ML20237G663
| ML20237G663 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 09/08/1986 |
| From: | Stier E GENERAL PUBLIC UTILITIES CORP., MPR ASSOCIATES, INC., STIER, E.H. |
| To: | |
| References | |
| LRP-A-001A, LRP-A-1A, NUDOCS 8708140144 | |
| Download: ML20237G663 (114) | |
Text
_ _
,)
l
/ -
/A V.44)' _m.
\
wiy
.,,, l
'87 AUG 11 AU :21 :1 i
i'
.w n .
h 0 C I+, , .' f
- v. ,,
t
'" - )
i, i.
TMI-2 -
L u,
~
REACTOR COOLANT L o,1 INVENTORY BALANCE TESTING .
H
'3 ..
NUCLEAR l'.CULATDtf CD1dMit$105 ~ '
Docket h ___Q f.s, - M, ti Esh. - .f -.
In the mattee of . h_# UMPg l l i
IDE C Est D
$t#t.
Apr.:4 4at __._ _// RECEIVIS .
- l ,
l ADECTID ,
late -
ces , ta IsAIE ContrattOf i wtnem7 -,.,' .(
oner
,,,,s., 0 && -
I i!
1 i
f ,
I
)
I B708140144 860908 PDR ADOCK 050003DO ~ ~ ~ ~ " " ~ ~ ~
,.+
G PDR
l i
l 1
TMI-2 ;
I REACT 0R C00LANT INVENT 0RY BALANCE TESTING l
PREPARED FOR GPU NUCLEAR CORP.
1 BY EDWIN H. STIER INVESTIGATIVE STAFF:
FREDERICK P. DE VESA ROBERT T. WINTER i
SEPTEMBER 5, 1985 j VOLUME IV (B)
MPR ASSOCIATES, INC. REPORT Section IX Appendices A through J i
r g
I MPR ASSOCIATES, INC, TECHNICAL REVIEW OF REACTOR COOLANT-SYSTEM LEAKAGE AT TMI UNIT 2 MPR'- 875 SECTION IX APPENDICES A THROUGH J-i l
l June 1985 1
I i
i I
i Prepared for )
Edwin H. Stier l
)
1050 CONNECTICUT AVENut, N.W. WA6HINGTON, D.C. 20036 202 659 2320 i
M PR ASSOCIATES lPU .
IX. APPENDICES i A. Evaluation of TMI-2 Reactor Coolant System Inventory Balance Test i
B. Corgarison of Requirements and Guidance on TMI-2 Reactor Coolant Leakage Limits and Surveillance Methods i j C. Recalcula' tion of Flaw Size Associated with a Leak of One GPM at Room Temperature D. Evaluation of Makeup System Batch Controller as a Possible Source of Observed Differences Between Indicated and Logged Water Additions to the Makeup Tank E. Evaluation of Methods to Average Leakage Estimates by l Alternate (" Slope" and " Water Addition") Methodr.
F. Sample Calculation of Total Reactor Coolant System Leakage Based on the Rate of Makeup Tank Level Change
(" Slope" Method)
G. Sample Calculation of Total Reactor Coolant System Leakage Based on Summation of Water Added (" Water 1 Addition" Method) '
H. Summary of Specific Allegations of Water Additions Not Properly Included in RCIB Test Calculations and the Results of MPR Review of Those Allegations I. Summary of Specific Allegations of Hydrogen Additions to the Makeup Tank During RCIB Tests and the Results of MPR Review of Those Allegations J. Correction of Recorded Total Leakage Values in Individual RCIB Tests K. Reviews of Records of Individual RCIB Tests l
I i
i
- Appendix A EVALUATION OP TMI-2 REACTOR COOLANT SYSTEM INVENTORY BALANCE TEST 1 Introduction and Purpose The purpose of this appendix is to report the results of an evaluation of the TMI-2 reactor coolant inventory balance (RCIB) procedure. Specifically, analyses were performed to assess the sensitivity of the results of the procedure to the following: i l
modeling approximations inherent in the procedure; instrument inaccuracies, and oscillations in plant conditions, i
- 2. Description of TMI-2 Reactor Coolant Inventory Balance \: i Procedure i
)
The TMI-2 RCIB procedure (Reference 9) was used to l obtain the leakage rate from the reactor coolant system. Leakage rate values were compared with limits established by the TMI-2 technical specifications. As discussed in Section III, the purpose of having these limits in the technical specifications is to ensure the plant complies with regulatory requirements during i operation. These regulatory requirements apply to leakage from the " reactor coolant pressure boundary,"
______--____--_______'_-____________ _ - _. . _ _ - _ _ _ _ _ _ = _ _ _
which is defined in Reference 10 as "the reactor coolant system plus those systems attached to the reactor. coolant system out to and including:
the outermost isolation valve, I
the second of two valves normally closed during normal operation, or
- I the pressure relief valve (s)."
]
l Figure A-1 shows schematically the TMI-2 reactor coolant system and those systems attached to the reactor coolant system. The reactor coolant system '
J includes the reactor vessel, two hot legs, two steam l generators, four cold legs, four reactor coolant pumps l
and a pressurizer. For simplicity, only one of each of the major components is shown in Figure A-1. The systems attached to the reactor coolant system include l the decay heat / low pressure injection (DH/LPI) system, the core flooding tanks, the letdown / makeup /high pressure injection (LD/MU/HPI) system and the reactor coolant drain system. Figure A-2 shows the extent of the reactor coolant pressure boundary based on the regulatory definition. All of the components within l
A.2 l
l l
I
the heavy dashed line are part of the reactor coolant -
pressure boundary. Accordingly, leakage from these components is subject to control by regulations.
The fundamental basis of the TMI-2 RCIB procedure is conservation of mass within a closed system, or
" control volume." The boundaries of the control volume l
t are determined by the procedure. Ideally, this boundary should be identical to the one shown in i' Figure A-2, i.e., the boundary to which the leakage regulation applies. However, the boundary for the i
i TMI-2 procedure, which is shown in Figure A-3, had a greater extent than the boundary required by i
regulations (Figure A-2). This approach meant that the l
TMI-2 leakage values determined using the control volume shown in Figure A-3 were greater than or equal to the actual values of leakage from the reactor coolant pressure boundary.
