ML20099E682
| ML20099E682 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 03/13/1981 |
| From: | Lippe L JUSTICE, DEPT. OF |
| To: | Jamarl Cummings NRC OFFICE OF INSPECTOR & AUDITOR (OIA) |
| References | |
| NUDOCS 8411260018 | |
| Download: ML20099E682 (2) | |
Text
f,\
,o .
6 i
.s ,
f, Mj;. UNITED STATES DEPARTMENT OF JUSTICE
- u. * )
=i;-
WASillNGTON. D.C. 20530 6'
Ahe. Repey se the o6..;. r e .a
.a n.s. :.4 4.i. .a n*
LL:F0B:jmm March 13, 1981 Mr. James Cummings Director Office of Inspector and Auditor Nuclear Regulatory Commission
. Washington, D.C. 20555
Dear Mr. Curaings:
I am writing in regard to the ongoing investigation of possible . .
criminal activity at the Three Mile Island Nuclear power station. As -
you of course recall, your office referred this case to the Criminal Division in April 1980, af ter conducting your own investigations into certain allegations by former Control Room Operator Harold Hartman.
When you z eferred the case, you sent to us transcripts of several -
statements made by Mr. Hartman, as well as summaries of interviews with several other CRO's. In addition, of course, you provided us with a variety of test records, logs, and other material essential to under-standing the case.
Attorneys from this Section, in cooperation with the United States Attorney for the Middle District of Pennsylvania, began a grand jury in-ves*.igation during the second week of May 1980. The investigation is continuing. At the time we went before the grand jury, we were led to believe that the NRC had provided us with all the statements pertaining to this matter which had been given to NRC investigators by Metropolitan Edison employees. Many of the decisions about how the investigation should proceed were based on the assumption that we were cognizant of all relevant statements in the possession of the NRC.
7
/'i t
1 8411260010 010313 PDR ADOCK 05000209
- l. P PDR Attachment 17
~ . _ _
V......
- f. . ,
(.
. (, <
. .e
. )
It is with considerable' surprise, and not a little dismay, therefore, that we now find the NRC took eighteen statements from Het Ed employees of which we were not informed until a few weeks ago.- In early February 1931,
- k. Frank Bowman, an attorney with this Section, called Mr. John Sinclair to verify certain facts which were to be included in an affidavit to be presented to the U. S. District Court in Harrisburg, Pennsylvania. One of these facts was the number of Not Ed employees who had been interviewed by the NRC regarding the Hartman allegations. Af ter doing some checking (apparently including a call. to Mr. R. Keith Christopher, an investigator in Region I), Mr. Sinclair reported that in addition to the statements already in our possession, NRC investigators had conducted fourteen
" screening type interviews" with TM1 CRO's and shift foremen, plus four in-depth taped interviews with Mr. Jim Floyd, Mr. Kenneth Hoyt, -
Mr. Bernie Smith, and Mr. Brian Mehler, all supervisory personnel at TML. All eighteen men are potential witnesses or targets of the grand jury investigation. Mr. Sinclair quickly provided us transcripts of the four- taped interviews and copies of notes on tha:results of the fourteen
" screening type interviews." However, the fact remains that these potentially crucial statements did not reach us until ten months after the start of grand jury proceedings.
It is too early to tell whether our lack of awareness of the additional eighteen statements will have any adverse effect on the outcome of the Three Mile Island investigation. Certainly, had we proceeded to trial and not supplied the defendants with pretrial statements of government witnesses *,
which would have been (unbeknownst to the Department of Justice) in the N Covernment's possession, we would have risked reversal of any conviction.
Regardless of the result in this particular case, however, we cannot stress too strongly that omissions of this kind are extremely grave and every effort should be made to avoid a similar future incident. If the Criminal Division is to do a competent and professional job of representing the Nuclear Regulatory Commission in those criminal matters you refer to us, we must be in possession of all relevant information in the NRC's control.
Very,truly yours,
' //
LAWRENCE LIPPE, Chief General Litigation and Legal Advice Section Crimins1 Division O
. ../...
p,
{ UNITED STATES. b NUCLEAR REGULATORY COMMISSION - M. J f , WASMNeGTON, o. C.30505 4
~$
Mt
=....
- L NOV 'l 1983
- - i
! I i '
t MEMORANDUM FOR: William J. Dircks i i: Executive Director for Operations l
3 FROM: Richard C. DeYoung, Director l i Office of Inspection and Enforcement :
1 i j
SUBJECT:
EXAMINATION OF REACTOR COOLANT SYSTEM LEAK RATE TESTING AT ,
i RANCHO SECO AND DAVIS-BESSE j i
Your memorandum of September 20, 1983 directed the Office of Inspection and .
Enforcement to review the reactor coolant system (RCS) leak rate test procedures !
j and calculational methods in use in 1978 at two B&W plants. Any significant !
I deficiencies were to be identified and compared to the previously identified !
deficiencies in the TMI procedures and calculational methods. ,
t
! Rancho Seco and Davis-Besse were selected at random for the review. Records !
- and procedures relating to primary coolant leak rate tests for a ont-month +
1 period of stable operation in 1978 were examined. The objectives were to .
establish whether these plants could. adequately determine the leak rates to l within the applicable limits and whether the problems previously identified :
at TMI were unique to TMI or were common to other B&W plants. Independent ,
calculations were performed with the NRC-developed leak rate computer program f
. to assess measured leakage and evaluate methods. l The details of our evaluation are presented in Enclosure 1. Our conclusions, based on this evaluation are that the deficiencies identified in the Davis-
- Besse and Rancho Seco plants are relatively minor and resulted in only small !
errors in the calculation of RCS leak rate. There were no inadequacies in the ;
Rancho Seco and Davis-Besse procedures sufficient to provide a motivation for ,
operator falsification of the leak rate test records, as apparent 1/ was the t case at TMI. No evidence of any falsification of records was identified at ;
either Rancho Seco or Davis-Besse.
