ML20211N667
| ML20211N667 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 07/01/1986 |
| From: | Goldberg J NRC OFFICE OF THE EXECUTIVE LEGAL DIRECTOR (OELD) |
| To: | Bright G, Kelley J, Kline J Atomic Safety and Licensing Board Panel |
| References | |
| CON-#386-814 LRP, NUDOCS 8607030178 | |
| Download: ML20211N667 (46) | |
Text
\ our Ue ,ih, UNITED STATES JEp wititurunL4M i[ g NUCLEAR REGULATORY COMMISSION O : E WASHINGTON, D. C. 20555 July 1,1986 '
0{CMETED gg James L. Kelley, Chairman Jerry R. Kline Presiding Board Presiding Board GF U.S. Nuclear Regulatory Commission U.S. Nuclear Reg @FICE
@tetyn;qoor 9merk2.l,ssion Washington, DC 20555 Washington, DC 205555RANoi
Glenn O. Bright Presiding Board ,
U.S. Nuclear Regulatory Commission Washington, DC 20555 In the Matter of INOUIRY INTO THREE MILE ISLAND UNIT 2 LEAK RATE DATA FALSIFICATION f Docket No. LRP M' ,
Dear Judges:
In accordance with the Board's May 22, 1986 Memorandum and Order, enclosed is the " Testimony of Donald C. Kirkpatrick and . Jare'd S. Wermiel" with attachments.. This testimony constitutes the NRC Staff's technical " overview" testimony in this proceeding. We intend to . ask Mr. Kirkpatrick a few additional questions, as supplemental direct testimony, in order that he may identify and sponsor as exhibits some photographs of the TMI-2 control room.
Sincerely, f
ack R. Goldberg Acting Deputy Assistant General Counsel
Enclosure:
As stated cc w/ encl. : Service list T
TSO7
. WTED CORRESPONDEng(
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION h0 ,
a at
~
BEFORE TIIE PRESIDING BOARD u
k '2 g 6 o
- S!ct -
In the Matter of ) S
,ge[!I y INQUIRY INTO TIIREE MILE ISLAND ) Docket No. LRP UNIT 2 LEAK RATE DATA ) -
FALSIFICATION )
) .
TESTIMONY OF DONALD 'C. KIRKPATRICK AND JARED S. WERMIEL Q.1 Please state your name, your present employer and your cur-rent professional position.
A.1 (Kirkpatrick): My name is Donald C. Kirkpatrick. I work as a nuclear engineer for the NRC in the Engineering and Gendric Conimunications Branch, which is under the Office of Inspection and Enforcement. A copy of my professional qualifications is attached. -
n (Wermiel): My name is Jared S. Wermiel. I am currently a Section Leader in the Plant , Electrical, Instrumentation and Control Systems Branch, Division of PWR Licensing-B, Office of Nuclear Reactor Regulation, U.S. Nuclear Regulatory Commis-sion. A copy of my professional qualifications is attached.
Q.2 Generally, what is your knowledge, training and/or experience regarding reactor coolant system (RCS) leak rate testing.
n
I I
A.2 (Kirkpatrick): My first detailed involvement in reactor coolant system leak rate testing was in 1980 when I reviewed the leak rate testing at Three Mile Island, Unit 2 (TMI-2). As a result of this review, NRC embarked on a program, which I helped to develop and maintain, to independently calculate the reactor coolant system leak rates at all pressurized water reactors. As a part of this effort I helped to write computational programs .
for programmable calculators and personal computers to assist NRC inspectors in making the leak rate determinations. I coauthored NUREG-0986 and NUREG-1107, which are the users' >
guides for RCS leak rate .dctermination programs for the .
Osbourn and the IBM personal computers ,~ respectively. In conjunction with these and other efforts, I have ma'de detailed reviews of the RCS leak rate tes, ting practices' at six plants'. I liave also been involved in the general review of' the results of the independent measurement of RCS leak rates for all
^
pressurized water reactors.
(Wermiel): In my professional capacity, I am. responsible for reviews for compliance of various nuclear power plants with the licensing criteria in the area of reactor coolant leakage detec-tion, including leak rate testing.
~
Q.3 Do you have specific knowledge of the leak rate testing re-f quirements and practices at Three Mile Island Unit 2 during the time period from February 2,1978 (the date TMI-2 received its
. operating license) until March 28,1979 (the date of the accident at TMI-2)?
A.3 (Kirkpatrick): Yes. I participated in the initial NRC investi-gation of Mr. Hartman's allegations regarding RCS leak rate testing at TMI-2 in March and April,1980.
(Wermiel): , I have knowledge of the licensing requirements associated with leak rate testing at TMI-2, but not the specific practices associated with the implementation of those requirements.
Q.4 Why is leak rate testing a safety requirement?
A.4 (Wermiel): Loss of reactor coolant inventory is, a major concern
~
in nuclear power plants because a proper 'water volume is need-ed to maintain core cooling. Thus,' it is important ~to detect leakage before unacceptable losses ' occur. In this regard, Gen-eral Design Criteria (GDC) 30 " Quality of Reactor Coolant Pressure Boundary" of Appendix A to 10 C.F.R. Part 50,
" General Design Criteria for Nuclear Power Plants" requires:
Components which are part of the reactor coolant pressure boundary shall be designed, fabricated, erected, and tested to the highest quality standards practical. Means shall be provided for detecting and, to the extent practical, identifying the location of the source of reactor coolant leakage.
The safety significance of leaks from the reactor coolant pressure boundary can vary widely depending on the source of the leak as well as the leakage rate and duration. Therefore, the detection and monitoring of leakage of reactor coolant into
the containment area is necessary. Methods for quantifying the leakage from identified sources are necessary so that the !
leakage from an unidentified source can be determined by the e operators and immediate corrective action implemented.
Q.5 Where is the NRC guidance for the control of RCS leakage set forth?
A.5 (Wermiel): Specific guidance governing leakage detection system design and capability is contained in Regulatory Guide 1.45, " Reactor Coolant Pressure Boundary Leakage Detection Systems." This Regulatory Guide provides details of -
acceptable methods for implementing the broad requirements of GDC 30. In addition, individual plant' technical specifications
.proyide the following: -
- 1. the definition of the various forms of 'reacter coolant system leakage, .
- 2. the maximum leakage limit on the various forms of reactor coolant system leakage permitted for safe operation of the facility (limiting conditions for operation or LOOS),
- 3. the bases for the leakage limits,
- 4. the required action in the event the leakage limits are exceeded (action statement), and,
- 5. the required surveillance tests necessary to demonstrate that the various forms of reactor coolant system leakage are within the limiting conditions for operation.
d
. Q.6 Where in the TMI-2 Technical Specifications are items 1 through 5 above found.