The control volume used at TMI-2 (Figure A-3) results in a more accurate and reliable procedure than can be conducted using the control volume of Figure A-2 because of flow across the boundaries. In the control volume of Figure A-2, there is normally flow through the lines connecting the LD/MU/HPI system and the reactor coolant system, and also a small amount of flow A.3 l
.j i
through the lines connecting the reactor coolant drain system and the reactor coolant system. The control volume shown in Figure A-3 is normally closed, which makes it more suitable for the RCIB procedure.
Occasionally, a " batch" transfer was made to the control volume during the RCIB procedure, (e.g., vetor addition to the makeup tank). These " batch" transfers could be easily _ accounted for in the mass balance..
For the control volume shown in Figure A-3, conserva-tion of mass states that for any two points in time:
i (Change in mass of water in reactor coolant system, except pressurizer)
+ (Change in mass of water in pressurizer)
+ (Change in mass of water in LD/MU/HPI system)
+ (Change in mass of water in reactor coolant drain system)
+
(Transfers in or out of the control volume)
+ (Leakage)
= 0 l
l l
Theoretically, the above acuation is exactly true, and given the values of the first five terms a value for leakage (the sixth term) can be determined. This was essentially the approach used in the TMI-2 RCIB proce-dure, where measurements were used to determine the values of the first five terms. There are inherent A.4 ;
l l l i
- i
1 limitations on the ability to precisely determine each l l
term, and this results in a potential error when the equation is used in this manner. This error is ;
l i
discussed in this Appendix.
The leakage which is determined using the method described above is subclassified according to " types" of leakage before comparison with permissible values.
Specifically, allowance is made for certain "identi-fled" leakage which is known and poses no threat to safe operation of the plant. The total identified
\
leakage is compared to a separate, higher permissible value than the residual, " unidentified" leakage, i.e.,
ten gpm rather than one gpm. It should be noted that the fourth term in the above equation (change in mass of water in the reactor coolant drain system) is included in the total " identified" leak rate, since the 1 purpose of the reactor coolant drain system is to collect several small known leakages from the reactor coolant system. Accordingly, the procedure as used at TMI-2 is perhaps most accurately described by the above
) equation written in the following form:
(Change in water mass in reactor coolant system except pressurizer)
+ (Change in water mass in pressurizer)
A.5 1
l l
_ _o
L
! +
(Change in water mass in MU/LD/HPI system)
+ (Change due to transfers)
+ (Change in water mass in reactor coolant drain system) i
+ (Other identified leakage) '
+ (" Unidentified" leakage)
= 0 i
The sum of the, terms: (Change in water mass in the i i
j reactor coolant drain system') and (other identified i leakage) is the total identified leakage ' for comparison - )'
with a permissible value.
It is also noted that the
" unidentified" leaxage defined above is actually unidentified plus intersystem leakage as defined in NRC regulations and guidance- documents (See Appendix B). _
At TMI-2 the following measurements were used to deter-mine the values of the terms in the leak rate equation:
I Reactor coolant temperature in the two hot legs; Reactor coolant temperature in two of the_ cold legs; Level in the pressurizer;
, Level in the. makeup tank; and I
Level in the reactor coolant drain tank. '
A.6
i i
These measurements were made at two different times, l
normally one hour apart. The measurements from the two !
I times were used to determine the total amount of I leakage during the interval between the measurements. l
)
The total leakage was then divided by the time to obtain a leakage rate. This entire process of gathering the measurements and performing the leak rate l
calculations was normally performed by the plant computer, although it was possible for the operators to read the values and perform the calculation manually.
The computer method was used in all but a few isolated cases; accordingly, the computer method is considered in the evaluations of this section. However, the following discussion applies to the manual calculation except as noted.
I i
- 3. Modeling Approximations Inherent in the TMI-2 Procedure The following modeling approximations, or " method l errors" were identified in the TMI-2 RCIB procedure:
Changes in pressure were not accounted for from the beginning to the end of the test in determining the changes in the density in the reactor coolant system and pressurizer. l A.7 i
. _ _ . _ _ _ _ _ _ _ _ _ _ - l
t I
I .
The reactor coolant density was determined using l
the average of the hot leg and cold leg l temperatures. This average temperature is not necessarily the one corresponding to the true mean reactor coolant.sys em density.
Operator-caused additions or removals from the makeup tank, the reactor coolant drain tank (RCDT), or any other part of the system were not converted to volumetric units at reactor coolant system conditions.
Leakage into the reactor coolant drain system was not converted to volumetric units at reactor coolant system conditions (This was corrected by a change which took effect March 16, 1979).
Identified leakage other than that collected in the RCDT was not. converted to volumetric units at-reactor coolant system conditio..s.
Changes in the reactor coolant system volume with pressure were not accounted for.
A.8 l .
l i
i I
I'
- l Some constants used in the calculation of mass changes in the reactor coolant system, pressurizer, makeup tank, and RCDT were known with s only limited certainty. I The changes in mass in the steam space of the pressurizer were not accounted for.
1 i
The interpolation table used in the computer I i
program to determine reactor coolant system density did not correctly calculate density above 582*F. For temperatures above 582*F, the density l at 582*F was used. l l
l The level versus volume relationship used to !
establish the volume collected in the reactor l coolant drain tank was slightly incorrect. i These modeling approximations potentially lead to results which differ from the true value. The amount i of deviation is dependent on the particular conditions for each test. In general, it can be stated that as l
l long as the measured conditions (temperatures, levels, i
pressure, etc.) do not change from beginning to end of test, there are no additions to or removals from the
~
A.9 E_ ,
reactor coolant system, and the identified leakage is small, then these errors all tend toward zero.
Essentially, these approximations represent inherent limitations in the ability to account perfectly for changes in conditions which occur during the RCIB test.
Calculations were performed to determine the effect of the above approximations based on a per unit change in the key conditions. The results of these calculations are summarized in Table III-1. Table III-1 gives the l " sensitivity coefficients" of the leak rate test to
)
each of the sources of error mentioned.
i Using the sensitivity coefficients in Table III-1, analyses were carried out to determine the magnitudes of the errors that may have occurred in the TMI-2 RCIB tests. This was done by evaluating the. actual changes in the key conditions from beginning to end of test.
In most cases, the results of the RCIB tests, as printed by the plant computer, were used to deteciine these changes. For RCS pressure, which was not printed by the computer, sample time-history traces of reactor coolant system pressure were used to determine the normal change which occurred over a one-hour period. A A.10
summary of the effect of the method errors is given in i
Table III-2. The discussion below covers the evaluation of each term.
