inspection.
W :
- .g 4 PERSONNEL CONTACTED GJ Nuclear Enerav i
R.H. Buchholz, Advanced Reactor Programs (ARP), Manager, SBWR Joe Case, Manager, Nuclear Quality Assurance (NQA)
Ken Drayman, Quality Systems Manager, NQA forrest Hatch, Manager, Services & Projects Quality i
Philip Novack, Quality Assurance Manager, ARP Don Kaye, Quality Assurance, ARP Jay Murray, QA Audits Manager, NQA Paul F. Billig, Senior Engineer, ARP i
Mohammed Alamgir, Safety & Thermal Hydraulic Methods David Foreman, SBWR Licensing Bob Mitchell, Safety Evaluation Programs Jim Klapproth, Fuel Licensing Frank Paradiso, ABWR System Integration Gary Dix, Manager, Fuel Quality Assurance Norman E. Barclay, Service & Projections Quality Jim Shaug, Safety & Thermal Hydraulic Methods Albert Yang, ABWR System Integration Jeff Baechler, SBWR Certification David Sandusky, Materials Applications and Test Operations Bill Zschaler, Test Facilities Engineering Nuclear Reaulatory Commission Richard McIntyre, Team Leader, Vendor Inspection Branch (VIB)
Joseph Staudenmeier, Analytical Support Group Uldis Potapovs, Section Chief, VIB Frederick R. A11enspach, Performance & Quality Evaluation Branch Billy Rogers, Vendor Inspection Branch Alan Levin, Reactor Systems Branch Bob Jones, Chief, Reactor Systems Branch (conference call)
Mark Rubin, Section Chief, Reactor Systems Branch (conference call)
U.S. Rohatgi, Brookhaven National Laboratory Denartment of Enerav Kashmira Mali, San Francisco Field Office Trevor Cook (conference call) i
,