A.6 (Wermiel): Items 1 through 5 above are found in the following sections of the TMI-2 Technical Specifications:
- 1. definition of the various forms of reactor coolant system leakage -- Section 1.14 through 1.17,
- 2. the maximum leakage limit on the various forms of reactor coolant system leakage permitted for safe operation --
Section 3.4.6.2,
- 3. the bases for the leakage' limits --
Bases Sec-tion 3/4.4.6.2, -
- 4. the required action in the event the leakage limits are exceeded -- Section 3.4.6.2,
- 5. t'he required surveillance tests nece'ssary to demonstrate that the various forms of reac'ort coolant system leakage are within the limiting conditions for operation -- Section 4.4.6.2.
In addition, reporting requirements (notification to NRC) are contained in Section 6.9 and the requirements for retention of records are contained in Section 6.10.
Q.7 What are the different forms of reactor coolant leakage?
I A.7. (Wermiel): The TMI-2 Technical Specifications define four different forms of reactor coolant leakage. They are as follows:
)
! 1. IDENTIFIED LEAKAGE shall be:
. a. Leakage (except CONTROLLED LEAKAGE) into closed systems, such as pump seal or valve packing leaks that are captured and conducted to a sump or collecting tank, s b. Leakage into the containment atmosphere from sources that are both specifically located and known either not to interfere with the operation of leakage detection systems or not to be PRESSURE BOUNDARY LEAKAGE.
- c. Reactor coolant system leakage through a steam generator to the secondary system. (Technical Specification 1.14).
- 2. UNIDENTIFIED LEAKAGE 'shall be all leakage which is not IDENTIFIED LEAKAGE or CONTROLLED LEAKAGE.
(Technical Specification 1.15).
- 3. PRESSURE BOUNDARY LEAKAGE shall be leakage (except steam generator tube leakage) through a non-isolable fault ,
in a Reactor Coolant System com.ponent body, pipe wall or vessel wall. (Technical Specification 1.16).
- 4. CONTROLLED LEAKAGE shall be that seal water flow supphed from the reactor coolant pump seals. (Technical Specification 1.17). ,
Q.8 Ilow is reactor coolant leakage detected?
A.8 (Wermiel): There are several methods of detecting reactor cool-ant system leakage. A separate (individual) direct monitoring system is provided to detect Controlled Leakage. In addition, closed systems providing for collection of Identified Leakage have individual monitoring capability. The technical specifica-tions are primarily concerned with assuring operability of RCS leakage detection systems and performance of tests utilized for detecting Unidentified Leakage. These leakage detection systems consist of containment atmosphere (gaseous and/or particulate) monitors, the containment sump level and flow l
l
. monitoring system, and the containment air cooler condensate flow monitoring system. In addition to the above systems, the performance of a RCS water inventory balance (leak rate test) provides a method for quantifying total system (gross) leakage and unidentified leakage.
Q.9 What limits do the TMI-2 Technical Specifications set for leakage i and what is the required action should these limits be exceeded?
A.9 (Wermiel): Section 3.4.6.2. of the TMI-2 Technical Specifi-cations limit the Reactor Coolant System leakage to: .
j
- a. No PRESSURE BOUNDARY LEAKAGE.-
- b. 1 GPM UNIDENTIFIED LEAKAGE. .
- c. 1 GPM total primary-to-secpndary leakage through
'[the] steam generators. .
- d. 10 GPM IDENTIFIED LEAKAGE.
- e. 8 GPM CONTROLLED LEAKAGE at a Reactor Coolant System pressure of 2155 + - 50 psig.
These limits applied at all times when the Reactor Coolant system temperature exceeded 200 F.
If these limits were exceeded, Section 3.4.6.2 of the TMI-2 Technical Specifications required the following action:
- a. With any PRESSURE BOUNDARY LEAKAGE, be in at least HOT STANDBY within 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> and in COLD SliUTDOWN within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
- b. With any Reactor Coolant System leakage greater than one of the above limits, excluding PRESSURE BOUNDARY LEAKAGE, reduce the leakage rate to within limits within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> or be in at least HOT t
-g-STANDBY within the next 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> and in COLD SIIUTDOWN within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
Q.10 What are the technical bases for the leak limits identified in Th11-2 Technical Specification 3.4.6.2.
A.10 (Wermiel): The technical bases for the leakage limits identified in Th11-2 Technical Specification 3.4.6.2 are found in Section 3/4.4.6.2 of the Bases Section of the Technical Specifications, i .
This section states that:
PRESSURE BOUNDARY LEAKAGE of any magnitude l
is unacceptable since it may be indicative of an impending gross failure of the pressure boundary. -
j Therefore, the presence of any PRESSURE-BOUNDARY LEAKAGE requires : the unit to be promptly placed in COLD SHUTDOWN.
Industry experience has shown that, while a lim'ited amount of leakage is expected from . the RCS, the
. UNIDENTIFIED LEAKAGE portion of this can be reduced to a threshold value. of less than 1 GPhf.'
This threshold value is sufficiently low t~o ensure early detection of additional' leakage.
The total steam generator tube leakage limit of.
1GPM for all steam generators ensures that the dosage contribution from tube leakage will be limited to a small fraction of Pait 100 limits in the event of either a steam generator tube rupture or steam line break. The 1 GPM limit is consistent with the j assumptions used in the analysis of these accidents.
The 10 GPM IDENTIFIED LEAKAGE limitation provides allowance for a limited amount of leakage from known sources whose presence will not inter-fere with the detection of. UNIDENTIFIED LEAKAGE by the leakage detection systems.
The CONTROLLED LEAKAGE limit of 8 GPM restricts operation with a total RCS leakage to all RC pump seals in excess of 8 GPM.
i i In addition to the above, Section 5.2.7 of the TMI-2 Final )
l i Safety Analysis Report (FSAR) describes the reactor coolant !
l l
--- c .-. - - . . - . - . - . . - , , - - - - - - - - . , , ,
. -g-pressure boundary leakage detection systems that were relied upon by the staff during the operating license review of TMI-2.
Based upon the information contained in the FSAR, the staff concluded in its Safety Evaluation Report (NUREG-0107) that the licensee's design and description of the reactor coolant pressure boundary leakage detection systems constituted an acceptable basis for satisfying the requirements of GDC 30.
Included in the material rell.ed upon by the staff was the licensee's analysis of the maximum allowable leakage (Sections 5.2.7. 3 and 5.2.7.4 of the FSAR). The maximum allowable unidentified leakage was specified as 1 gpm.. The -
value of 1 gpm was selected for the following reasons: This value was well below the leakage associated with~ a crack of critical size in the reactor coolant pressure boundefy; this
.value of leakage could be detected 'within a reasonable period of time; and it was believed that continued operation at this level for some period of time to allow for corrective action would not jeopardize plant safety or result in external releases that would exceed 10 C.F.R. Part 20 limits. Based upon this analysis, the maximum value of 1 gpm for unidenti-fied leakage was incorporated into the TMI-2 Operating License as part of the Technical Specifications.
