ML20082J036

Responds to NRC Request for Info Re Tech Spec Change Request 210 for Outlining Administrative Procedures or Guidelines Established to Assist Plant Operators in Translating RMA5 Trends Into Indications of Significent Changes

ML20082J036
Person / Time
Site:	Three Mile Island
Issue date:	08/22/1991
From:	Broughton T GENERAL PUBLIC UTILITIES CORP.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
C311-91-2098, NUDOCS 9108270111
Download: ML20082J036 (6)

Text

'

GPU Nuclear Corporation tl Nuclear i  :::,nnereo

%ddletown, Pennsylvania 170$7 0191 717 944 7621 TELEX 84 2386 Writer's Direct Dial Number:

August 22, 1991 717-948-8005 C311-91-2098 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, DC 20555

Subject:

Three Mile Island Nuclear Station, Unit 1 (1M1-1)

Operating License N0. DPR-50 Docket No. 50-289 Amendment No. 163 - OTSG Leak Rate Limit Response to NRC Request for Information During a telephone conference on August 9,1991 regarding Technical Specification Change Request (TSCR) 210, the NRC staff requested information outlining the administrative procedures or guidelines that have been established to assist plant operators in translating RMA5 trends into

, indications of significant changes in primary to secondary leakage rate such l that a reactor shutdown would be warranted. Amendment 163 requested that this t

information be provided by letter within two weeks of the issuance of the amendment.

Attached are two memoranda dated February 22, 1991 and August 1, 1991. These guidance documents are those which were telecopied to the NRC and discussed with the NRC Project Manager on August 9, 1991. The limits on primary to secondary leakrate stated in the August 1, 1991 memorandum are the limits which existed prior to the issuance of Amendment 163. This guidance has been l reissued to reflect the new limits. Also, it is noteworthy that the procedure

! changes proposed in section V. of the February 22, 1991 memorandum have been l incorporated.

Sincerely, co m

• Y .,

38 T.G.Brok,hton o4 Vice President and Director, TMI-l 1 g MRK

-o OQ Attachment c4 l $z cc: NRC Region I Administrator

& TMI Senior Resident Inspector THI-l Senior Project Manager .c l g GPU Nuclear Corporation is a subscary of Generai Pubhc Ut.ht.es Corporation 1

t

. Shift Suporvisors .

. ' August is 1991 r 3210-91-0236- +

Page i Z. TITLEt ,

RCS PRI-SEC Leakrate Validation ZZ. PURPOSE Provide on-shift crews with guidance on the performance of validations of primary to secondary leaktates in excess of the license condition j limit.

III. GENERAL DESCRIPTION:.

Procedure 1301-1 (Shift and Daily Checks) provides guidance on the  ;

calculation of primary to secondary leaktate as well as defining the actions to be taken in the event the license condition limit was ,

exceeded. The previous revision to the procedure indicated that the

• plant had 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to validate leaktates identified to be in excess of -

the lics.se condition limit. This has been modified to provide up tc 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> to perform the validation, as long as the leaktate was not  ;

calculated to be more than 0.1 gpm (6 gph) above the limit (12 gph total). If the leskrate was calculated to be equal to or greater than .

0.1 gpm above the limit, the validation period would be limited to I 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />. It should be noted that the current license condition limit .

remains the controlling condition for operation and that if that  :

condition is clearly exceeded the plant shall be shut down.

The difference in validation periods recognizes the idea that the higher a value is above the limit e the more confidence we should have that the limit has actually been exceeded. Not providing specific procedure steps for the validation process recognizes the individual nature of each OTSG 1eakage event. The current leak is increasing slowly and the calculated values vary by more than 1 15% about the mean. It is '

estimated that it may take from 1 to 3 weeks for the average value to exceed the license condition limit after the first single value exceeds ,

the limit. In contrast to this condition, the event in the beginning of ,

cycle 8 took less than an hour to exceed the limit from the first indication.

In the future the following guidance should be used when validating. ,

primary to secondary leaktates calculated by "Heth B* to be in excess of the license condition, when the Shift Supervisor does agg feel an urgent need to shut down the unit because of leakage or other circumstancess

1. With plant conditions stable perform at least a one (1) hour computer leakrate. Training Handout 3210-91-0060 (attached) provides additional guidance on leakrate determination.

2. Determine if the calculated " Method B" value is consistont with I other available information (i.e. , increasing RMA-5/RRA-15 values, .

the computer'leakrate above and a comparison of previous sets of input data to the calculation).

3. Determine if there are any sampling or analysis problems that could have lead to the initial high value.

i 1

[

~ ' Shiie Supervisors

, August 1, 1991 3210-91-0236 Page 2

4. Schedule additional sets of PCs and condenser offgas samples (2 sets per shif t is a reasonable rate over the validation period) to provide additional calculated values (Method B) to use in a statistical evaluation of the data.

5. Although there is no set ncthod to be used in determining if the limit is exceeded, the following should be consideredt

a. How much variation is there in the data?

b. Is the arithmetic mean of the data obtained during the validation period greater than the limitt

c. Is there an in :reasing trend in the calculated leakrate valuest

d. Are three (3) or more values obtained in succession above the lim 1tt

e. How much are the calculated values above the limit?

