ML20204F978

Forwards NRR Evaluation of Util Reanalysis Rept on Verification Testing for Auxiliary Power Sys Voltage Study in Compliance W/Nrc Branch Technical Position PSB-1. Further Testing & Analyses Recommended

ML20204F978
Person / Time
Site:	Sequoyah
Issue date:	08/01/1986
From:	Youngblood B Office of Nuclear Reactor Regulation
To:	White S TENNESSEE VALLEY AUTHORITY
References
NUDOCS 8608070021
Download: ML20204F978 (14)

Text

i s Docket Nos.: 50-327/328 lAl1G1986 Mr. S. A. White Manager of Nuclear Power Tennessee Valley Authority 6N 38A Lookout Place 1101 Market Street Chattanooga, Tennessee 37402-2801

SUBJECT:

VERIFICATION TESTING FOR AUXILIARY POWER SYSTEM VOLTAGE STUDY FOR SEQUOYAH Enclosed is the NRR evaluation of the TVA reanalysis report on verification testing for the auxiliary power system voltage study in compliance with NRC Branch Technical Position PSB-1.

We find that the new TVA computer program can adequately predict the steady state response of Sequoyah's power system down to the 480 volt level.

However, there is insufficient basis for us to conclude that the computer program can predict the transient response of the system or the steady state response at the 120/208 volt level.

We recommend further testing and analyses as described in the enclosed evaluation report.

Sincerely, B. J. Youngblood, Director PWR Project Directorate #4 Division of PWR Licensing-A

Enclosure:

As stated cc: See Next Page DISTRIBUTION: See Other Page

(

I Il" / \ I .l -

. /p .

,) , /

PWR# 1, +A PWR P4 -A PW f CSta. e':) h k

MDuncan ' BJYoin/DPWR-A gblood

/86' 7/ /86 7/ /86 7/}

\

8608070021 860801

,hDR ADOCK 05000327 PDR j

DISTRIBUTION:

Docket File NRC PDR Local PDR PRC System NSIC PWR#4 Rdg.

MDuncan BJYoungblood Rdg.

CStahle TVA0G (3) S. Richardson 5029 HDenton JTaylor BHayes GZech NGrace LSpessard SWeise, RII SAConnelly DMuller TNovak BJYoungblood JHolonich TKenyon WLong TAlexion BKSingh KHooks ACRS (10)

OELD JPartlow BGrimes EJordan 6

l l

, Mr. S. A. White Tennessee Valley Authority Sequoyah Nuclear Plant I cc:

Herbert S. Sanger, Jr. , Esq. Tennessee Department of Public General Counsel Health Tennessee Valley Authority ATTN: Director, Bureau of 400 West Sumit Hill Drive, E 11B 33 Environmental Health Services Knoxville, Tennessee 37902 Cordell Hull Building Nashville, Tennessee 37219 Mr. K. W. Whitt Mr. Michael H. Mobley, Director Tennessee Valley Authority Division of Radiological Health 400 West Sumit Hill Drive, E3A8 T.E.R.R.A. Building Knoxville, Tennessee 37902 150 9th Avenue North Nashville, Tennessee 37203 Mr. Bob Faas Westinghouse Electric Corp. County Judge P.O. Box 355 Hamilton County Courthouse Pittsburgh, Pennsylvania 15230 Chattanooga, Tennessee 37402

- Mr. Jerry Wills

- Tennessee Valley Authority SN 133B Lookout Place Chattanooga, Tennessee 37402-2801 Mr. Donald L. Williams, Jr.

-Tennessee Valley Authority 400 West Sumit Hill Drive, W10B85 Knoxville, Tennessee 37902 Resident Inspector /Sequoyah NPS c/o U.S. Nuclear Regulatory Commission 2600 Igou Ferry Road Soddy Daisy, Tennessee 37379 Regional Administrator, Region II U.S. Nuclear Regulatory Commission, 101 Marietta Street, N.W., Suite 2900 Atlanta, Georgia 30323 l

l l

L __ ___ _ _ . _ _ _ ____

j ENCLOSURE 1 EVALUATION REPORT VERIFICATION TEST FOR AUXILIARY POWER SYSTEM VOLTAGE STUDY UNDER PSB-1 SEQUOYAH UNITS 1 AND 2 BACKGROUND In reference to verifying the validity of voltage drop calculations for Sequoyah Nuclear Plant Auxiliary Power System, a letter from B. J. Youngblood to S. A.