Q.11 What methods do the TMI-2 Technical Specifications provide for monitoring leakage to make sure that the five leakage limits are not exceeded?
l i
1
. A.11 (Wermiel): Section 4.4.6.2. of the Technical Specifications, under the Surveillance Requirements, requires that the Reactor i Coolant System leakages be demonstrated to be within each of the limits by:
4
- a. Monitoring the containment atmosphere particulate radioactivity monitor at least once per 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />,
- b. Monitoring the containment sump inventory and dis-charge at least once per 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />,
- c. Measurement of the CONTROLLED LEAKAGE from the ' reactor coolant pump seals when the Reactor Coolant System is 2155 + - 50 psig at least once per 30 days.
~d. Performance of a Reactor Coolant' System water in-ventory balance at least once per .72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> during-steady state operation. -
Of these methods, the water inventory balance is very impor-tant, si"ce it is the most accurate method for measuring'.'the "unillentified" leakage. The uniden,tified leakage, in turn, is the best means of identifying pressure boundary leakage.
The containment atmosphere particulate radioactivity monitor accuracy and sensitivity for detecting small reactor coolant system leakage rates is dependent upon many factors which
]
- reduce the accuracy of leak rate determination. Some of these factors are: the specific particulate activity of the coolant leaked, any filtration associated with the reactor coolant leakage e
pathway (e.g., valve packing or pipe lagging), time delays associated with radioactivity buildup in containment, the rate of sampling of the containment atmosphere, and the sensitivity of the monitor. For these reasons, this method is the least 9
1
accurate and sensitive for demonstration of compliance with unidentified leak rate limits.
The accuracy of containment sump inventory monitoring is similarly limited because sources of water leakage other than reactor coolant system leakage collect in the sump. These leakage sources must be identified and accurately measured in order to determine reactor coolant system leakage into the sump.
The controlled leakage limit- is associated only with reactor ;
coolant pump seal performance and is thus not related to
" unidentified" leakage. ,,
Q .12 What reporting and record retention ' requirements, applicable to leak rate testing, are included in the TMI-2 Technical Specifications?
A.12 (Wermiel): Section 6.9.1.8.b requires a report to the Director of the NR C's Region 1 Office within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> when any parameter subject to a limiting condition for operation is less conservative than its limit.
1 Section 6.10.1.d requires that records of surveillance activities required by the Technical Specification be retained for at least five years.
Q.13 What is the concept behind the reactor coolant system water inventory balance?
A.13 (Kirkpatrick'): The performance of the water inventory balance is a method of measuring the reactor coolant system leak rate by determining the change in the inventory of water in the primary coolant system during a specified time period and then dividing this change by the time period to derive the leak rate.
The leak rate derived this way is known as the " gross leak rate." Most of this leakage normally comes from valve stems or pump seals, where it is collected and piped to a holding tank,
~
such as the reactor coolant drain tank (RCDT). The leak rate -
into such tanks is also measured and is designated as the
" identified leak rate," because the sources .of this type of leakage are known. Any other . leakage, 'where the source can be identified and which can be quaniified, may also be included in the identified leak rate. The ' difference between the gross leak rate and the identified leak rate is designated as the
" unidentified leak rate." The unidentified leak rate is the most important of these values, because it could represent leakage from a primary coolant system component body, pipe wall or vessel wall of the reactor coolant pressure boundary, Q.14 What is the makeup tank (MUT) and what are its functions?
A.14 (Kirkpatrick): The makeup tank collects the reactor coolant system letdown and provides a suction reservoir for the makeup pumps. This tank also provides a hydrogen cover gas which is
- dissolved into the reactor coolant to scavage oxygen from the reactor coolant system.
Q.15 What is the reactor coolant drain tank and what are its functions?
A.15 (Kirkpatrick): The reactor coolant drain tank collects, condenses and cools the discharge from the pressurizer safety and power-operated relief valv,es. It also collects discharge or leakage from a number of other primary system valves and pump seals that is piped to the reactor coolant drain tank.
This tank is normally kept partially . filled with about six ;
' thousand gallons of cold water which 'is provided to quench the steam that is emitted from these various leakage sources. The rise, in the level of this tank . during ~ the inventory balance measurements provides the principle' ~means of -determining the
. Identified leakage. Water can be ~ transferred by the operators from the reactor coolant drain tank to the makeup and puri-fication system.
Q.16 How is the level of water in the make-up tank measured?
A.16 (Kirkpatrick): Most of the tank level measurements, including the makeup tank, are made by taking advantage of the fact that i
i the change in the pressure of the water at the bottom of the tank is directly proportional to the change in the level of the tank. A differential pressure detector, connected by metal tubing to the top and bottom of the tank, is used to measure 1
- the difference in pressure between these two points. This pressure difference is then divided by the incremental pressure per unit height of water to derive the level. Two separate level transmitters (LT-1 and LT-2) are provided for the makeup tank. (See Attachment 1 at the end of this testimony.)
Q.17 How is the water inventory balance test performed?
A.17 (Kirkpatrick): The change in, inventory in the primary system is determined by measuring levels of the various tanks and ves-sels in the reactor coolant system at the beginning of the time period and comparing them to the levels. at the end of the time -
period. The effect of temperature and pressure on the quantities of primary water in the reactor vessel, the steam generators, the piping and the .p'ressurizer are also accounted for. Of course, if water is added "to or removed from any of these tanks or vessels by the operators during the
~
measurement , it must also be accounted for in order to have valid test results. Since the density of the water varies from one part of the reactor coolant system to another, all of these quantities are normally converted to pounds of mass before they are summed.
The levels of the primary leakage collection tank (or tanks) at the beginning and end of the test period are also compared to determine the identified leak rate. Any leakage from the reactor coolant system not collected in these tanks, for which the source can be identified and quantified, may also be added
o
- to the identified leak rate. Again, if any fluid is removed from a collection tank during a test it must be accounted for.
The water inventory balance may be done by a computer or by a hand calculation. If a computer is used, the operator simply enters the appropriate command for the test at the computer terminal. The computer prompts the operator for a test dura-tion, then gathers most of th,e necessary parameters from the automatic data acquisition system at the beginning and end of the test period. The computer then prompts the operator for any necessary additional information, such as the quantity of -
water added or removed during the test and additional identi-fied leakage. The computer then makes all of the necessary
~
calcul'ations and prints out a repoit of the key parameters used in the calculation as well as the variods types of leak rates. If hand calculations are performed, the operator reads all of the necessary parameters directly from indicators (normally) in the control room at the beginning and end of the test period and enters them on a work sheet. The work sheet is then followed in performing the necessary calculations.
Q.18 How was the requirement to perform an RCS inventory balance test implemented at TMI-2?
A.18 (Kirkpatrick): Both a computer program and hand calculation work sheets were provided. However, practically all of the tests were done using the computer. The operator entered the
- . _ = _ _ . . - .