Your memorandum also requested that IE provide a summary of the RCS measurement f program initiated by IE after the accident at THI-2. This is provided in Enclo-sure 2. The results of the trial use of the program by the regional inspectors ,
have not disclosed any indications of falsification of the test.results by licen-sees. However, these inspections were not specifically structured to uncover
- falsification of data. A number of minor errors and inaccuracies of the type
- identified at Rancho Seco and Davis-Besse have been identified. In one case,. .
l l
. .m 7______ _ _ . _ _ ___
[
. .. .. . . ...k NOV 71983 2
because of rounding of input data, the calculations were not sufficiently accurate to provide reliable information on unidentified leakage. Deficiencies at oper-ating plants uncovered during the trial use period have been corrected, f :: =?
Richard C. oung, rector Office of pection and Enforcement Enclosure's: As stated cc: V. Stello, OEDO J. Roe, OEDO J. Taylor. IE E. Jordan, IE R. Baer, IE L. Cunningham, IE R. Woodruff, IE
D. Kirkpatrick, IE .
H. Denton, NRR B. Hayes. 01 J. Axelrad. IE T. Murley, RI
- J. O'Reilly, RII -
J. Keppler, RIII J. Collins, RIV J. Martin, RV en .
B 4
i
Enclosure 1 EXAMINAT10it OF REACTOR COOLANT SYSTEM LEAK RATE TESTING AT RANCHO SECO AND DAVIS-BESSE The performance of the primary coolant leak rate tests at Rancho Seco and Davis-Besse during the period shortly prior to the TMI accident was examined. The objective of the examination was to determine if the leak rate test problems previously identified at TMI were unique to TMI or if they were comon to other B&W plants. The two plants were chosen at random from among the B&W plants in operation prior to 1979. The reactor coolant system (RCS) leak rate test pro-cedures and calculational methods used during that period were reviewed to find out if they could adequately determine the leak rates to within the applicable limits. The surveillance test records during a one-month period of stable power operation in 1978 were selected for review for each plant. Those deficiencies that were identified were compared to deficiencies previously identified at TMI.
The licensee calculations were reviewed. Independent calculation:, were performed These results agreed closely with the KRC-developed leak rate computer program.
with the licensee's calculations.
- 1. PurposeandDescriptionofk.eakRateTestProcedure The reactor coolent system inventory balance procedure is designed to measure the RCS leakage, it is a surveillance procedure that is required by the .
Technical Specifications of essentially all nuclear power plants. The' applicable limits for PWR plants are typically zero for throughwall leakage
- in vessels or piping,10 gpm for identified leakage from seals and valves, and 1 gpm for unidentified leakage. The required frequency of testing is at least once every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />. Since the unidentified leakage can be converted to identified letkage at any time by locating and quantifying its source, the 1 ppm limit can be considered to be the allowable tolerance limit for the entire surveillance test procedure. This limit is used in judging the signifi-cance of the combined errors and inaccuracies of the procedures evaluated below. The procedure involves determining an RCS water inventory at the beginning and at the end of a test interval. The net change in the inventory is the end of a test interval. The not change in the inventory is then used to determine the gross leak rate. The inventory should account for water mass changes due to level changes in the pressurizer and makeup tank (volume control tank). It should account for density changes in the RCS and pressurizer (bothwaterandsteamspace)fromchangesintemperatureandpressure. It should also account for any RCS water addition or removal made during the test interval. The identified leakage is determined by measuring the level changes in RCS leakage collection tanks such as the pressurizer relief tank (alsocalledthequenchtankorreactorcoolantdraintank). Any other RCS allowable leakage that can be quantified is included in the identified leakage.
The unidentified leakage is then determined by subtracting identified from .
gross leakage.
p , .
2 Enclosure 1 2.. Ctissions and Errors in the TMI Procedure The TMI leak rate test computer program which was used before the ac-cident for all of the leak rate calculations had at least six significant errors. Some were large enough to practically ensure that an unacceptable unidentified leak rate would be calculated even if there were no actual unidentified leakage. These are summarized in Tables 1 and 2 ar.d detailed in the following paragraphs. -
- a. Omissions at TMI The TMI procedere neglected the cffeet of prc::urc cher.;:: cn the reactor coolant system as well as temperature changes in the pres-surizer.
each other.TheseThetwo items error resulting haveisopposing effects,i and has littletending to can on y 4 lb/ps effect at most plants. However, because of a feedwater control problem, TMI was experiencing large pressure variations during much of the time. These averaged about 20 psi, resulting in an error of 80 lb per test. TMI used a reactor coolant weight of 5.84 lb/ gal in calculating its leak rates, resulting in an average error (from neglecting pressure) of about 0.2 gpa per test. As is discussed below, this was not very significant. compared with other errors in the TMI procedure. -
- b. Errors in calculations at TMI Inconsistent densities were used to convert mass of water to gallons of leakage. The gross leakage from the RCS was determined by summing '
the mass changes (calculated in pounds) in the various primary spaces and multiplying by a gallons-per pound factor, based on the water density at RCS temperatures (5.86 lb/ gal at 582*F). The identified 1eakage, however, was derived from the leakage collection tank level change converted to gallons by use of a table in the computer. The calibration for this level measurement was based on cold water density (8.29 lb/ gal at 70*F). Since the unidentified leakage is ' defined as
- the difference between gross and identified leakage, this inconsistency 1eads to an erroneous increase in the unidentified leak rate of about 40% of the identified leak rate. Before March 16, 1979 (when hand-calculat " corrections for this item were begun) the identified leak rate was averaging 3 gps. This resulted in the calculation of an un-identified leak rate which was too large be 12 ppe. .