White, dated March 26, 1986, recommended that TVA perform a new verification test as prescribed in Branch Technical Position PSB-1 (part B.4) since the staff could not conclude that TVA had adequately demonstrated that a new verification test was not necessary. This recommendation was largely based on Electrical, Instrumentation and Control Systems Branch's (EICSB) finding of the inconsistency between the measured load data during the test and how it was used in the analysis and the questionable accuracy of the test instrumentation recording. A meeting was held on April 16, 1986, during which TVA presented additional information and clarification to their October 2, 1980 test report in support of its position that performing additional verification testing of the Auxiliary Power System at Sequoyah was not necessary. Subsequently, TVA agreed to provide additional information for our further consideratian as follows:

e

1) Confirmatory analysis to demonstrate that the new computer proaram is comparable to the old computer program which was used in the original test report by inputting the same load values into the new computer program and comparing the calculated voltages.

r --

t

2) Analyses to demonstrate that there is no significant configuration change between the 1980 and 1986 systems by inputting the data from the July 12 & 16, 1980 tests into the 1980 & 1986 system models and comparing the calculated voltages.

3) More detail on how the two tests (July 12 & 16, 1980) were conducted and description of how the circuit breakers were aligned for each configuration.

I On June 2, 1986, TVA provided a response to our previous concerns on the changes along with a report titled "NRC Branch Technical Position PSB-1 Reanalysis."

EICSB has reviewed the additional information and has re-evaluated the previous position on the need for the verification test.

6

- - - , - , - , - - - -- -- ~-, . - , . . -,.n.-, , -n - ,--,. -- ,. - ,- - - , . , . . , - -

I EVALUATION I) THE COMPUTER HARDWARE AND PROGRAM CHANGES Since the mainframe computer (MC) and its program "VNEW" which were used for the previous verification test have been replaced by the personal computer (PC) with a new program called " RADIAL", EICSB was concerned whether the program RADIAL /PC is equivalent to the old program VNEW/MC regarding analytical techniques and assumptions. At the meeting, we requested a confirmatory analysis (Item 1) using the July 12, 1980 test configuration to demonstrate that there is no appreciable difference in the calculated voltages of the' two (i.e., RADIAL /PC vs. VNEW/MC).

? Our review of the latest TVA comparison analysis also included a Stone & Webster computer program. All three computer programs were run by usir.g an identical set of loads for each board. The summary table I (page 10) of the report showed the following:

BOARD VNEW RADIAL STONE & WEBSTER 6.9kv Start BUS A 7152, 7151 7148 6.9kv Start BUS B 7011 7008 7005 6.9kv Unit Bd IB 7011 7008 7005 6.9kv Shutdown Bd 1A-A 7004 7002 6998 480v Shutdown Bd 1Al-A 495 495 495 480v RX Vent Bd 1A-A 483 483 Not Conducted

t Our review of the results of these analyses finds that the voltage values obtained from the three computer programs show no appreciable voltage differences. This indicates that both computer programs' analytical techniques and assumptions are equivalent for the steady state. However, this comparison requires verification of the steady state and transient response characteristics of the two computer programs. This test result failed to demonstrate the transient response and the steady state at 120/208 volt level. Therefore, the replacement of the old computer program with the new computer program appears to be verified for the steady state case only down to the 480 volt level.

II) CHANGE OUT OF 100 VALVE MOTORS:

We were concerned that the replacement of 100 valve motors with the motors of different electrical characteristics may affect the plant steady state load making it necessary for the new system loadings to be re-analyzed. However, TVA indicated that this change out will only affect the transient loading and voltage whereas the steady state load remains the same. Therefore, we find that this change out of 100 valve motors represents no overall load increase for the steady state condition.