O
. code "rcsl" at the appropriate computer terminal. The program then prompted the operator for the test interval. A period of from one to eight hours was permitted, but in all reviewed cases an interval of one hour was entered. The program then 1
collected data on the reactor coolant system temperatures, pres-surizer level, make-up tank level and reactor coolant drain tank level for the data acquisition system. At the end of the inter-val the computer collected another set of these parameters. At this time the computer also prompted the operator to enter the operator-caused changes to the drain tank and .to the primary i
any identified leakage (other tiian leakage collected in ' ;-
system ,
~
the drain tank), and primcry to secoddary- leakage. When any of these did not exist, the operator-was required to i*e'spond by ,
~
entering a zero. The com pu.te'r' thed'made the necessary l .
calculations, and printed out . the pa'rameters collected', as well as the gross, identified and unidentified leak rates. This printout, which also contained the operator's entries, as well as places for the operator and a supervisor to sign off, then became the official test record. (A sample . copy of a test j record is provided as Attachment 2 at the end of this testimony. )
Q.19 How were leakages calculated at TMI-2 between February 7, 1978 and March 28, 1979?
i A.19 (Kirkpatrick): Detailed worksheets were included in the surveillance test procedure that originally guided the operators i
I
- through the calculations. The computer calculations were based on these worksheet calculations, except that the computer did not make any corrections for pressure variations as was provided for in the worksheets. The equations, on which both the worksheet and computer calculations were based, have been abstracted from the worksheets and are detailed in Attachment 3 at the end of this testimony.
Q.20 Ilow accurate was this test procedure at TMI-2?
A.20 (Kirkpatrick): The test procedure at TMI-2 was not suffi-ciently accurate to determine unidentified leakage of 1 gpm due :
to uncertainty in the data and computational errors. In entering a guilty plea in United States of A'merica v.
Metropolitan Edison Company, Criin No. ' 83-00188, Metropolitan Edison acknowledged, inter alia, th'at it was on notice, from sometime before mid-October 1978 up to the time of the acci-dent at TMI-2, that this procedure did not accurately and meaningfully measure the amount of unidentified reactor coolant leakage.
Q.21 What caused the uncertainty in the test procedure?
A.21 (Kirkpatrick): Uncertainty was caused by the periodic oscilla-tion of some of the parameters, as well as instrument uncertainty. This oscillation was significant because the auto-matic acquisition - of a data set was made by taking three readings at one minute intervals for each parameter and averaging the readings. Because of the oscillations, some of
I 1
- the parameters would change while the data set was being taken. The estimated variation of all the calculated leak rates, due to data uncertainty, was of the order of one gpm for the usual one hour test.
The uncertainty in the test results was magnified by the short duration of the test interval chosen. With the plant in steady state, the expected variation, in the leak rate is inversely proportional to the test interval. Therefore, the variation could have been halved or quartered by doubling or quadrupling the test interval. However, a one hour interval .
was almost always used. (The effect on the leak rates caused by the expected variation in the different parameters is shown in Table 2 of Attachment 4 at the bnd of this testimony.)
Q.22 What significant computational errors did the computer program contain?
A.22 (Kirkpatrick): Seven computational errors in the computer program were identified by NRC inspectors. Some were relatively minor, but the three which could cause errors of one gpm or more in all of the calculated leak rates are discussed below . (Table 1 of Attachment 4 gives the estimated variation caused by the computer errors.)
In the first case, the tables in the computer program used to convert temperature to density terminate at 5820 F .- When the l
l reactor coolant system temperature exceeded 582o F, the
. _ 19 _
temperature corresponding to 582o F was eclected. Out of the 161 tests reviewed , 49 had temperatures exceeding 5820 F (30%), resulting in errors as high as one gpm.
The two most serious errors were caused by the use of incon-sistent densities to convert mass of water to gallons of water.
Unter outside of the reactor weighs about 8.29 lbs/ gallon (at 70o F). At the average reac, tor temperature, however, water weighs about 5.86 lbs/ gallon (at 5820 F). In the first density error, the computer used a table based on ambient temperature conditions in establishing the number of. gallons of water, which -
entered the reactor coolant drain tank during the test. How-ever, the gross leakage was calculated by dividing the pounds of water lost from the primary . system by a density based on the average reactor coolant temperafure. The result was that for every 100 gallons of gross leakage from the primary system that entered the drain tank, the computer calculated about 70 gallons of identified leakage. While this caused a deficit in the identified leakage of 30 gallons, it also caused an excess in the unidentified leakage of 30 gallons , because the unidentified leakage is determined by taking the difference between the gross and identified leakages. i l
l A similar error was made in establishing the quantity of water added to the coolant system by the operators. The value that j was entered into the computer by the operators was based on j 1
l
- ambient gallons of water. The computer added these gallons of water to the gross leakage summation without converting them to reactor conditions. As a result of this error,100 gallons of water added to the reactor coolant system , and properly entered into the computer, would expand to 140 gallons at average reactor temperature, but would only be counted as 100 gallons in the inventory balance.
Q.23 What would be the effect of these errors on the likelihood of calculating leakage rates that exceeded the acceptable limits?
A.23 (Kirkpatrick): As a result of the drain tank density error, any time the gross leakage exceeded ' about 200 gallons (3.3 gpm) and provided no water was added during the test, the
~
computer should theoretically have calculated an unidentified leakage which exceeded the allowable value of one gpm, even if the true value was zero. Ilowever, due to the large random uncertainty in the calculation, a test having gross leakage of this magnitude would still have about a 50/50 chance of producing an " acceptable" unidentified leak rate if no unidenti-fied leakage actually existed. When the identified leakage in-creased above this level, it became increasingly improbable that an " acceptable" unidentified leak rate result would be calculated, regardless' of the true unidentified leak rate.
Q.24 Were any changes ever made by the TMI 2 staff to correct these errors in the leak rate calculation?
e - A.24 (Kirkpatrick): A procedure change, which required the use of work sheets to make hand corrections to the computer program,
- was introduced on March 16, 1979. This was done to correct i
the reactor coolant drain tank density error I discussed in 1
Answer 23, which overstated the unidentified leakage by not correcting the quantity of the identified leakage back to reactor i -
conditiens. The procedure change was accompanied by a written evaluation signed by ,the Unit Superintendent. This correction amounted to multiplying the computer-derived i identified leak rate by the ratio of the drain tank water density
- ~
to reactor coolant water density. This procedure did provide a :
l
! more accurate identified leak rate. However, the corresponding correction needed to adjust - the water added to the reactor coolan't system by the operators 'for ex~pansion in the reactor 2
was not made. During the time period in which the hand corrections were made, water was' being added to the reactor i .
coolant system , during every test, in amounts that were roughly equal to the identified leak rate. Therefore, the com-
- puter errors in the identified leakage and the computer errors in the water added, roughly cancelled each other. By correcting the identified leakage, but not the water added, the new procedure had the effect of understating the more l important unidentified leak rate. ,
l Q.25 Were the results of the RCS inventory balance required to be l
-logged? ,
i s
A.25 (Kirkpatrick): Yes. The performance of all surveillance tests required by the Technical Specifications was required to be logged in the control room log. This was required by the licensee's Administrative Procedure 1012, " Shift Relief and Log Entries." (Item 3.3.17) . [
i l Q.26 What actions were required to be taken for leak rate test results that showed unidentified leakage was greater than i !