There was a similar failure to correct the volume of water added by the operators to the RC5 for expansion to reactor density. This omission results in an erroneous decrease in the unidentified leak rate of the same magnitude. During March 1979, 15 water additions, averaging 250 gallons each, were made during leak rate tests, resulting in an error of 1.7 spm. The majority of these additions were made after March 16, 1979, when the hand-calculated corrections for the first item were begun. However, no corrections for the density of the water additions were made.
V. , .
?
3 Enclosure 1 l The tables in the program used to convert temperature to density terminate at 582*F. When the RCS temperature exceeds this value, the density corresponding to 582*F is selected. Twenty-two of the tests reviewed had temperatures above 582'F and resulting errors averaging 0.2 gpm.
An incorrect RC5 volume was used in the calculation 3of the mass change ,
in the RC5. The computer 3used a value of 10,673 ft whereas the SAR ave a value of 10,346 ft . For the average temperature variation of
.2'F this caused an error of 0.013 gpm.. :
i lhe table in the computer memory used to convert reactor. coolant drain i tank (RCDT) levels to gallons of water differed from the equivalent i table used by the operators in the control room. As an example, for an RCDT level of 76 in., the table in the computer memory gave a value of 6,605 allons whereas, the value used in the hand calculation was 6,411 gal ons. Yhe average drain tank change was 2 in, causing an error ,
cf 0.14 gpm.
- 3. Evaluation of Davis-lesse i The evaluation of Davis-lesse included a review of the procedures in ,
existence during 197g, an evaluation of the leak rate calculations .
prformed by the licensee during that period of time, and the performa'nce of independent calculations using the NRC computer program.
- a. Davis-lesse Procedures The Davis-lesse procedures were reviewed to determine to what extent '
- they accounted for those items necessary to make an accurate water
. inventory determination. The results are sunnarized in Table 1. The i procedures accounted for level changes in the pressurizer and makeup l tank. It also accounted for density changes in both the RC5 and the pressurizer due to temperature variation. Adding or remov ng water' ,
during the test was prohibited. As discussed in the fo11 ing paragraphs, .
three items needed for accurate leak rate determinations were omitted from the Davis-lesse procedure. None of these proved to be very signifi-cant under the usual conditions of the test.
The. Davis-lesse procedure omitted the correction for the effect of i pressure on the RC5 causing an error of about 10 lb/ psi. However, the l average pressure variation recorded on the surveillance test sheets was !
only 0.3 psi. Also, Davis-lesse used a relatively long test interval
- averaging two hours. The resulting error on the average test was a ,
negligible 0.003 gpm. l The Davis-tesse procedure also neglected the mass change in the pres- !
surizer steam space which occurs witg pressurizer leve change. Steam I
has a significant density of 6 lb/ft at operating pressure. The
!i i i
I ,-
i -l
i
~4 . Enclosure 1 -
f steam volume changes by 3.173ft per inch of level change. The average level change was 1.4 in, resulting in an error of about 27 lb or 0.03 gpm for a two-hour test..
Davis-Sesse did not,use the pressurizer relief tank (PRT) level changes -
in calculating its identified leakage. The main component of Davis-Besse's identified leakage is pump seal leakoff, which was not collected in the PRT but went to the sump (where it mixed with non-reactor-coolant
. water). This identified leakage was measured manually at each reactor -
shutdown and used unchanged in the leak rate calculation until the next
- shutdown. This is conservative with respect to the unidentified leak l
. rate, since the seal leakoff tends to-increase, but it does introduce some error in the calculation. The average increase in the seal leakoff measurement, following a shutdown, was 0.15 gpm. Assuming that the 3 i
seal leakoff increased linearly. -the average error caused by this *
- practice would have been 0.075 gpm.
i
- b. Davis-Besse Leak R, ate Calculations The calculations used with the Davis-Sesse procedure were reviewed to 1 I determine if correct values and methodology were used. The results are i shown in Table 2. The calculations at Davis-Besse were normally done-
! by a cceputer after the test parameters had been entered manus 11y by *,
j -
the cparators. The exact version of the computer program used at Davis-i Sesse in 1978 could not be obtained. However, manual calculation sheets
! were provided by the procedure. Also, the computer printout of the test res'ults provided water mass quantities for each of the items a' included in the inventory, so it could be shown that the computer
,l prcvided exactly the same results as the manual calculation. i The Davis-Besse procedure calculated all inventory quantities in pounds and summed the results before making the conversion to gallons, so there was no problem in comparing the different quantities. The ASI'E steam table values were used to derive the density of. pressurizer wa te r'. The mal 1 accounted for. jaup tank and Appropriate the pressurizer values were used inlevel changes converting netwere masscorrectly i j changes to leak rates. !!o significant inaccuracies were found in the s Davis-Sesse calculations.