IV. COMPONE'NTS t N/A V. OPERATIONt N/A VI. ALARMSt N/A VII. INTERFACING SYSTEMSt N/A VIII. ADDITIONAL INFORMATION:

N/A IX. OPM EEMNbHt N/A

shift Superv1 sors

' ?ebruary 2:,' 1991 1110-91-0060

? age ;

.  ::::Et RCS FRI-SEC Leakrate Determination

. PCf PPOS E t Prov1de on-shift crews with gu1 dance on how to quantify and respond to RCS PRZ-SEC leakage.

Z. GENERAL DESCRIPTIONS Recently the PRO has completed an evaluation to determine what methods are ava11able to the on-shift operators to quant 1fy a small RCS PRI-SEC leak (i.e., one which is close to or exceeding the i gpm threshold) similar to the one which occacred on March 6, 1990. During the leak which occurred on 3/6/90, the ACS activ1cy had not yet reached the plateau for perform 1ng a method "A" type PRZ-SEC leaktate determination.

n order to deter 1ne the s1:e of the leak, several people were performing ' hand" calculan1ons using a mass balance approach. Based on the results of the " hand" calculations, the PRG has determined that these calculations are not accurate enough to use as a bas 1s for determ1nang 1! a plant shutdown and/or an E-Plav declaration are required based on exceeding one '), gym PRI-SEC leakage.

In the future, the following guiuance st uld be followed when trying to quantify an RCS PRI-SEC leakt

1. Once at has teen determined that a small RCS PRI-SEC leak has occurred (computer po1nt C4042 and appropriate RMS tcadings would g1ve the first indication of increased leakage) the operators should maintain the plant at steady-state conditions and perform at least a one (1) hour computer leakrate. The Shift Supervisor and Operations Management would then use the results of the computer leakrate to determine 1! any E-Plan act1on and/or a plant shutdown should be initiated. After the leakrace is completed, samples of the RCS and off gas activity should also be obtained for follow-up analysts of leak size.

2.  :! the Shift Superv1sor feels an urgent need to shut down the unit because the leak appears bad or there are other c1rcumstances which would prohibit the performance of a i hour (min 1 mum) computer leakrate, the supervisor would automac1cally take the E-Plan action appropria te for a leakra te of >l gpm. Additionally, grab samples of the RCS and off gas activ1ty should also be obta1ned !ct follow-up analysis of leak s12e. This gu1 dance w111 simplify the dec1sion to make an E-Plan declarat1on and will also insure a conservative approach is used in making the E-Plan declaration.

Shift Superv1 sors

,. Pebruary 22, 991 3210-?l-0060

? age 1 The ctmputerzoed method of determin1ng the method *A" PRI-SEC leaktate (computer point C4042) is an accurate method when the RCS activ1cy 1s maintained in a stable cond1 tion. Any plant upset or power changes which could cause fluctuations in RCS act1v1ty will cause misleading readings on computer po1nt C4042. Therefore, the computer point (C4042) snould mainly be used for event recogn1 tion and trending, and would st111 need a backup validation by the computer mass balance leakrate calculat1on.

Once a plant shutdown is initiated, any change in the PRE-SEC leaktate could be masked by the additional makeup required to compensate for the RCS cooldown (i.e., cooling Tavy down 1*T requires 95 gal, of makeup). Therefore, 1t may be difficult to determine if a leak that was <$0 gpm increases to >$0 gpm which would requ1:e an escalation in the E-Plan class 1!ication. The only method ava11able to the crew in this situation 1s the ' hand

• mass balance calculation. 'he person do1ng the mass talance must be sensat1ve to the correlation between the apparent leaktate and the RCS cooldown race. :n order to standardite the method to be used for perf orming a " hand

• mass balance calculation the following guidelines should be followedt

a. The longer the czme interval between collect ng data points, the more reliable the leaktate estimate w111 be. Therefore, leakrates should not be calculated for a time interval of

<3 minutes .

b. :n order to collect data points at the same time, the required plant parameters should be put on the same computer group so they can be printed out and then used in the calculation.

c. A standardized formula should be used. The following formula is the preferred one to uset it.P L *!!2 call"' ~  !?.M T L *t30 call *? -

'hTava *?)*'95 call *?)

(TIME minutes)

= gal / minutes leakrate b = Differential as determined by (initial value - final value)

PL = Pressuri:er Level (Computer Point C4017)

M T L = Makeup Tank Level (Computer Point A4098)

Tavy = RCS Average temperature (Computer Point A5066) l l

29IE: The coefficient of 95 gal /*T for cooldown applies when Tavy is $79'F, to determine the coeffic1ent for other values of l Tavg refer to Page E3-1 of 1210-10.

s

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

- . - - . _ . . . . . . - . -. . . _ . . - . . ~ . .. . - . _ ~ . _ .

e, ..

LShift Superv1scrs February 12, ;391.

' j 2210-91-0060 '

Page 4- '?

e

Y . . COMPOVENTSt -

N/A - [

.7+ OPERAT:OVt The information in this tra1ning handout will be added to the alarm $

respcnse (MAP C-2-1) for RM-A-b. Additionallyr OP 110$-10A will be' i revrsed to refer the cperator to the alara response for RM-A-5 if I computer point C4042 goes Anto an alarm condition.

}

t VI. M ut t I

There are no new alarms associated with th1s training handout. b VII. INTERFACING SYSTEMSc l i

N/A >

VI:I. ADDITIOVAL INTot b i

PRO meeting minutes !ct meeting #91-006 t'

\ h S

ZX. OPM SECT:0Mt I t

The OPH is not affected.  :

.h 1

I L

t

+

9 I

)

I F

6 i

h 4,

l I I t

1 P

I

. . . _ - _