1 gpm?
A.26 .(Kirkpatrick): The immediate actions to be taken if the RCS unidentified leakage is excessive are listed in item 7.2 and item 6.4 of Surveillance Test Procedure 2301-3D1. They are I
listed in successive steps as follows:
7.2 If unidentified reactor cooiant leakage excee'ds 1 gpm i
~
proceed with Action Statement 3.4.6.2.b of the !
i technical specifications. (This required reactor
, i shutdown, if the leakage could not be identified, is ;
described below.)
6.4 If the net reactor coolant system leakage is excessive t
as defined by the acceptance criteria in Section 7, ,
i
- proceed as follows:
i i
! l i 6.4.1 Perform another determination of the RCS leak rate. ;
4 l 6.4.2 Insure that no unaccounted for operator action has !
! occurred that would change the RCS inventory.
t
- 6.4.3 Initiate action to determine the source of leakage. (A series of items to check are listed.)
6.4.4 If the sources of the leakage are found, initiate Data Sheet 3, Identified leakage. (This required the determination and documentation of the leak rate, and a determination by the Shift Supervisor of the safety
, implications of the leak.)
If' the source of the leakage could n'ot be identified or reduced within four hours, then the Action Statement of Technical ;
Specification 3.4.6.2.b required that the reactor be -taken to hot standby (at operating temperature ' but suberitical) within ,
the , next six hours and to cold. 5hutdowri within the following 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />. In addition, Section 6.9.'1'.8.b required th'e prompt notification of the NRC within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> with a followup report within 14 days. If the source of the leakage could be iden-tified , it could be included with the rest of the identified leakage and no further action had to be taken unless the total identified leakage exceeded 10 gpm.
Q.27 What information is reflected in the makeup tank level recorder strip chart?
A.27 (Kirkpatrick): This strip chart contains a continuous time line of the level of water in the makeup tank; ir, effect, it records a :
continuous history of the quantity of water in the makeup tank.
J I
Q.28 Generally what type of plant, equipment or instrumentation con-ditions can affect the validity of a leak rate test?
A.28 (Kirkpatrick): In general, anything, other than actual water mass changes, that causes the plant parameters used in the calculations to change or changes in the sensors used to measure the parameters, can affect the validity of the test.
The plant conditions that can affect the leak rate test results include plant transients and oscillations. Plant transients and oscillations increase the uncertainty in the test results as I discussed previously. Therefore, this surveillance test was required to be run only during steady-state conditions.
Any. obstruction in the sensing ' lines leading from the various tanks to the differential pressure detectors used for level measurement can cause an erroneous level measurement. In particular, water in the line leading from the makeup tank level detector to the top of the makeup tank, the so-called " loop seal", sometimes caused errors in the makeup tank levels used by the computer.
Imprecise calibration or " noise" in the instruments providing the data to the computer can also cause errors. These types of problems in the makeup tank level instrumentation did cause errors in the leak rate test results by causing erroneous i
l 1
- changes in the apparent makeup tank levels between the beginning and end of the test.
Q.29 By what methods can operators minipulate leak rate tests to bring them within the Technical Specification limits?
A.2S (Kirkpatrick): There are at least five that I know of:
(1) By adding hydrogen to the makeup tank during the test, (2) By adding water to the reactor coolant system during the test, (3) By taking advantage of the plant oscillations during the test, (4) By switching makeup tank level detectors during the test, (5) And, prior to mid-November 1978, incorrectly reporting the reactor coolant drain tank' voltage'.
~
Q.30 What is the effect of adding hydrogen to the makeup tank dur-ing the test?
A.30 (Kirkpatrick): The addition of hydrogen during the test could sometimes cause a reduction in the measured gross leak rate
! and, consequently, the unidentified leak rate. Hydrogen is routinely added to the makeup tank in order to provide a cover gas and to reduce the free oxygen in the reactor coolant system. With the level instrumentation operating properly, the addition of hydrogen should have had no effect on the makeup tank level, because the pressure would be applied equally to the top and the bottom of tank, resulting in no change to the
differential pressure. However, if water collected in the low spot in the line leading from the top of the make-up tank to the differential pressure detector which provided the level indica-tion, a water seal could form. (See Attachment 5 at the end of this testimony.) Because two separate level detectors are used to measure makeup tank level, a water seal could be present in either detector reference leg or both. Thus, the effect may be ,
different between detectors or.may not exist at all. When this happened, the water seal, if it existed, isolated some of the pressure change at the top of the tank from the differential pressure detector. The result was an apparent upward shift in -
the makeup tank level equivalent to th'e relative change in the levels of the two ends of the water seal.' This . increase in le. vel caused an apparent addition to the water inventory at the end of the test.
With the water seal in place, it takes a very small pressure change to make a significant difference in the measured quanti-ty of water in the make-up tank. Theoretically, a pressure increase of only 0.036 psi could cause an upward shift of one inch in the measured tank level. This represents a water vol-ume of approximately 30 gallons. However, the upwaid chift in the tank level is limited by the inaximum height of the water seal, and usually did not exceed several inches. A similar downward shift in the makeup tank level occurred when the pressure decreased and the water seal shifted back to its l
l l
. original position. Any actual increase or decrease in the tank level could also cause the water seal to shift as the cover gas was compressed or expanded by the level change. As a result, the indicated change in level would be greater than the actual change in level. If hydrogen were to be added to affect the outcome of a leak rate test it generally had to be added toward the end of the test so that the normal downward trend in the tank , due to leakage , would, not reverse the upward shift before the final data set was acquired by the computer.
Q.31 How was water added to the makeup tank? -
A.31 (Kirkpatrick): Water could be added to the makeup tank from various sources. If it was necessary to reduce the RCS boron congentration, water was added .from th' 'edemineralized ' service water system. If it was necessary to ' increase boron concentra-tion , boric acid was added from tiie boric acid transfer pumps.
Borated water could be added from one of several primary water collection tanks located outside of the containment, known as reactor coolant bleed tanks. Normally, water from all of these sources was put into the makeup tank through a batch con-troller. As the water went in, a counter, known as a totalizer, counted the . number of gallons of water that went into the makeup tank. The operators could determine the exact number of gallons of water that went into the makeup tank by taking
- the difference in the totalizer readings between the beginning i
j and end of the water addition. Operators could also
- automatically add a predetermined amount of water using the batch controller. However, the leak rate test procedure did not specify that the amount of water that was added be deter-mined by the totalizer, so if the totalizer values were not recorded, the operator could determine the water addition by observing the recorded shift in the makeup tank level.
The batch controller control valve could be throttled down so that the water entered the makeup tank gradually. If this were done, the change in the recorded makeup tank level would also be gradual, so that the fact of the water addition would not be ,
obvious from an examination of the makeup tank chart.