, c. Calculation of Davis-Resse Leak Rates Using the NRC Program <
1
! . A one-month period of stable power operation at Davis-Besse was selected
- and copies were made of the surveillance test records for-the entire '
i month. The surveillance test input values were used to run leak rate j test calculations using the l'RC Computer Program (described in HUREG-0986).
! These calculations were made for the entire month. The results are shown l in Table 3. These results are compared with the results calculated by I the Davis-Besse procedure in Table 4. These two sets of results agree l quite well. In only 2 cases out of the 29 results calculated did the
- results differ by more than 0.2 gpm. The average difference in the two sets of results was about 0.1 gpm.
' . i 0
,,,,,ne,--m. _nm,.-nn,wc.,e.-,,wm,-,,,n.. ,n,,4n__,n__ _,nn,.,,,r,,,,n,,,-. --~.-
5 Enclosure 1
- 4. Evaluation of Rancho Seco Like the evaluation of Davis-Besse, the Rancho Seco evaluation included a review of the licensee's procedures and calculations during 1978 and independent calculations using the NRC Computer Program,
- a. Rancho Seco Procedures Rancho Seco actually used two procedures, a complete test done weekly and an abbreviated test done daily. The daily procedure required that the more comprehensive test be run immediately if the daily test pro-l duced an unacceptable result. As shown in Table 1, both of these
> p ocedures accounted for level changes in the pressurizer and makeup tank. In both cases adding or removing water during the test was pro-hibited. The weekly test j and the pressurizer becaus, accounted for e of temperature density However, variation. changes neitherin both the RC of the procedures included corrections'for RCS pressure. None of the items omitted in these procedures had a very significant effect on leak
- rate test resultsv- Each is reviewed briefly in the following sections.
(1) Rancho Seco Comprehensive Test
! The only omission in this test was the lack of a pressure cor- .
I rection for the reactor coolant system density. The effect bf L compressiononthewatermasscontentofthereactorcoglant system is slightly less than 10 lb/ psi for the 10578 ft RCS L at operating pressure. The average pressure variation recorded
- on the surveillance tests reviewed (discussed below) was about
- 3 psi, resulting in an error of 29 lb. This caused an error of about 0.06 gpm in the average one-hour test result.
(2) Rancho Seco Daily Test Procedure The abbreviated Rancho Seco test omits pressurizer temperature as well as pressure corrections. The lack of pressure c*orrection j causes the same 29-1b error in RCS water inventory. However, any 1 pressure rise is caused by an increase in pressurizer temperature,
- which is accompanied by a pressurizer water expansion that results i in a loss of 6 lb/ psi. This is 18 lb for the average 3 psi pressure
- rise. The result is that the omission of a pressure correction only
- has a 0.02 gpm effect on the average test when pressurizer tempera-e ture is also neglected. .
! The daily test also omitted the effect of temperature on the -
RCS density. This causes a larger error of 808 lb/*F. The .
average temperature change recorded on the daily tests was 0.2'F l causing an error of 162 lb and 0.3 gpm on a one-hour test.
9
, , , - - - - - + .
, , . , - - - . - - . - e~w - - . --e---,-.m.
,. --.-.--.%-%---,-~m. c,.,.-.- - -. .-v--- ,,----.,-,.-.-r.
! . * ; ., .K
.I ,
j .
6 Enclosure 1 l V
- b. . Rancho Seco Leak Rate C h:ulations
.Th'e calculations used with the Rancho Seco procedures were also reviewed.
The results are shown in Table.2. The calculations at Rancho Seco were all done manually. The romprehensive procedure calculated all inventory Q" ~
quantities in pounds and summed the results before making the conversion
- , , to gallons. The simplified procedure did carry the various quantities
( in gallons, but appropriate corrections were made to the high temperature r RCS quantities to account for.the reduced density. The ASME steam table _.
' values were used to derive the density of pressurizer water. In both '
of the procedures the makeup tank and the' pressurizer level changes were ;
. correctly accounted for. Appropriate values were used in converting-the net mass changes to leak rates. Two small inaccuracies were identified .
in the Rann o Seco tests.
On[of the deficiencies identified was the use of a constant value of 2
. lb/*F to correct for RCS temperature changes. The calculation i
666.7,'
using the specific. volume equations for compressed water from the ASME
~
steam tables give~s a value of 808 lb/*F. The resulting error of 141 lb/*F' combined with the average temperature variation of 0.2*F resulted in an average deviation of 28 lb or 0.06 gpm per test.'
The other deficiency was the use of a 33.3 gal /in. value to account for -
changes in the PRT level in the Rancho Seco daily test. The value used with the comprehensive test, as well as the currently accepted value is
< 33.8 gal /in. The 0.5 gallon error combined with the average PRT level
- change of 2.6 in. resulted in an estimated error of 0.02 gpm.
- c. Calculation of Rancho Seco Leak Rates Using the NRC Program The surveillance test input data for a one-month period of stable power >
operation at Rancho Seco was also used to run leak rate test calculations using the NRC Computer Program. The results are shown in Table 5 for the-daily tests and Table 6 for the weekly tests. These resul.ts are compared with the results calculated by the Rancho Seco procedure in Tables 7 and 8.
. The 'two sets of results for the weekly test differ by an average of only 0.01 gpm. The two sets for the' daily test also agree quite well in most j cases. However, in 5 cases out of the 31 daily results calculated, the results differed by more than 0.3 gpm. These were all cases in which there was a relatively large temperature change recorded between the '
beginning and end of the test. The inaccuracy was due to the omission of an RCS temperature correction by this procedure. The reviewe. was able to include the temperatures in the NRC calcolations because these were included on the test sheets "for information only." In two cases j the temperature variation was about l'F., resulting in an unidentified i leak rate which exceeded the one gpm limit. Since the temperatures were
- not normally used in the calculation, it is possible that the operators may not always have recorded them with sufficient accuracy for this purpose. However, the two sets for the daily test still agree to within an average of about 0.2 gpm.
i .