III) ADDITION OF TWO START BUSES AND ONE COMMON STATION SERVICE TRANSFORMER:

TVA explained that the two additional start buses were not actually added, but instead the original two start buses were split into four buses, thus adding no new loads or impedances. Although the third common station service transformer

g has been added, the circuit breakers are normally open making the transformer available as a backup for either of the other station service transformers. TVA demonstrated that this change has little effect on the overall Sequoyah Auxiliary Power System configuration by performing a comparison of the voltage analyses (Item 2) between the 1980 (two start buses) and 1986 (four start buses) configu-rations. The comparison was performed using the July 12 & 16, 1980 test data and the new computer program. The results were shown in the Summary Table I and II (page 10) of the report. We summarize them as follows:

TEST I TEST II (July 12, 1980) (July 16, 1980)

CONFIGURATION CONFIGURATION BOARD 1980 1986 1980 1986 6.9kv Start BUS A 7154 7156 7045 7041 6.9kv Start BUS B 7051 7045 7067 7062 6.9kv Unit Bd IB 7051 7045 7067 7062 6.9kv Shutdown Bd 1A-A 7044 7038 7060 7055 480v Shutdown Bd 1Al-A 501 500 501 501 480v RX Vent Bd 1A-A 493 500 494 501 Start of ERCW pp (Term. V) Not Conducted 6705 6695 Start of Aux. BLG EXH FAN 1A Not Conducted 495 458 We find that the above analyses (I & II) show no appreciable voltage difference (max.

1.5%) between the 1980 and 1986 configurations. Therefore, this indicates that the new configuration has not changed the old electrical system configuration significantly.

I IV) REANALYSIS OF THE 1980 VERIFICATION TEST RESULTS As for the actual verification tests performed at Sequoyah in 1980, TVA, in their latest response, explained how the circuit breakers were aligned for each test' configuration (Item 3).

Our review of the test procedure found that a comparison was made between the calculated board voltages, based on load values derived from breaker alignment and the supply voltages, and the board voltages obtained from the tests. We found that this procedure deviated from our Branch Technical position PSB-1 t

. (Part B.4) which requires measurements of the loads and voltages for a given test configuration and then, using the measured load values on each board as input to the computer model to calculate the voltages. Subsequently, the analytically derived voltage values and the test results are compared. To be consistent with our present PSB-1 procedure, during our meeting on April 16, 1986, we requested new analyses be performed by using the load values obtained during the tests as input to the new computer program. The results are as follows:

TEST I TEST II (July 12, 1980) (July 16, 1980)

BOARD MEASURED ANALYSIS  % MEASURED ANALYSIS  %

6.9kv Start BUS A 7200 7154 0.6 7000 7045 0.6 6.9kv Start BUS B 7000 7051 0.7 7000 7067 1.0 6.9kv Unit Bd IB 7100 7051 0.7 7090 7067 0.3 6.9kv Shutdown Bd 1A-A 7000 7044 0.6 7100 7060 0.6 480v Shutdown Bd 1Al-A 495 501 1.2 500 501 0.2 480v RX Vent Bd 1A-A 484 493 1.9 489 494 1.0 Start of ERCW pp (Term V) Not Conducted 6787 6705 1.2 Start of AUX. BLG EXH FAN 1A Not Conducted 466 459 1.5

r TVA used board meters, test meters and brush recorders for taking test measurements. However, due to calibration problems the brush recorder failed to yield consistent results. As a result, TVA's latest response did not include the measured values obtained from the brush recorder. It is also indicated that CT and PT inaccuracies were present. Furthermore, the allowed accuracy limit was not specified in Position 4 of our FSAR Question 8.33. TVA stated that all measure-ments were taken by the board meters whose accuracy was limited to 5%. Therefore, a 5% tolerance was established as the maximum acceptable difference between the measured voltag'es and calculated vn1tages.

V) BRANCH TECHNICAL POSITION PSB-1 (PART B.4)

TVA performed the 1980 verification tests at Sequoyah in response to FSAR Question 8.33. This question was a precursor to Branch Technical Position PSB-1 (Part B.4) which was issued as part of the SRP in July 1981. Part B.4 of PSB-1 provides detailed guidance for performance of their verification tests. Although Question 8.33 does not explicitly include all of the guidance of Part B.4 of PSB-1, it does so by clear implication. Therefore, our evaluation of the 1980 tests was based on establishing a correlation between these tests and the testing and expected test results specified in Part B.4 of PSB-1. In our April 16, 1986 meeting, TVA i concurred that the intent of position 4 of Question 8.33 requirement is the same as Part B.4 of PSB-1 requirements, even though the PSB-1 requirements are more specific.