Q.32 What 'is the effect of addition .o'f water' to 'the makeup tank
.during the running of the lenk rate fe'st?
A.32 (l;irkpatrick): If water is added during the running of the test but is properly accounted for in the test data or calcula-1 tions , including appropriate correction for reactor conditions, then there would be no effect on the test results. However, if water is added but not accounted for in the test data or calculations, then the test result showing the measured leak rate would be under-reported. Furthermore, the addition of water during the test could cause a reduction in the measured leak rate, even if the amount of the addition was entered into the computer, if appropriate corrections for reactor conditions are not made. The failure of the calculational procedure to
l
. account for the volume increase of the water as it was heated up in the reactor resulted in a smaller amount of water being credited to the inventory than was actually added. >
As I discussed previously, 100 gallons of water added to the reactor coolant system would expand to about 140 gallons in the reactor and reduce the apparent gross leakage by that much.
However, only 100 gallons would be accounted for. If the presence of a water seal caused an additional upward shift in th'e makeup tank level that persisted' to the end of the test, an additional reduction in the measured leakage would occur.- If'a ;
zero, instead of the 100 gallons, were' 'en'tered in response to the computer prompt for operator-caus'ed cha.n ges , a further
. reduction in the measured leakage would ~occt$r, since none of the addition would be accounted for tiy the computer.
s -
The normal addition of water causes a sudden characteristic change in the recorded makeup tank level. Ilowever, it is possible to enter a smaller quantity of water into the computer than was actually entered into the reactor coolant system. This would also cause a reduction in the calculated leak rate.
Q.33 What is the effect of taking advantage of the plant oscillations during the test?
A.33 (Kirkpatrick): In general, plant oscillations should have no effect on the test other than to increase the uncertainty in the n
.--_.m , ,
test results. The leak rate calculation is designed to account for all of the changes in the significant plant parameters. For .
example, an increase in the makeup tank level should appear as a reduction in another parameter affecting inventcry, such as the pressurizer level. However, if a water seal is present in the makeup tank level measuring system, the measured rise and fall of the makeup tank levels are exaggerated compared to the actual rise and fall. Under these circumstances, the apparent gross leak rate can be reduced by timing the test such that the test is started .while the makeup tank level is decreasing (when the measured level is lower than it should be), and the test is 4
ended while the makeup tank level is increasing (when the measured level is greater than it should be).
Q.34 What is the effect of switching makeup tank level instrument i during the tests?
A.34 (Kirkpatrick): The makeup tank level detection system included two level detection instruments. One provided the signal for the strip chart recorder, and the other provided the information for the data acquisition systerr. The allocation of these instruments to the recording devices could be reversed by means of a switch in the control room, so that the first instrument fed the data acquisition system and the second fed the recorder.
t Due to errors in the instruments, they did not always read the same level. Under these circumstences, the calculated gross leak rate could be reduced by starting the test
with the lower measuring instrument connected to the data acquisition system , and switching the higher measuring instrument to the data acquisition system prior to the end of the test.
Q.35 What is the effect of incorrectly reporting the reactor coolant drain tank voltage?
A.35 (Kirkpatrick): Prior to mid-November 1978, the reactor coolant drain tank levels were not gathered by the computer from the data acquisition system. They were derived from level detector output voltages that were measured at .a patch panel and en- -
j tered into the computer manually by the operators. The com-puter then determined the quantity of water in ' the reactor coolant drain tank from these voltages. The calculated idenii-
.fied leak rate could be increased bp under-reporting the volt-age at the start of the test or by'over-reporting the voltage at the end of the test. This had the effect of reducing the calcu-lated unidentified leak rate because this is determined from the difference between the gross and identified leak rates.
Q.36 Were hydrogen additions to the makeup tank required to be recorded?
A.36 (Kirkpatrick): There is no specific requirement in Administra-tive Procedure 1012, " Shift Relief and Log Entries," for the Control Room Operator or Shift Foreman to log hydrogen added
. to the makeup tank. However, hydrogen additions were frequent-
- ly logged in both the Control Room Log and the Shift Foreman's Log.
Q.37 Were water additions required to be recorded?
A.37 (Kirkpatrick): Item 3.3.11 of AP 1012, " Shift, Relief and Log Entries ," requires the Control Room Operator to record the addition or dilution of RCS boron concentration in the Control Room Log. Since the addition of water or boric acid to the makeup tank . changes the RCS boron concentration, it is reasonable to conclude that water additions and boric acid additions were required to be recorded in the Control Room Log. Such a requirement does not exist for recording water or boric acid additions in the Shift Foreman's' log.
-, - e ---m
4 t
)
1 i
STATEMENT OF PROFESSIONAL QUALIFICATIONS OF DONALD C. KIRKPATRICK My current position is as a nuclear engineer in the Engineering and Generic Communications Branch , in the Division of Emergency Preparedness and Engineering Response, in the Office of Inspection and Enforcement. My main duties involve the development of calculational techniques for the inspection of nuclear power reactors, the evaluation of nuclear reactor events to identify significant generic -problems and recommend the necessary corrective action, and participation in the . review and development of proposed codes and standards.
I was a reactor inspector for startup testing ar)d operation, in the Atlanta Regional Office of the Atomic Energy Commission,' from 1970 to -1972. In 1972 I became a reactor inspection specialist in the Compliance headquarters office in Bethesda, Maryland and held that position when this office became the Office of Inspection and Enforcement within the Nuclear Regulatory Commission. ,
From 1961 until 1970 I was employed as an' drigineer for the Los Alamos Scientific Laboratory, in the power reactor development division. There I worked as a power reactor shift supervisor' and in the development and testing of nuclear fuel and reactor coolant systems. -
I graduated with a Bachelor of Science degree in Engineering Physics i from the University of Oklahoma in 1961.- Following this, I took 32 hours3.703704e-4 days <br />0.00889 hours <br />5.291005e-5 weeks <br />1.2176e-5 months <br /> of graduate course work in nuclear engineering from the University of New Mexico extension at Los Alamos.
[
Jared S. Wermiel Professional Qualifications Plant, Electrical, Instrumentation and Control Systems Branch .
Division of PWR Licensing-B Office of Nuclear Reactor Regulation I am a Section Leader in the Plant, Electrical, Instrumentation and Control _
Systems Branch in the Division of PWR Licensing-B, Office of Nuclear Reactor -
Regulation, U. S. Nuclear Regulatory Commission. In this position I supervise a group of engineers in perfonning technical reviews, analyses, and evaluations of reactor plant features within the plant systems areas of responsibility pursuant to the construction and operation of reactors.
I received a Bachelor oU Science Degree in Chemical' Engineering from Drexel University in 1972. Since 1972, I have taken courses in PWR and BWR System Operation, Reactor Safety, Fire Protection, and Systems Reliability.
My experience includes seven years with the Bechthl Power Corporation as a Systems Design Engineer engaged in the design of various nuclear power plant auxiliary and balance of plant systems. These have included cooling. water systems, water treatment systems, and fire protection systems.