? I
. l
7 Enclosure 1
- 5. Conclusions The Rancho Seco and Davis-Besse RCS leak rate test procedures and calcu-lation~al methods used in 1978 were.sufficiently comprehensive and accurate to determine the leak rates to within the applicable limits. The omissions and inaccuracies in the Rancho Seco weekly test and Davis-Besse test resulted in estimated average errors of less than 0.1 gpm. The estimated average error in the Rancho Seco daily test was about 0.3 gpm. These estimates were confirmed by the results of the NRC computer calculations which used the same input data as the licensees used. The errors identified in these procedures are at least on an order of magnitude less than the errors previously identified in the TMI procedures. There were no inadequacies in the Rancho Seco and Davis-Besse procedures sufficient to provide a motivation for the operator to falsify the leak rate test records, as apparently was the case at TMI. No evidence.of any falsification of records was identified at either Rancho Seco or Davis-Besse.
e me e
e '
l I
I
_ ~ _ . ..._ .
~
m . ,. .
. _ l i
Enclosure 2 PROGRAM FOR INDEPENDENT MEASUREMENT OF RCS LEAK R'ATES BY NRC INSPECTORS Following the IE-investigation of the alleged leak rate test falsification at TMI in March 1980, a program was written to calculate reactor coolant system (RCS) inventory balances using the Hewlett Packard 41C programmable calculator.
The adequacy of this program was verified by conducting leak rate test calculations at the Farley and Calvert Cliffs stations in July and August 1980. On the basis of this work, it was concluded that the expected errors for the licensee's cal-
- culations were 1.9 gpm for TMI 2, 2.3 gpm for Farley 1, and 0.1 gpm for Calvert Cliffs 1 and 2. The Resident Inspector stated that the licensee subsequently took corrective action for Farley 1.
In March 1981 a Temporary Inspection module (TI2512/48) was issued for the independent calculation and verification of RCS leak rates by NRC inspectors using the Hewlett Packard 41C program. All PWRs in Region III were found to have gross leak rates of less than I gpm and there was acceptable agreement between the licensee's' and the inspector's calculations. The program was also appliedtothePWRsinReg1}onVandnoadverseresultswerereported.
l In 1982 an expanded program was writter. for the Osborne portable computer in order to expand the capabilities of the calculations, facilitate data entry, and simplify storage of the plant-specific parameters. Osborne computers were l procured for both IE and the regional offices in early 1983. Regional inspectors .
were trained in the use of this program in April 1983. This program was sub-sequently documented in NUREG-0986 and is currently available for the NRC independent measurements program.
A permanent inspection module for the IE manual has been written, providing for regular (an annual frequency has been proposed) NRC verification of the adequacy of the licensees' RCS leak rate test procedures. This module is presently being processed by the Division of Quality Assurance, Safeguards and Inspection
' Programs.
At this time, there has been no documented feedback on the use of this program by the Regional Inspectors.
. . . , - . - - ~ _ , _ - . . _ _ _ _ - _ - _ - . , . . . , . - , _ _ . - . - , . _ . , , , - . , , _ . , ,
TABLE 1 .
CHANGES ACCOUNTED FOR BY PROCEDURES -
IMI Davis-Desse Rancho Seco (Weekly) Rancho Seco (Daily)
Avg. Error l
Avg. Error Avg. Error Avg. Error
- cceter Included (gpm) Included (gpm) Included (gpm) Included (gpm)
!i dznsity - temperature yes --
yes --
yes --
no 0.30 i , density pressure
- no 0.60 no 0.003 no 0.06 no 0.06 t
rssurizer water density
- no -0.40 'yes --
yes --
no -0.04 yes no 0.03 yes --
yes --
- !ssurizer steam volume yes yes --
. yes --
yes --
- rssurizar level --
- (e'up tank level yes --
yes --
yes --
yes --
yes no 0.075 yes --
yes --
lief tank level --
- ter cddition or loss yes --
prohibited -- prohibited --
prohibited --
2 hese two items tend to cancel each other because RCS pressure increase is caused by a ressurizer temperature rise resulting in reduced pressurizer densi ty.
?
l TABLE 2' .
! CORRECTNESS OF VALUES USED IN LEAK RATE CALCULATIONS IMI Davis-llesse Ranclio~Seco (Weekly) Rancho Seco (Daily)
Avg. Error Avg.Lrror Avg. Error Avg. Error Correct (gpm). Correct (gpm) Correct (gpm) Correct (gpm) .
,. sme ter no 0.013 yes --
yes --
yes --
volume 0.14* yes --
no 0.06 no 0.06 d:nsi ty no yes ye's --
no -0.04
- ssurizer water density -- --
yes .
ssurizer steam density yes --
yes --
yes --
yes --
ssurizer volume vs 1cvel yes --
yes --
yes --
yes --
yes yes --
yes --
yes -- .
icup tank mass vs level --
yes yes --
yes --
yes --
. cup tank density --
ior tank mass vs level yes --
no 0.02 drcin tank) no 0.14 yes --
lef tank density no 1.2 yes --
yes --
yes --
isity of water added prohibited N/A er Icst no 1.7 N/A prohibited N/A prohibited
-ivarsion of mass to .