- a. E'rrors from a Change in Pressure Changes in reactor coolant pressure were not accounted for in the RCIB test procedure. This led to errors when the pressure changed during a test because of changes in the water density in the RCS, changes in the steam and water densities i
in the pressurizer and a change in the volume of "
the RCS, (items 1,2,& 8 of Table III-1). The net {
l influence coefficient of the leak rate to a I i
pressure change was 0.0216 gpm/ psi. From examination of time-history pressure data, ISO psi was used as the range of typical pressure changes in the RCIB tests. The resulting influence on the calculated leak rate was *1.08 gpm. This was a randomly distributed error (i.e., pressure did not systematically increase or decrease during RCIB j tests) with *1.08 gpm corresponding roughly to a two-standard-deviation limit. >
A.11 l
l
- b. Errors from a Change in RCS Average Temperature Changes in reactor coolant temperature were normally accounted for in the RCIB test. Errors arose from the fact that the RCS average tempera-I ture was used to determine the density for the entire RCS, and the value of the volume of the RCS (10678 ft3) was not a precise value (items 4 and 9 of Table III-1). The net influence coefficient from these two sources is 0.263 gpm/*F, most of which is attributable to the uncertainty in system volume. (Note: Based on our calculations of the actucl volume and consideration that some portions l of the system may not participate fully in a
, temperature change we estimated the "true" volume j
may be as much as 10% lower than the value of 10678 ft 3 used at TMI-2. A precise value, however, cannot be determined.)
A review of the actual RCIB tests at TMI-2 showed I
the temperature increased and decreased during the I tests with roughly a normal distribution. A temperature change of iO.6*F corresponded to the two-standard-deviation limit. This gives an error range of i0.16 gpm.
l l
A.12 !
i,
)
i
- c. Errors from a Change in Pressurizer Level Pressurizer level changes were nominally accounted for in the RCIB test. However, errors arose from the fact that the change in mass in the pres-surizer steam space was not accounted for and that the value of the pressurizer cross sectional area j was not precise (items 3 and 9 of Table III-1).
The net influence coefficient is 0.070 gpm/in, most of which is attributable to ignoring the mass in the steam space. l Results from the TMI-2 RCIB tests were analyzed to l
l determine the changes in pressurizer level which '
occurred. It was found that both positive and l
negative changes occurred but that there was a i l
bias to the positive changes. It is possible this result may be because tests with negative changes were more often among those not filed. The two-standard-deviation limit for the "true" normal distribution was estimated to be i4.0 inches. The resulting range of error in the calculated leak rate is *0.28 gpm.
- d. Errors From a Change in Makeup Tank Level Changes in makeup tank level were nominally accounted for during the RCIB test. Errors A.13
i attributable to a change in makeup tank level occurred as a result of uncertainty in the makeup tank cross sectional area and as a result of systemmatic deviations between the indicated and 3 true level change in the makeup tank level. The
'I former source of error is very small and is (
l neglected in this evaluation. The other source of ]
error is considerable and is discussed in Section l l
4.b(1) below. l l
l
.e . Errors From a Change in RCDT Level Changes in RCDT level were nominally accounted for as identified leakage in the RCIB test. Errors i
I occurred due to uncertainty in the level / volume 1
relationship in the tank and due to the fact that l t
the sensed level was not compensated for tempera-ture to obtain true level. This latter error is discussed in Section 3.f below.
The former source of error was investigated by J independently calculating the level / volume relationship in the RCDT for comparison with the relationship used in the TMI-2 procedure. The method of calculating volume of horizontal A.14
i I
cylindrical tanks presented in Appendix D of -
{
NUREG-0986 (Reference 8)'was used in the independent calculation. Also, it was noted there was a 2-inch offset between the measured level and the true level from the bottom inside surface of the tank, and this was accounted for in the independent calculation. The result of the 1
independent calculation gave a level / volume relationship very close to the curve.shown in the TMI-2 procedure. The identified leak rate 1 discrepancy was less than 0.1 gpm in almost all of the tests, but did approach several tenths of a gpm in a few of the tests with high identified leakage.
The relationship used in the TMI-2 computer program was a piecewise linear approximation (i.e.
interpolation); however, in the program the values of volume at level values of 72 inches-and 74 inches were higher than the smooth curve shown in the procedure by about 15 gallons. Below 70 inches and above 76 inches, the volume values in the computer program appear to match the curve.
This shortcoming would produce an error whenever the initial or final level was between 70 and 76 inches. The maximum error would be 15 gallons, A.15
when, for' example, the initial level was 74 inches and the final level was greater than 76 inches. A review of 170 documented leak rate tests between August 30, 1978, and March 28, 1979, showed that in 132 tests the initial and final levels were above 76 inches, and thus the error was zero on these tests. In 20 cf the tests, the initial and final levels were above 75.2 inches, giving an error less than 0.1 gpm. In the remaining j 9 tests, the initial level was between 72 and 75.2 inches. However, in 6 of these 9 tests, the level increase in the tank was very small, giving an error of virtually zero. In the remaining j three tests, the level change was significant and the error was near the maximum value of 0.25 gpr.. For the normal situation where the RCDT level woule go up during a test, this error (if l present) teiided to cause an understatement of identified leak rate and, consequently, an overstatement of unidentified leak rate.
- f. Errors from the Addition or Removal of Water from the System or the Collection of Water in the RCDT i
During the RCIB test, additions _or removals of water from the syetem were nominally accounted for. Also, as mentioned earlier, the water A.16 i
i collected in the RCDT was accounted for as identified leakage. Errors arose in accounting l
for these volumes of water because they were not at a temperature consistent with the volume changes determined in the remainder of the RCIB l
1 procedure. The measured inventory change in the RCS was determined at the average temperature of the RCS, while the water-additions, removals, and collection in the RCDT were determined at room temperature. Either the RCS inventory change should have been corrected to room ter/perature, or the water additions, removals and RCDT collection corrected to RCS temperature before these effects were combined. If there were no water additions or removals during the test and only a small amount of collection in the RCDT, then the cal-culated leak rate was reasonably interpreted as l being a value at the average RCS temperature.
This was the case during most of the early RCIB- l 1
tests, i.e., 1978 and early 1979. However, as the
- leakage collected in the RCDT became more significant and water additions more common, the effect of considering these terms at a different temperature had a stronger ef fect, potentially influencing the tests by 1 to 2 gpm. It should be i
A.17
noted that.in many of the tests with water addi-tions and significant collection in the'RCDT, these two terms tended to offset one another leaving a smaller residual effect. Faegre and Benson, (Reference 1, Exhibit 41) contains a breakdown of the effect of this error on many of the individual tests.