I joined the Auxiliary Systems Branch, Division of Systems Safety, NRR of the Commission in March, 1978. Since joining the Commission I have performed safety evaluations on nuclear power plant auxiliary, secondary, and plant systems areas of responsibility for a number of operating nuclear power plants and license applicants. I have also reviewed various topical reports and pro-vided comments on proposed ANSI Standards dealing with various auxiliary and plant systems.
l l
4 g
-~.
2 I have had responsibility for the review of the following nuclear power plant systems and areas of concern: new and spent fuel storage, spent fuel pool cooling, fuel handling, service water, component cooling water, condensate storage, ultimate heat sink, instrument air, ma'in steam isolation valve leakage control, heating ventilating and air conditioning, portions of the main steam system, main feedwater, auxiliary feedwater, tornado and internal missile protection, pipe break protection outside containment, flood protec-tion, reactor coolant pressure boundary leakage detection, post-fire safe shutdown capability, and containment isolation.
I am a registered Professional Engineer in the State of Maryland.
I I am a Member of the American Institute of Chemical Engineers.
t e
e e
d I
l
. Attachment 1 1 1
1
, l l
l l
Om X
-oo-X v
MAK E-UP
- TANK n .
4
'X x
! _0 CAL MOUNT A-32
, X m.s .
LT- 7.
LT-l Q q j- }.
I
' ' ' ' I FL.EL. ~
/
- .-, --, 305'-O
////// NNNNN //// NNNN l
' I L I J I e J
' ;.' s
~
t I J i '
-COMBINATION VENT, DRAIN & EQUALIZING VALVE I -
REVISED PIPING ARRAAGEMENT 11-16 -7 8 0 Diagram reproduced from Ex. 26 to the Faegre and Benson Report.
i
. . _ . - , . _ _ - . . . . . - _ _ _ . . ~ . _ , . _ _ _ _ . . . . . _ __ . , _ , _ . _ . _ _ . ...-._,-,,...__......__..4 .
,n
'. Attachment 2
- sl o
SATE: 12/27/78 REACTOR COOLANT LEAKAGE TEST
' I t 1E : 8:10:11 SP 2301-3D1 M
NOTE: IF OPERATOR ACTION DECREASES RCS VOLtNE THE DATA ENTRY FOR THAT ACTION MJSTBE NEGATIVE p ...YOU MJST ENTER DEC. PT. WITH LEAKAGE VALUES...
b DESIRED INTERVAL (1-8 HOURS) 9 1 ENTER OPERATOR CAUSED CHANGES TO THE RCDT FROM DS 4 (2301-3D1)
,O t# ENTER OPERATOR CAUSED CHANGES FROM DS 4 (2301-3D1) 0 ENTER IDENTIFIED LEAKAGE FROM DS 3 (2301-3D1) (GR1) ,
9 0 ~
ENTER PRIMARY TO SECONDARY OTSG TUBE LEAK (GPM) 0 D
TIME TCA THA TCB liiB TAVE ~PRZR LVL MJTK LVL RCDfLVL 3 (F) (F) (F) (F) (F)* (IN) (IN) (INCHES) 8:10:35: 556.945- 605.391 557.930 606.008 581.563 294.492' ~71.816 76.090 3
9:10:35: 557.609 606.195 558.648 606.852 582.320 299.109 69.524 76.252 3
3 GROSS LEAK RATE (<30 GPH): 1.1540 GPM
] TOTAL IDENTIFIED RCS LEAK RATE (<10 GPM): 0.2024 GPM
) NET UNIDENTIFIED EAK RATE (<1 GPM):
0.9516 GPM OPERATOR: -
D APPROVED: Md . .
J J
8EST COPY AVAllABLE i .
3
~ ~ " ~-~ ----
STOP 0 - '
)
- Attachment 3 LEAK RATE CALCULATIONS The Unidentified Leakage is based on the technical specification defini-tion, which is "all leakage which is not IDENTIFIED LEAKAGE or CONTROLLED LEAKAGE." CONTROLLED LEAKAGE, however, is seal water flow from the reactor coolant pump seals, which does not leave the system. Also, the leakage limits are all given as leak rates, stated in gallons per minute. The Unidentified Leak Rate (ULR), then, is the Total Leak Rate (TLR, also called gross leak rate) less the Identified Leak Rate (ILR). That is, ULR = TLR - ILR ,
The TLR and ILR are calculated separately, as follows:
Total Leak Rate.
TLR = 7.4805 (AMrcs + AMpzr + 4 Mmut)/[60(A t)g res]
+ OCRCS/60(6 t)
Where, 7.4803 = the number of gallons in a cubic foot of water (gal /cuft),
Qres = RCS density, taken from a table and based on the average of the average RCS temperatures, calculated from 'the initial and final data sets (lbs/cuft),
At = the time between the initial and final data sets (hours),
OCRCS = Operator caused changes to the RCS, that is, water put into or removed from the RCS by the operators (gal) ,
A Mres , 6Mpzr and AMmut are the mass changes in the RCS, the pressurizer and the makeup tank respectively (Ibs) . These are developed separately next.
6 Mrcs = 10678 ((ircs - ( frcs)
Where ,
10678 = The volume of the RCS water (cuft),
Qircs= Initial density of RCS water (Ibs/cuit),
& frcs = Final density of RCS water (Ibs/cuft).
The initial and final densities were derived from a table of water density versus temperature, using the average reactor water temperature. The hand calculation sheets provided for the effect of reactor pressure in the determination of density. Ilowever, the table used by the computer was based on the density of water at a fixed operating pressure.
. 4 I
Mpzr = 3.208(Lipzr - Lfpzr)/vpzr i Where, ;
3.208 = the volume change in the pressurizer per inch of level change (cuft/ inch)
Lipzr = Initial pressurizer level (inches),
Lfpzr = Final pressurizer level (inches),
vpzr = the specific volume of the saturated water in the pressurizer (cuft/lb).
The hand calculation sheets used an average between the initial and final pressures in determining the pressurizer water specific volume. However, the computer program used a value based on a fixed nominal operating pressure.
O Mmut = 255(Limut - Lfmut) .
Where ,
255 = the number of pounds of water per inch of makeup tank '
level change (Ibs/ inch)
Limut = , Initial makeup tank level (inches),
[fmut = Final makeup tank level (inche's )'.
Identified Leak Rate.
ILR = [(Gf - Gi) + OCRDT + SGTL + OIL]/(606t)
Where ,
Gf = final amount of water in the reactor coolant drain tank (gal) .
Gi = initial amount of water in the reactor coolant drain tank (gal) . ,
d (These tank volumes were derived directly from a table of tank level versus gallons.)
1 OCRDT = Operator caused changes to the reactor coolant drain tank, i such as tank pumpout (gal).
SGTL = Steam generator tube leakage (gal).