8.33**
gallens yes 5.86** yes 8.25** yes 8.33** yes lbsva 582'F.
. .b/gs1. .
l
TABLE 3 P
/
REACTOR COOLANT SYSTEM LEAK RATES Davis Besse Unit 1 RCS AVG FZR MUT PRT LEAK RATES PRESS TEMP LEVEL LEVEL LEVEL (- - - gom - - ->
(psig) (F) (in) (in) (in) GROSS IDENT UNIDENT MTE START START START START JR(Hr) END END END END 3/30/78 2215 582.2 203.5 48.66 0 1 2210 582.1 201.1 68.3 0 0.55 0.29 0.26 8/S1/78 2210 581 .6 206.3 73.8 0 1- 2208 581.6 206.1 72.21 0' O.88 0.37 0.51 9/1/78 2211 581.7 210.7 70.83 0 1.05 2207 581.7 211 68.59 0 1.09 0.37 0.72 9/2/78 2207 581.4 208.8 68.26 0 3.28 2213 581.6 210. 3 63.71 0 0.71 0.37 0.34 9/3/78 2215 582.1 212.2 74.65 0 4.433 2211 582.1 211.7 68.79 0 0.72 0.37 0.35
?/4/78 2215 581.7 210.8 71.84 0 .
1.E66 22:5 581.7 211.8 69.'09 0 0.66 0.37 0.29 9/5/75 2208 581.4 207.6 72.74 0 581 .5 2.967 2211 209.6 68.19 0 0.71 0.37 0.34 9/6/78 2239 581.9 211.6 72.68 0 4.45 2215 582 210.3 66.62 0 0.84 0.37 0.47 9/7/78 2215 581.6 206.5 70.62 0 1.4 2211 581.6 200.5 68.68 0 0.31 0.37 -0_.06.
9/8/75 221F 582 211.3 65.34 0 1.233 2210 581 .5 208.8 63.66 0 0.54 0.37 0.17
?/9/78 2219 581.7 206.2 61.69 0 1.617 2218 582 209.8 59.36 0 0.60 0.37 0.23 9/12/78 2216 580.9 201.6 70.9 0 1.367 2219 581.1 202.6 68.56 0 0.96 0.37 0.59 9/13/78 2215 581.1 202 78.17 0 1.9 2211 580.8 202.1 74.4 0 0.79 0.37 0.42 9/14/78 221o 580.0 203.4 78.25 0 1 2211 580.7 201.4 77.26 0 0.67 0.37 0.30 9/15/78 2210 580.7 208.7 69.45 0 1 2216 581.3 211.1 68.03 0 1.15 0.37 0.78 e n .:/ o oot t e.m t 9 pin,o /.p.41 0
~
' TABLE 3 ( Continued),
2 l REACTOR COOLANT SYSTEM' LEAK RATES i
De.v.i s Besse Unit 1 RCS AUG PZR .MUT PRT LEAK RATES PRESS TEMP LEVEL LEVEL LEVEL (- -
- gpm - - ->
(psig). (F): (in) (in) (in) GROSS IDENT UNIDDIT ATE START START START START JR(Hr ) END END END END
>/17/78 22!1 581.1 209.4 59.21 0 1' 2213 581.2 209.3 57.88 0 0.86 0.37 0.49
?/18/78 2218 581.4 211.9 61.44 0 0.37
~
1.067 2215 581.6 214.3 58.54 0 1.28 0.91
?/15/78 2219 581 .3 211.3 74.87 0 1.033 2210 581.3 210.9 72.93 0 1.12 0.37 0.75 2/19/78 2211 581.7 216.4 70.27 0 .
1.017 2225 581.5 215.9 69.02 0 0.35 0.37 -0.02 P/20/78 2216 581.2 211.4 67 0 4 2221 581.6 213.7 60.79 0 0.84 0.37 0.47
?/20'/78 2218 581 211.6 65.02 0 .
2.017 2216 581.2 210.1 63.07 0 0.81 0.37 0.44 9(21/78 2211 581'.4' 209.6 71.73 0 3.733 2215 581.9 213.5 64.91 0 0.95 0.37 0.58 9/22/78 2218 581.3 210 66.77 0-4 . 2219 581.9 212.2 58.83 0 1.16 0.37 0.79 9/23/78 2215 581.6 210.8 65.66 0 _.
2 2215 581.9 211.9 61.77 0 1.14 0.'37 0.77 9/24/78 2210 581.5 211.2 79.7 0 4 2216 581.6 211 71.23 0 1.14 0.37 0.77
, 9/25/78 2219 581.9 212.4 80.73 0 1.367 2219 581.7 213.1 78.25 0 0.62 0.37 0.25 9/2d/78 2211 581.3 211.3 66.05 0 4 2219 581.6 211.8 59.64 0 0.91 0.37 0.54 9/29/78 2208 532.7' 154 76.1 0 4 2208 532.6 154 68 0 1.03 0.52 0.51 9
--,---,ws- n -w-~~- - , - - ~ -w o--' e-+ -~---r--- u -b*- -
-TABLE 4 i COMPARISON OF UNIDENTIFIED LEAK RATE TEST RESULTS !