9 . Errors from Changes in RCS Temperature Above 582*F The computer program which performed the RCIB calculations incorrectly computed density when the RCS average temperature was above 582*F.
Specifically, the density at 582*F was used for all temperatures above 582*F. Thus, temperature ;
changes above 582*F were ignored. An error of l
2.49 gpm per*P of change above 582 *F resulted. In ,
i, a large number of the RCIB tests, neither the i l initial nor final temperature exceeded 582*F and hence the error was zero. Specifically, this was the case in 116 of 170 tests between August 30, 1978, and March 28, 1979. In the other tests, the changes above 582*F were usually a few tenths of a !
l degree, leading to errors up to about 1 gpm. In J
l A.18 .
i i
l one case (September 22, 1978), the change above q
582 *F was about 1.5' F, leading to an error of l I
several gpm. I 4 Instrument Inaccuracies
- a. Normal Instrument Errors The results of the reactor coolant inventory l 1
balance surveillance test are affected by the accuracy of the measurements of temperature, pressure and level used in'the procedure. As with I any procedure involving measurements, there is a limited certainty to which the "true"' values of j parameters are known. The uncertainty in the !
l measurements is propagated to an uncertainty in i the final result, in this case the leakage rate.
The influence of an error in a single measurement on the final result depends on the magnitude of the error and the " sensitivity coefficient" of the -I j result to that measurement.
Measurement errors come f rom many sources, as described below:
Nonlinearity: Most instruments, produce an output which does not vary perfectly linearly A.19 f
i-i
1 with the measured variable. The measurement is interpreted in a linear fashion, though, and an error results.
Hysteresis: Because of hysteresis, it is possible to obtain two or more different outputs for the exact same measurement condition, depending on the prior history.
This condition results in a potential error.
I Sensitivity to Effects Other than the Measured Parameter: With some instruments it is possible to obtain several different outputs from the instrument for the exact I same true value of the measured parameter, simply because of variations in "other" ef fects, e.g., change in the temperature around a level transmitter. This effect produces a potential error.
I Time-Dependent " Aging" of Transducer: Some instruments undergo a gradual change or
" shift" ~ in their output over long periods of time (many months, for example). At the time l 1
A.20 8
l l
l I
a measurement is made, any cumulative shift since the last calibration is a potential i
error.
l Limited Sensitivity of Instrument: It is ~
j possible to obtain measurement errors as a result of an instrument not responding to a small change in the measured parameter. i ll Calibration Uncertainty: Some instruments L are calibrated in-situ after their installa-tion to ensure the measurement is positively related to a known plant parameter, e.g.,
level measurements which are calibrated in-situ against actual tank level as determined i using a sightglass. In these cases, the errors in the reference measurements used in the calibration and in the calibration itself will carry over into the measurement.
Time Response Errors: Under transient conditions, the output indi.cated by an L
instrument can deviate from the true value as a result of the time response capability of i
A.21 i
the instrument. This produces an error which is dependent on both the nature of the transient conditions and on the instrument.
Instrument Application: It is possible to 1
misapply instruments such that the resultant measurement is not the " desired" measure-ment. Level measurements are particularly prone to this error; most level measurements are simply differential pressure sensors which convert the sensed pressure difference ~'
to a " level" based on an assumed liquid ,
density in the vessel. Error can result from conditions in the tank or reference leg being i different than those assumed. These errors can be difficult to quantify since a detailed "
analysis of each individual measurement is required.
Manufacturers specified accuracy information for instruments normally covers all of the sources of error described above except in-situ calibration, transient and application errors. Individual breakdowns of the magnitude of each effect are - ,
sometimes provided by manufacturers.
A.22 N
\
.: 9 A .
For the instrumentation used in the TMI-2 RCIB test, the sources of error discussed above were reviewed to determine which ones would affect the results of the test. It was determined that for all of the measurements except reactor coolant drain tank level and makeup tank level, the ,
nonlinearity,- hysteresis, sensitivity, and calibration errors would affect the test, but that the errors associated with dependence on other effects, aging of the transducer, time response, and application would not be expected to have an effect on the results. The reason the latter errors do not have an effect is because the test j results depend on the change in parameters during -
the test and this tends to exclude these -
sources. For example, the aging of a transducer -
might lead to an error in a measured value. How-ever, the error in the difference in two values over a one-hour time period would be virtually ,
nil. For the transient errors, the time period of the changes (principally oscillations) in the measured data is about 12 minutes (Reference 1).
At this slow rate, the instruments would be ex-pected to follow the changes with high precision.
P A.23 9
e . .
. 4>
For the reactor coolant (RC) drain tank level measurement, none of the sources of error could be excluded except the transient response error.
This is because the results of the test are determined not only by the change in level but also by the value of level; hence, the long term errorc are important. There is a potential application error with the RC drain tank level measurement in that the instrument senses differential pressure rather than true level; the measurement is " converted" to level based on an implicit water density assumption. Changes in the conditions of water in the tank or reference leg from the condition in which the system was calibrated lead to an error. This application error is discussed in below.
For the makeup tank level, it appears there was an application error which resulted in level changes being systematically overstated. This is discussed in Section 4.b(1).
Instrument accuracy information was reviewed to determine the potential magnitude of the errors identified for each instrument. Unfortunately, this information did not provide a breakdown of A.24
/
the error into various sources. It appeared the total error specified for each instrument over-stated the actual error which contributed uncer-tainty to the leak rate test. As an example, consider the reactor coolant system temperature i struments. The error specified for these instruments is i1% of range or 21*F for the instruments used in this test. If the short-term random error were really i1*F, then a time history plot of temperature would be expected to show a line whose width is of the order of at least 11*F. That is, repetitive measurements would randomly deviate from the true value, creating a
" cloud" of data about the true line. The TMI-2 reactimeter data used by Faegre and Benson (Reference 1) shows this random deviation is significantly less, on the order of 10.3*F. Some l
of this deviation may also be due to actual turbu-lent variations in the temperature. It was assumed, though, that all of the observed varia-l tion was short-term random instrument error.