OIL = Other leakage from the RCS that can be identified and quantified (gal).
l
. TABLE 1 1
EFFECT OF COMPUTER PROGRAM ERRORS ON LEAK RATE DETERMINATION CALCULATION OF ERROR ERROR IN LEAK
! ERROR TYPE , LARGEST CHANGE RATE (gpm)
RECORDED FOR ONE HOUR TEST 1.71 Failure to account 300 gal of water Note (1.)
fer density change added (300 gal.) (1.343-1)/60 min in water added ,
1.26 Failure to account 3 inches (prior to Note (1) for density change in 3/16/79) (3 in.) (73.33 gal /in) (1.343-1)/60 min.
RC drain tank 1.11 Failure to extend 0.5*F in T avg Note (2)
RCS density correction (0.5'F) (2.21 gpm/*F) abovs 582*F 0.27 Inc:rrect RC drain 4 in. in RCDT Note (3) Note (1) tank level to volume (from 74" to 78")' (302 gal - 290 gal) (1.343)/60 min.
ccny:rsion table .
Note (4) 0.19 Failure to account 80 psi in RCS pressure .
fer effect of pressure (0.884f/ psi) (80 psi) (0.1612 gal /#) -
chrnge ,
one *F in RCS T avg 0.07 Inccrrect RCS volume Note (327 ft(2)/10346ft)(2.21gpm/*F) 3 (327 cu ft error)
Note (4) 0.03 Inc:rrect makeup tank 5 in. in MUT level mass change rate 5(257f/in-255f/in) (0.1612 gal /#)/60 min. .
n A
. - _ _ . _ _ . _ _ _ _ _ _ _ _ _ _ _ . . _ _ _ . . _ .-_ _ _ _ _ _ - . _ _ _ _ . _ . . . ~ _ . _
TABLE 2 EXPECTED VARIATION OF LE'AK RATE CALCULATION DOE TO NORMAL VARIATION IN MEASUREMENT VARIATION NO.
CALCULATION OF VAR. IN PARAMETER BASIS FOR ESTIMATE IN TIMES VARIATION FOR LEAK RATE OF VARIATION MEASURE- VALUE (gpm)
MENT USED ONE HOUR TEST Note (6) Note (4) 0.18 RCS cyg. temp Repeatability =0.2% range O.115*F Note (6) 3 3 Range =520*F to 620"F 8 /8(0.115F)2(2587ft)(0.776#/*Fft)(0.1612 gal /#)
60 min.
VariationI.2%x100/f l O -
l-1 i
[2(1.44in)2(102f/in)(0.1612 gal /#) 0.56 Pressurizer Oscillation of 2.5" 1.44" 2 during measurement 60 min.
Lcval i
Variation =2.5"/ f
- 2. ./2(0.87in)2(257#/in')(0.1612 gal /f) 0.73
! Makeup Tank Oscillation of 1.5" 0.87"
-during measurement - 60 min.
Lcv21 Variation =1.5"//I~
2 Note (1) 0.26 RC Drain Repeatability =0.2% range 0.11" Tank Level .
/2(0.11in)2(1.343)(73.33 gal /in)
Range =0 to 92" - ,
60 min.
! Variation =0.2%x92/ C 0'.97 Combined . Square root of sum of - .- ['.(0.18)2+(0.56)2+(0.73)2+(0.26)2 Variation squares of individual variations l
4 Notes on Tables 1 and 2 (1) The ratio of the ambient water density to the RCS water density 3 3
= (62.31#/ft ) / (46.4#/ft ) = 1.343 (2) Effect of temperature change in RCS 3
= (Volume of RCS) (Density /*F) x (Gal per ft )/(time) (RCS water density) 3
= (10346 ft3 (0.776#/*F ft3) (7.4805 gal /ft ) = 2.2 gpm*F 3
( 60 min) (46.4#/ft)
(3) RCDT error = level change by correct tab'le 1.ess level change by computer table = (6558 gal - 6256 gal) - (6755 gal - 6465 gal)
(for 78"). (for 74") (for78") (for74")
3 3
(4) = Conversion from lbs to gallons = (7.4805 gal /ft )j (46.6#/ft )
= 0.1612 gal /#
(5) Each measurement is taken 3 times at one minute i$tervals (6) The RCS average temperature is derived from't'he hot leg and cold leg temperatures in each of the two loops. Each temperature measurement 3 represents one fourth of the RCS volume of 10346 cu. ft., or 2587 ft .
This results in a total of eight temperature values that are us~ed in the leak rate calculation, four for the beginning data set and four for the end data set.
e e
e
Attachment 5 Conceptua Diacram o _oo 2 Sea s To m To m
.)efector
_ m -
--+
Mageuph ;
_001 Sea Wi",rl makuptani m
q n fressure i A ..
~
- increasinq m
i To - Jo ,,
,)d eClor la Retts an (
i l
1 l
l 0 0 ) S e a, d r1
! maketta tank
.l
! L 7PCSSitre
, t L M. . . ..b J -
C,eCreasinC l
1 1
l )
4 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE PRESIDING BOARD In the Matter of )
)
INQUIRY INTO THREE MILE ISLAND ) Docket No. LRP UNIT 2 LEAK RATE DATA )
4 FALSIFICATION )
CERTIFICATE OF' SERVICE i
I hereby certify that ecies of " TESTIMONY OF DONALD C. KIRKPATRICK AND JARED S. WERMIET " in the above-captioned proceeding have been served on the followingr ' 3 deposit in the United States mail, first class, .
or, as indicated by an e-ter --k, by deposit in the Nuclear Regulatory Commission's internal ma. ,. .em, this 1st day "o f July,1986:
- James L. Kelley, Chairman
- Jerry R. Kline' Presiding Board , Presiding Board U.S. Nuclear' Regulatory Commission U.S. Nuclear Regulatory Commission Washington, DC 20555 Washington, DC 20555
- Glenn O. Bright Harry H. Voigt, Esq.
Presiding Board James W. Moeller, Esq. .. ~
U.S. Nuclear Regulatory Commission LeBoeuf, Lamb, Leiby & MacRae
, Washington, DC 20555 Suite 1100 1333 New Hampshire Avenue, NW Ernest L. Blake, Esq. Washington, DC 20036 Shaw, Pittman, Potts & Trowbridge 1800 M Street, NW James B. Burns, Esq.
Washington, DC 20036 Isham, Lincoln & Beale Suite 5200 Smith B. Gephart, Esq. Three First National Plaza Jane G. Penny, Esq. Chicago, IL 60602 Killian & Gephart Box 886 Michael W. Maupin, Esq.
216-218 Pine Street Hunton & Williams Harrisburg, PA 17108 P.O. box 1535 Richmond, VA 23212 i
Mrs. Marjorie M. Aamodt Mrs. Marjorie M. Aamodt 200 N. Church Street Box 652 Parkesburg, PA 19356 Lake Placid, NY 12946
- Docketing & Service Section Office of the Secretary U.S. Nuclear Regulatory Commission Washington, DC 20555 J/f5k R. Goldberg
- cting Deputy Assistant General Counsel 4
6
, , .