CALCULATED BY NRC COMPUTER AND BY DAVIS-BESSE Leak Rate (opm) Leak Rate (opm)
Date NRC Davis-Besse Date NRC Davis-Besse Duration (Hr) Duration (Hr) 8/30/78 0.26 0.30 9/16/78 0.60 0.71 1 1 8/31/78 0.51 0.515 9/17/78 0.49 0.32 1 1 9/1/78 0.72 0.126 9/18/78 0.91 0.755 1.05 ,
1.067 9/2/78 0.34 0.359 9/18/78 0.75 0.566 3.28 1.033 9/3/78 0.35 0.33 9/19/78 0.28 0.39 4.433 1.017 9/4/78 0.29 0.26 9/20/78 0.47 0.472 .
1.866 4 9/5/78 0.34 0.32 9/20/78 0.44 0.437 2.967 2.017 9/6/78 0.47 0.37 9/21/78 0.58 0.55 4.45 3.733 9/7/78 -0.06 0.17 9/22/78 0.79 0.77 1.4 4 -
9/8/78 0.17 0. 0 9/23/78 0.77 0.75 1.233 2 9/9/78 0.23 0.12 9/24/78 0.77 0.80 1.617 4 9/12/78 0.59 0.60 9/25/78 0.25 0.13 1.367 1.367 9/13/78 0.42 0.37 9/26/78 0.54 0.454 -
l .
- 1. 9 4 .
9/14/78 0.30 0.191 9/29/78 0.51 0.497 1 4 9/15/78 0.78 0.85 1
,- y .,-_.,,m -
.1 - T DLE 5 ,
REACTOR COOLANT SYSTEM LEAK RATES Rancho Seco Unit 1 (Daily Test)-
'RCS AVG P'Z R MUT PRT LEAK RATES PRESS TEMP LEVEL LEVEL LEVEL (- -
~ gpm - - ->
<tsig) (F) (in) (in) (ln) GROSS 1 DENT Lt41 DENT
. ATE - START START START START JR( Hr-) .END END END END 3/1/78 2186 582.1 180 80.9 39.9 1 2187 582.1 180.9 77.8 42.3 1.40 1.34 0.05 . ,
S/2/78 2205 582 180 77.9 24.2 1.05 2208, 582 180 73.4 26.9 2.17 1.44 0.73 8/3/78 2184 582 180 83 49.4 1.467 2184 582 180 76.4 54 2.30 1.76 0.55 8/4/78- 2207 582 182.2 79.7 40.5 1 2205 581.6 182.7 74.7 43.8 1.82 1.85 -0.03 8/5/78 21 87 582.2 180.7 69.9 33.9 1.033 2187 582.2 180.6 66.1 36.8 1.90 1.57 0.33 8/6/78 '2196 581.2 182 72.1 42.4-1 2201 581.2 182 69.3 45.3 1.40 1.62 -0.22 2/7/73 22~7 581.7 180.8 79.3 48.4 -
1 2234 582.7 181 75.2 51.7 3.71 1.85 1.86 S/8/78 2178 582.3 178.7 77.2 53.1 1 2131 582'.4 179.8 72.8 56.1 2.17 1.68 0.49 8/9/78 21S7 581.8 18C' . 2 78.8 35.8 1 2137 581.8 180.5 75.2 38.6 1.78 1.57 0.22 8/10/[8 2221 582 182 86.1 24.5 1 2205 582.9 182.4 82.6 26.4 3.29 1.06 2,.,22 E/1.1/78 2201 582.7 180.9 79.5 55.2 1 21 81 582.6 179.2 76 57.5 2.12 1.29 0.83 8/12/78 220E 583.3 181.2 90.6 47.5 1 2205 583.3 181.4 86.9 50.2 1.88 1.51 0.37 l
t
-- - - - - - , , - - 0 - , , _ , , . , , , _ - ,_ , _ , -
TABLE 5 (Cdhtinued) . 2 REACTOR COOLANT-SYSTEM LEAK RATES l Ranchc. Secc. Unit 1 (DailyTest)_ -
RCS AVG PZR MUT PRT LEAK RATES PRESS TEMP LEVEL LEVEL LEVEL (- - - gpm - - ->
(ps.ig) (F) (in) (in) (in) GROSS IDENT UNIDENT ATE START START START ' START' Fd Hr ) END END END END
- /13/78 2181 583.3 181 86 26.8 1 2202 583.2 181.4 82.1 29.1 1.59 1.29 0.30-
- /14/78 2181 582.4 180.4 80.? 29.7 1.017 2181 582.6 180.3 78.6 32.3 1.50 1.43 0.07
./15/75 2207 583.1 181.6 81.1 49.4
-1 2201 583.1 181.2 78 51.8 1.71 1.34 0.37
- /16/78 2207 582.5 179.2 73.9 32.1
.1 2207 582.7 179 71.4 34.3 1.65 1.23 0.41
- /17/78 2207 583.1 181.8 77.9 27.6 1 2208 583.1 182.1 . 74.9 30 1.47 1.34 0.12
../18/78
- 2180 '582.7 181.5 77.4 36.9 1 2180 582.6 180.6 74.5 39.4 1.50 1.40 0.10 3/1F/79 2202 583.1 182.7 77.4 41 1 2202 582.9 182.7 73.4 43.5 1.72 1.40 0.32 l
3/20/78 2181 582.5 181.5 80.8 43.5 1 . 2181 582.7 180.8 77.6 46.5 2.11 1.68 0.43 5/21/78 .2202 583.2 183 80.6 39 1 2205 583.1 182.8 76.7 41.5 1.85 1.40 0.45 E/22/75 2207 582.4 182.2 74.3 38.4 1 2181 582.3 181.1 70.5 40.8 2.19 1.34 0..85 B/2S/78 2201 582.1 182.6 80.8 30.4 1 2199 582.1 182.1 77.8 32.9 1.65 1.40 0.25 i
t
- l. ITABLE 5 (Continued) 3' REACTOR COOLANT SYSTEM LEAK RATES Rancho Seco Unit 1 (Daily Test)
RCS AVG P2R MUT PRT LEAK RATES PRESE TEMP LEVEL LEVEL LEVEL (- -
gpm - - ->
(psig) (F) (in) (in) (in) GROSS 2 DENT UNIDENT
> ATE START START START START JR( Hr-) END END END END
-3/24/78 2167 581.8 182.1 82.8 31 1 2169 582 182.4 80.1 33.5 1.64 1.40 0.24 3/25/78 2178 . 582.4 180.4 76.4 31 1 2180 582.4 179.1 74.5 33 1.22 1.12 0.10
, S/26 /78 2178 582.2 182 84 32.7 1.333 2178 582.2 181.6 80.3 35.5 1.48 1.18 0.30 8/27/78 2195 582.7 182.'1 71.1 37.9 1 219e 582.6 182.7 67.4 40.2 1.58 1.29 0.29 8/28/78 2103 582.4 182.3 73.5 27
- 1 .21o2 582.4 182.2 70.3 29.5 1.67 1.40 0.27 8/29/78 2204 582.1 183.3 80.3 33.7 1 2205 582.1 183.o 77 36.2 1.56 1.40 0.16 .