Bas"d on the reactimeter data, the following errors were determined:
l A.25
f reactor coolant temperatures *
'0.3*F reactor coolant pressure *20' psi prest,urizer level *1.5 inches j makeup tank leve.1 l
- 0.25 inch 1 No such data were available for the RCDT level measurement. Information on this instrument indicates an accuracy of *0.5% of range or 1
- 0.125 inches. {
.)
Utilizing the propagation of error technique, a leak rate procedure arror analysis was performed using the uncertainty estimates described above '
l for the measured parameters. The results of this calculation showed that the random error in gross leak rate was *0.69 gpm, the random error-in identified leak rate was *0.15 gpm, and the error in unidentified leak rate was t0.71 gpm. Because of the manner in which the uncertainties were determined, these uncertainties correspond j approximately to two-standard-deviation (95%
confidence) values. The corresponding value calculated in Reference 1 for this level of confidence is *0.94 gpm for unidentified leak rate. In addition to the values discussed above, A.26 A
r -
s the potential application errors in the RC drain .
{
tank level and makeup tank levels could cause I
systemmatic errors, as discussed below.
I 1
- b. Special Sources of Instrument Error at TMI-2 l
This section discusses two instrumentation 1 application errors which were particularly -
important at TMI-2. One of these affected the makeup tank level and the other affected the reactor coolant drain tank level.
1 1
(1) Makeup Tank Level The problems with the makeup tank level instrument and the resultant errors are partially discussed in other pa'ts r of this report, e.g., see the Section IV.C discussion of the alternate methods (" slope" and " water addition") of determining actual leakage. '
The principal effect of the problems is that l level changes could be significantly overstated. The hypothesized eource of the l problem with the makeup tank level instrument is discussed in Faegre and Benson (Reference
- 1) It is related to the collection of water in the tubing leg of the level instrument A.27 l
l t _ _ _ _ - _ _ _ _ _ _
I which is intended to be dry. The collection of water potentially trapped a pocket of gas in the transducer and tubing. Figure A-4 shows a schematic picture of this cause.
The presence of a slug of water in the dry leg would make a shif t in the absolute value of the level reading. That is, as water I
collects, the pressure on the " dry" side increases and the indicated level in the tank goes down, i.e., it tends to read less than the true level. Consequently, an operator who was " keeping" the indicated makeup tank level in a certain level range, e.g., 60 to 80 inches, would actually be keeping the l level in a range higher than these values.
This is of some importance as will be shown below.
Assuming a slug of water has formed in the dry sensing line as shown in Figure A-4, this produces a situation where the level measurement is sensitive to the pressure in the tank gas space as well as to the level in the tank. Thic occurs because changes in tank pressure cause the slug to displace and A.28
)
)
i i
change the measurement " offset" created by l 1
the slug. Specifically, increases in l l
pressure displace the slug toward the
{
transducer and hence increase the indicated level. Decreases in pressure decrease the indicated level. The sensitivity coefficient of the level measurement to pressure depends j on the amount of gas trapped in the I
transducer and sensing line and the amount of l
l gas in the top of the tank. Specifically, l the sensitivity coefficient is greater if there is more gas trapped at the transducer and is also greater if there is less gas i trapped in the tank. This latter relation-s ship shows the importance of the fact that the actual tank level was probably higher than indicated, i.e., this reduced the gas space and helped accentuate thi~ pressure effect.
The tank pressure changes when gas is added to or removed from the tank, and also when the tank level changes, thereby changing the volume available for the gas. Assume first that no gas is added or removed. If the tank level increases, the gas volume decreases and l A.29 1
9
i A
1 the pressure increases. The level instrument will measure the true level increase plus an additional apparent level increase due to increasing pressure in the tank. The total measured level increase will exceed the true level increase; similarly measured decreases would exceed actual level decreases. The above effect occurred at TMI-2, as evidenced by comparing the makeup tank level increases associated with water additions to the i
i magnitudes of the water additions recorded in 1
1 the CRO log. The level rose more than would j
! be predicted based on the amount of water l
l added. This is discussed more extensively in Section IV.C, and the apparent errors for the March 1978 through March 1979 time _ period are shown in that section.
The effect of the makeu~p tank level instru- l ment error on the unidentified leakage calculated in th'e RCIB test depends on both the changes in the makeup tank level during the test and the error ratio at the time of the test. A severe example would be in late March 1979; the ratio of the indicated to the l
logged amount of a water addition was about A.30
_,._.______._._._-.-a-
i l
1.3 and the indit ted drop in makeup tank I
level (if no wate; were added) in an hour was about 12 inches. The indicated drop was equivalent to 370 gallons (30.8 gallons per inch) while the actual drop was about 284
! gallons (370/1.3).
i The difference of 86
)
.)
gallons is equivalent to an overstatement of
! the total leakage by 1.43 gpm. This over-statement was converted to a volume at i
l reactor coolant average temperature in the l
RCIB calculation, giving an overstatement of I I
l the total leakage and, consequently, the unidentified leakage by about two gpm. This is an example of a time when the effect was I
severe. During other times when the makeup tank level changes were smaller, tbs error was less. Also, when water was added to the makeup tank in the course of the RCIB test I (as was often the case in March), this tended to reduce the makeup tank level changes and, therefore, reduce the effect of the makeup tank level instrument error. !
The makeup tank level instrument error is i also discussed by the NRC (Exhibit 1). Both the Faegre & Benson and the NRC tend to show l
A.31 i
I J
a simplified version of the actual situation and caution must be exercised in applying these idealized models to the actual case.
The actual case is more complex than the 1
simplified models because the legs of the j "U-bend" are neither straight vertical runs nor constant in cross section. Therefore, the l
effect of pressure changes which compress or expand the trapped gas is highly dependent on the exact locations of the water / gas inter-
?
faces and is not readily predictable. The effect also becomes a function of the history of tank pressure and water accumulation, since this affects how much gas is trapped at the transducer. These complicating features make it difficult, if not impossible, to theoretically calculate the effect of water collection or its variation.