t Part B.4 of PSB-1 states the following:

a) loading the station distribution buses, including all Class IE buses down to the 120/208 y level, to at least 30%;

b) recording the existing grid and Class 1E bus voltages and bus loading down to the 120/208 volt level at steady state conditions and during the starting of both a large Class IE and non-Class IE motor (not concurrently);

~

Note: to minimize the number of instrumented locations, (recorders) during the motor starting transient tests, the bus voltages and loading need only be recorded on that string of buses which previously showed the lowest analyzed voltages.

c) using the analytical techniques and assumptions of the previous voltage analyses and the measured existing grid voltage and bus loading conditions recorded during conduct of the test, calculate new set of voltages for all the Class 1E buses down to the 120/208 volt level; d) compare the analytically derived voltage values against the test resul ts .

, _ _ . - - _ ., .r.

_9 With good correlation (within 30 between the analytical results and the test results, the validity of the mathematical model used in the voltage analysis is established.

However, the above procedure involves testing of both the steady state and transient response characteristics. The transient testing requires starting both a large class IE and non-class IE motor. TVA stated that " Voltages were measured over the same path and at the motor terminals of an emergency raw cooling water (ERCW) pump motor (700hp) while it was started and at the terminals of an auxiliary building general exhaust fan motor (150hp) while it was started." TVA, further, stated that " Starting

_ motor voltage dip values are not reliably established by test because of limitations

~

in the instrument capabilities. In both cases, the dip was only for approximately one cycle."

The intent of such a transient test requirement is to detect potential spurious load shedding or separation of class 1E system from offsite power when a large motor is started. The ability of the computer model to predict the effects of the motor transient in the system is verified by comparing the data measured during the transient test with the computer predicted transient values. Upon completion of both the steady state and transient analyses, the validity of the mathematical model is verified. ,

-- - _ _ , - -____,__,,__.___m _ _ , , . _ _ __,_,,._,_,,__7 _ , , _ _ _ ,_ _ . _ , ,

, FINDINGS For the steady state aspect of the test, EICSB finds that:

1. The new configuration has not affected the overall voltage profile of the 6.9kv boards.

2. The change out of 100 valve motors represent no overall load increase.

3. The replacement of the old computer program with the new computer program appears to be acceptable.

4. Despite the fact that our Position 4 of Question 8.33 contained no specific accuracy requirement (applies to Sequoyah) and the measure-ment accuracy of Sequoyah metering was 5%, the consistency of the results between the analyses and test values (within 2%) show that the model consistently predicts steady state system performance.

5. No test and analyses were performed down to 120/208 volt level where the ability of the class IE control circuit to pick up the control devices such as starter, relay, and solenoid is determined.

With exception of item 5, there is reasonable assurance that TVA's new computer program can adequately predict Sequoyah electrical system steady state response characteristics.

l As for the transient aspect of the test, TVA started one ERCW pump motor (700hp) l and an auxiliary building exhaust fan motor (150hp); we find that:

• l

f ' ~

o ^ - ,

1. ~

>, 1. The test report indicated instrument recording problems such that starting motor dip values were not reliably established (i.e., no transient data for the motor and the class IE buses). -

2. The selected motor sizes (700hp & 150hp) were not large enough to show any significant transient effect (the dip was only for one cycle).

The Branch Technical position PSB-1 (part B.4) requires starting of both a large class 1E and a large non-class IE motor (not concurrently)

3. No transient voltage analysis was performed by comparing results of calculations performed by the new computer program with the data obtained during the starting of large motors.

In the absence of the transient analysis, we cannot determine that TVA's new computer program can adequately predict Sequoyah's electrical system transient response characteristics.

CONCLUSION Based on our findings, we conclude ~ that: (1) TVA's new computer program can adequately predict the steady state aspect of the Sequoyah electrical system.

However, this is conditioned on either additional testing and analyses being performed to demonstrate acceptable system response down to the 120/208 volt level or providing proper justification. (2) There is insufficient bases to conclude that the computer model can accurately predict the electrical system response during transients such as starting large motors. (3)TVAshould demonstrate new computer program's capability to predict the transient response of the Sequoyah electrical system.

Therefore, EICSB reconsnends that TVA perform additional testing in accordance with Branch Technical Position PSB-1 (part B.4) as described above.

__ .