8/30/76 21 *
- 582.3 183 75.1 48.1 -
1 21o? 582.2 183.3 72 50.5 1.37 1.34 0.02 S/31/78 2199 581".9 183.2 80.8 34.2 1 2201 581.6 183.5 76.8 36.4 1.49 1.23 0.26 I
O
\
- l l
l .
TABLE 6' REACTOR COOLANT SYSTEM LEAK RATES.
Rancho Seco Un i t 1 '(h'eekly Test)
RCS AUG PZR MUT PRT LEAK RATES PRESS TEMP LEVEL LEVEL LEVEL (- -
gpm - - ->
(osig) (F) (in) (in) (in) GROSS IDENT UNIDENT DATE START START START START
)UR(Hr> El!D END END END S/9/78 2199 582.3 180.5 77.7 26.5 1.017 2204 5E2.1 181.3 73.9 29.3 1.40 1.54# -0.14 #
S/1S/78 21?S 582.8 181 84.1 31.8 1 21?2 592.6 181.4 80.3 34 1.59 1.23re 0.35 #
8/2?/78 2207 582.6 ISS.4 75.S 47.5 1 2207 582.1 182.9 71.7 49.9 1.38 1.34 # 0 . 0 4+
S/31/78 2183 582 182.9 82.9 45.1 1 2181 , 582.1 182.1 79 5 47.5 2.08 1.34,5 0.73g
- Does not include 0.5 gpm correction for evaporative losses. This value was identified during hot functional testing and routinely added to the identified leakage calculated in the. weekly pr6cedure.
w.
e 9
, . , . - - - . - . - ._ _ . . ,, .,.__,7- _ , _ _ _ . . _ _ . - _ , , _ , . . _ . , , - . . _ , ,,eo_~, . . m-
TABLE 7 C0!4 PARIS 0N OF UNIDEITIFIED LEAK RATE TEST RESULTS CALCULATED BY HRC COMPUTER AND BY THE RANCHO SECO DAILY PROCEDURE Date Leak Rate (opm) Date Leak Rate (opm)
Duration (Hr) NRC Rancho Seco Duration (Hr) NRC Rancho Seco 8/1/78 0.05 0.03 8/17/78 0.12 0.12 1 1 4
8/2/78 0.73 0.77 8/18/78 0.10 0.36
, 1.05 1 8/3/78 0.55 0.56 8/19/78 0.32 0.64 1.467 . 1 8/4/78 -0.03 0.60 8/20/78 0.43 0.61 1 ,.
1 8/5/78 0.33 0.34 8/21/78 0.45 0.49 1.033 1 8/6/78 -0.22 -0.14 8/22/78 0.85 0.87 -
1 1 .
8/7/78 1.86 0.21 8/23/78 0.25 0.26 '
1 -
1 8/8/78 0.49 0.32 8/24/78 0.24 0.08 1 1 i
'8/9/78 0.22 0.21 8/25/78 0.10 0.17 1 1 -
8/10/78 2.22 0.64 8/26/78 0.30 O.34 1 1.333 8/11/78 0.83 0.90 8/27/78 0.29 0.49
. I 1
! 8/12/78 0.37 0.34 8/28/78 0.27 0.27 l
1 1 l 8/13/78 0.30 0.62 8/29/78 0.16 0.15 1 1 l
1
'. 8/14/78 0.07 0.23 8/30/78 0.02 0.18 t
1.017 1 8/15/78 0.37 0.34 8/31/78 0.26 0.35 1 1 8 0.41 0.16
. /16/78 .
TABLE 8 COMPARISON OF UNIDENTIFIED LEAK RATE TEST RESULTS CALCULATED BY NRC COMPUTER AND BY'THE RANCHO SECO WEEKLY PROCEDURE Date Leak Rate (GPM)
Duration (Hr) NRC Rancho Seco 8/9/78 .64 .65 1.017 8/18/78 .15 - .20 1
8/23/78 . .46 .46 1
8/31/78 .23 .24 1..
e S
9 Ne 1
4 e
l O
, , - - --e - - --,~-e - -- ,. , , , . - -- -