(2) Reactor Coolant Drain Tank Level There was an application error in the RC drain tank level measurement in that the i measured level was not temperature ]
compensated even though it was treated as a i l
true level in the RCIB calculations. This A.32 ' 'i
level.was measured by a differential pressure instrument which is sensitive to the head of water in the tank. This head is dependent not only on the level but also on the density, i.e., temperature of the water in the tank. At one temperature (the l
i calibration temperature) the measurement will, be a true indication of level but at other temperatures the measured level must be temperature-compensated iY the true value is I
/
desired.
i l
l For tanks whose surface area does not change l
with level the mass change in the tank can be l
i computed from the measured level change without compensating for temperature. l However, for tanks such as the RCDT, the true' t )
levels are needed to determine the mass change. Accordingly, deviations between the true and measured levels will give rise to errors. Unfortunately, RC drain tank i
temperature data were not available for the RCIB tests to check the magnitude of this error for each test. A " worst-case" was formulated in which the temperature in the i tank was 100*F versus a calibration A.33 b
a
l temperature of 50*F, and the measured tank level change was for 76 in, to 78 in. The error was found to be 0.02 gpm. Hence, this error appears to be very small.
i i
- 4. Effects of Oscillations in Plant
- Conditions Oscillations in plant conditions were observed at TMI-2, as evidenced by time history data. For example, a review of data from January 4-5, 1979, (See Reference 1, Faegre & Benson, Exhibit 37) shows oscillations in reactor coolant temperature, pressurizer level, and makeup tank level, which have a period of about twelve minutes and peak-to-peak ranges of about 1*F for temperature, 5 inches for pressurizer j i
level, and 2.5 inches for makeup tank level. The oscillations are not in phase but appear to have a constant phase relationship indicating they are likely all attributable to a common source.
One concern with oscillations of this type is that the time response of the instruments may not be rapid enough to completely detect the changes. In this case, because the oscillations are relatively slow, the instruments are responsive enough to very accurately measure these changes and the error attributable to A.34
l i
this source is negligible. The limiting instrument in this regard would be the temperature sensors, which are j normally mounted in thermowells. l Although MPR did not i review th'e specific details of these thstaowells, it-is noted'that the conduction response time of one inch of :
steel is about 20 seconds -- a small time compared to' the 12 minute period of the oscillations.
Another concern with the oscillations is that when three values one minute apart are averaged in the RCIB test a true value assignable to the median time is not obtained. This was investigated by examining a sine wave, and determining the maximum deviation of the i
average of three values from the median value, where '
the three values were separated by times equal to 1/12 of a period. The result was a maximum error of +4.5% of the peak-to-peak amplitude. Accordingly, maximum errors of 0.045'F in temperature, 0.225 inch in pressurizer level and 0.113 inch in makeup tank level could be expected due to this source.
Anot'nr possible error is introduced because the values l
)
used in the leak rate test are not instantaneously j sensed, but are valces previously sensed and stored in the computer. The " freshness" of the values depends on A.35 l
l q
I J
l the update rate and the timing of the tes' call with respect to the last update. The update rate ranges 4
{
from 15 to 30 seconds for the TMI-2 instruments. Thus, i i
the values could at worst be 30 seconds (1/24 of a i period) old. l The maxiloum error in one measurement for an offset of 1/24 of a period is 13% of the peak-to peak )
range. An analysis of the combined effect of this error and the previous one was performed. The maximum
{
l l total error if both effects occurred was found to be l 0
I
- 13.2% of the peak-to-peak range. This is a maximum j i
error of about *0.13*F for temperature, 10.66 inch for )
I I
pressurizer level and *0.33 inch for makeup tank -{
level. l These maximum values are about the same as the j typical variations due to instrument error discussed above. Thus, at worst the effect of the oscillations would cause an additional error of the same magnitude of the instrument errors, on a " typical" basis the additional error would be less. Since it is independent of the instrument erors, it would combine with them on random basis, producing only a minor effect on the overall expected error.
A.36
~ l WPA ASSOCIATES l
-F-co-s?-7 l e/2/85
/ -
. 4 a;
N HOT d LEG E
CONTAINMENT ra m
i e
DECAY HEAT / o l 1 LOW PRESSURE INJECTION v j SYSTEM REACTOR R.C. PUMP e VESSEL m
11 X C >
CORE FLOODING COLD LEG TANKS M : .
X 2 COLLECTED l I
X c LEAKAGE l v LETDOWN / MAKEUP /
HIGH PRES 3URE X =
COO INJECTION SYSTEM M J: 1:
REACTOR COOLANT WATER AND CHEM D N SYS M l ADDITION SYSTEMS l
REACTOR COOLANT RADWASTE )( ~
DISPOSAL SYS
)
1 1
i TMI-2 ;
SIMPLIFlED DIAGRt.', ;
OF REACTOR COOLANT SYSTEM i AND ATTACHED SYSTEMS FlCURE A-1 i
i
l .I WP2 ASSOClOTES r-so-ef a 4/2/05 l
\
j a:
= ~~g HOT g i
CONTAINMENT o 7ssssassa l ""
g ll l DECAY llEAT/ !o l d LOW PRESSURE INJECTION j
hv j w g ,
k REACTOR R*C. PUtiP u I l
q g j i
I VESSEL di 5 N
4 l
t = l I
l a ALL COMPONEN"' SHOWN h , I C J g !
WITHIN THIS 1 E ARE haslp( asssaq \ l l PART OF THE REACTOP $ g F OD NG COLD LEG y py TANKS sassm l
ramasssamI %
%:t l 1 M COLLECTED St .: LEAKAGE i l se LETDONN/FAKEUP/
HIGil PRESSURE M- 4 p INJECTION SYSTOP /
f
, uu u assanssssssad l l
ATER AND CHEM DRAIN SYSTEP1 ADDITION SYSTEtts i 1
1 I REACTOR COOLAFT RADWASTE )(
DISPOSAL SYS i
TMI-2 l SIMPLIFIED FLUID EYSTEMS DIAGRAM i SHOWING EXTENT OF REACTOR COOLANT PRESSURE BOUNDARY I
FIGURE A-2
CD0 0ssocAvts i
l I
4/2/89 f
l i
rI__
Il .
M __ TlHOT g I
CONTAINMENT g I g m n (o________d --
i "
l DECAY HEAT / ~~
F ^
1-il LOW PRESSURE l I l
INJECTION ~
I y I i SYSTEM ~I {t I
I -
REACTOR R.C. PUMP o Il VESSEL i
n
! i 4 ;
I L 4 ALL COMPONENTS SHOWN I J WIT!!IN THIS LINE ARE h__)a( r ._C ___
\ g INCLUDED IN THE REACTOR COOLANT INVENTOPY CORE BALANCE PROCEDURE FLOODING COLD LEG TANKS g ,
___________as er_s._____
u _ %
COLLECTED g M c LEAKAGE g si l
% LETDOWN /ttAKEUP/ LETDOWN
<!IGH PRESSURE "
I COOLERS INJECTION SYSTEM h___m .mWA m_av. l WATER AND CHEM ADDITION SYSTEf1S
" I REACTOR COOLANT I g
=P-t m +L_______________________J, TMI-2 SIMPLIFIED FLUID SYSTEMS DIAGRAM SHOWING BOUNDARY CONSIDERED BY REACTOR COOLANT INVENTORY BALANCE PROCEDURE FIGURE A-3
CPo cesoclAtt s i F-so-es-? 4 4/10/89 1
I I
i 4
i l 1 l
UPPER SENSING L'h* <
(CONTAINS A YlATEH I
, SLUG WHICH TRAPS g j l
A GAS SPACE AT THE HYDROGEN GAS TRANSMITTER) t l
l
- !!!!n lll*llll:ll!!!'iiji
....9 LEVEL & iiiiW ~ili TRANSMITTER .....
fi@i!!!![!);pi !!!!!!:
tillii. ..;g WATER ,
! :..... fh' - i
.' !!\llyijhili.ii 5Eb! ri;;: '~
c-- i!!!!!;..
l )
l [
r: :iM! ,gji@,_,_
' ~
QM j MAKEUP TANK LOWER SENSING LINE (WATCR FILLED) l l
I i
l l
TMI-2 MAKEUP TANK AND LEVEL INSTRUMENT SHOWING HYPOTHESIZED SOURCE OF LE!EL MEASUREMENT ERROR (WATER SLUG IN UPPER >3ENSING LINE)
FIGURE A-4
Appendix B COMPARISON OF REQUIREMENTS AND GUIDANCE ON TMI-2 REACTOR COOLANT LEAKAGE LIMITS AND SURVEILLANCE METHODS
- 1. Definitions of Terms f For the purpose of this report the major terms used c
l will be defined as indicated below, These are
\
1 1
consistent with the same definitions as currently in l use in the nuclear industry as included in the governing industry standard (Reference 5). It should be noted that these definitions are not necessarily consistent with the use of these terms in the TMI-2 technical specifications or the TMI-2 reactor coolant inventory balance (RCIB) procedure. They are intended, however, to be consistent with the NRC's regulatory guidance documents as will be discussed in part 2 of this appendix.
Unidentified Leakage This is leakage from the reactor coolant system into containment which .is unidentified as to its )
source and thus could be through a crack or a flaw in the reactor coolant pressure boundary *.
r The reactor coolant pressure boundary is as defined in Reference 10. See the discussion in Appendix A.
m.
- l Identified Leakage i
This is leakage from the reactor coolant system into leakage collection systems, or leakage from j l the reactor coolant system into _ containment which- !
has been identified as to its source, has been quantified, and is not through a crack or. flaw in the reactor coolant pressure boundary. l
- 1 Intersystem Leakage 'l This is leakage from the reactor coolant system l into other plant. fluid systems.
Total Leakage This is total. leakage from the reactor coolant system, i.e., the sum of the previous three quantities.
- 2. Leakage Limits and Required Surveillance Methods There are a number of documents which, in effect, )
define or explain the limits on leakage from the 1 1
reactor coolant system, the bases for those limits, and what type or surveillance activities are appropriate I and required. As a means to make a direct comparision among the various sources of requirements and guidance, Table B-1 has been prepared.
Table B-1 includes excerpts from those documents which I
l were in existence when'TMI-2 was placed in operation, 1 namely, l
4 B.2 L
I
i 4
NRC Regulatory. Guide 1.45,= Reactor Coolant Pressure Boundary Leakage' Detection Systems, May 1973 (Reference 2);
NRC Standard Review Plan, NUREG 75/087, November 24, 1975 (Reference 3); and TMI-2 Technical Specifications, licensed February 8, 1978 (Reference 6).
Table B-1 also includes excerpts from those documents which have been issued since the TMI-2 accident, namely, 1' l
NRC Standard Review Plan, NUREG-0800, Rev 1,
-s July 8, 1981,-(Reference 4) and Standard for Light Water Reactor Coolant Pressure l
Boundary Leak Detection, Instrument Society of
} America Standard ISA-567.03, 198/ (Reference 5).
In Table E-1 the pertinent material is extracted on the specific categories of: " identified", " unidentified",
B.3 l
u_.______.____._.. -r -- -
-l and "intersystem" leakage as well as general-requirements or guidance particularly important to TMI-2 practice.
- Several major conclusions and observations can be made from the comparison: 1 The TMI-2 technical specifications (and the NRC
]
l standardized technical specifications) do not have l a one-to-one correspondence between the limiting aonditions of operation and the surveillance requirements. That is, there is no consistent specification of the surveillance method (or !
l me thod s) which are to be applied to eac" leakage limit.
The guidance documents and the ISJ, Standard indicate that sump collection and airborne radio-activity monitoring are the primary means to monitor unidentified leakage. The inventory balance is evidently only a secondary means to I monitor unidentified leakage and is discussed by the NRC only in conjunction with intersystem leakage.
B.4
1 l
l I
Other than steam generator leakage there are no operational limits which specifically identify {
intersystem leakage at TMI-2 nor are there any in the NRC standardized technical specifications.*
Overall, the reactor coolant inventory balance (RCIB) test is only one of many methods which should be used to monitor the plant for signs of reactor coclant i pressure boundary leakage. It is not identified by the NRC as a primary means to detect unidentified leakage.
l l l
- 3. Temperature at Which Leakage Limits are Evaluated
- a. Background and Introduction The TMI-2 Technical Specifications (Reference 6),
Section 3.4.6.2, indicate that a limiting condition for operation is that the reactor coolant (RC) system unidentified leakage be less than one gpm. The technical specifications do not indicate a specific temperature or density which ;
is applicable to this one gpm limit. It is necessary to determine an applicable temperature (or density) for this requirement since the mass j
' *As pointed out by Faegre and Benson (Reference 1) the TMI-2 technical specification definitions, in effect, made intersystem leakage part of what was termed " unidentified" leakage.
1 B.5
- 1
.a i
p
~of water in one gallon' varies with temperature.
The operating temperatures in the reactor coolant
/ ]
system, the systems connected,to the reactor I coolant system, and the leakage collection systems i cover a wide range from about room temperature up to 650*F. The density of water varies consider- l ably over this range of temperature' - the weight of water in one gallon' varies frem 8.33 lbs at room temperature to 5.01 lbs at 650*F. Hence, it '
is necessary to determine a single temperature for <
.i all measurements so that the leakage measured or i calculated'can be compared to en unambiguous limit.
The TMI-2 RCIB calculation used the average reactor coolant system temperature for the particular test to determine the " gallons" of leakage. Since this temperature was usually about 580* F, the TMI-2 RCIB test results are reported in gpm based on about 5.99 lb/ gal rather than 8.33 lb/ gal at room temperature. Consequently, the TMI-2 leakage limit of one gpm at reactor coolant system temperature was equivalent to a limit of i 1
0.72 gpm at room temperature.
B.6 1
-j
_-_ ___- E O
~
The applicable temperature for the one gpm unidentified leakage limit was determined by:
l l
Evaluating the applicable Nuclear Regulatory Commission's (NRC) regulations and guidelines associated with establishing the one gpm limit used in the TMI-2 technical l
specifications. ,
l Evaluating the interpretation of the ,
regulations and guidelines by other nuclear !
plants operating in the same time period as TMI-2.
Checking to determine if there are any safety requirements or concerns which would uniquely 1
tie the one gpm limit to a specific temper- I ature.
These evaluations are described below. ,
- b. Evaluation of NRC Regulations Regulatory Guide 1.45, May 1973, " Reactor Coolant Pressure Boundary Leakage Detection Systems,"
(Reference 2) describes the methods acceptable to B.7
- __________,m_
1 the NRC for implementing the Commission's regula-tions with regard to detection and measurement of l
reactor coolant leakage. This document estab- I lishes one gpm as a " preferable" limit for un- :
identified leakage during operation. Quoting from Regulatory Guide 1.45:
" Uncollected leakage to the contaittent
! atmosphere from sources such as valve stem l packing glands and other sources that are not )
i collected increases the humidity of the con-l tainment. The moisture removed from the atmosphere by air coolers together with any i' associated licuid leakege to the containment is known es " unidentified leakage" and should be collected in tanks or sumps where the flow {
1 rate can be established and monitored during I i
plant operation. A small amount of uniden- a tified leakage may be impractical to !
eliminate, but it should be reduced to a !
small flow rate, preferably less than one gallon per minute (gpm), to permit the leakage detection systems to detect pos-itively and rapidly a'amall increase in flow ;
rate."
8.8 i
l
1 Although the Regulatory Guide establishes that one gpm is an acceptable limit for unidentified leakage during operation, it does not state a specific applicable' temperature for this one gpm !
limit. The regulatory guide indicates that the t
unidentified leakage "should be collected in tanks or sumps where the flow rate can be esta-blished." That is, the measurement of flow is where the unidentified leakage is collected. The design of TMI-2 was such that unidentified leakage would collect in the reactor building (RB) sump.
l The RB sump was instrumented so that the amount collected could be monitored. At this location, the water is at normal reactor building temper-ature (i.e., sbout 70-110*F depending on the time of year and the exact operating conditions inside !
I containment). Hence, any flow rate measured at this location would be in terms of gallons per minute at a temperature in the range of 70-110*F. By stating that one gpm is the preferable limit for a measurement made in the j j
sump, the Regulatory Guide implies that normal 1 i
reactor building temperature is the applicable i
temperature for the one gpm limit.
i B.9
$?
,3 4# t@#'
t O IMAGE EVALUATION ////gf 84 ,
/o N/74 Yk" '4Y#if'
/
y f}/
TEST TARGET (MT-3) f,
% /g4, 4<[4[)
'Y &g, l.0 l2 m Esa ij 'y 22g i,1
[m lll!E
- 1.8
'l I.25 '
l.4 1.6 1, I l
4 150mm >
< 6" >
- W+ h///? h p,,,,, % </ 4
%/; / %y f4
~ ~
The NRC Standard Review Plans (NUREG-75/087 dated November 24, 1975, and NUREG-0800 dated July 1981, References 3 and 4) provide guidelines for reviewers of plant technical specifications.
These review plans indicate that each plant is to have in its technical specifications an operating i
limit for unidentified leakage, (although one gpm is not mentioned as a specific limit). These review plans also mention the unidentified leakage should be collected in a sump and the flow monitored during normal operation. The 1975 review plan mentions (under Areas of Review): the following in connection with monitoring unidentified leakage:
"The primary monitors determine flow rates and flow rate changes to tanks and sumps."
The 1975 review plan further states (under Acceptance criterf_a) the following regarding monitoring unidentified leakage:
"At least three separate detection methods should be employed and two of these methods should be (1) sump level and flow monitoring B.10
and (2) airborne particulate radioactivity monitoring. The third method may be selected from either the monitoring of condensate flow from air coolers, or monitoring of airborne gaseous radioactivity."
Thus, the review plans, in accordance with Regulatory Guide 1.45, identify the sump monitoring as the principal method of measuring the flow rate of unidentified leakage.
i As L
mentioned earlier, the temperature in the sump is normal reactor building temperature. Therefore,
- it is reasonable to conclude that a litiit l
established for this measured flow rate is in trms l of gallons per minute at this temperature.
Finally, Instrument Society of America (ISA)
Standard ISA-S67.03, " Standard for Light Water Reactor Coolant Pressure Boundary Leak Detection,"
(Reference 5) states that " Leakage Rate" is the
" Leakage expressed in volumetric units per unit of time at 20*C and one atmosphere pressure." The NRC participated in the preparation and acceptance for issue of this standard. The designation of these specific conditions is in accordance with B.11.
4 4
i the previous discussion, i.e., 20*C is 68'F which is very close to the expected temperature in the reactor building sump.
The conclusion of the above evaluation of NRC regulations and industry standards is that normal ;
i room temperature, i.e., about 70*F, is the applic-l able temperature for the one gpm unidentified
'I leakage limit. When the reactor coolant inventory j balance procedure was used at TMI-2 to determine unidentified leakage, the leak rate was calculated in gpm at the reactor coolant system average i temperature and not corrected to room tempera-ture. The net result of this approach was, in l 1 effect, that an overly stringent application of the one gpm unidentified leakage limit was being used. S pecifi cally, the hotter reactor coolant system temperature was being applied to the limit rather than room temperature, which produced an i unnecessarily conservative reduction in the permissible leakage to 0.72 gpm rather than one ,
gpm at room temperature.
B.12
i
- c. Ev81uation of Interpretation of, Regulations by Other Nuclear Power Plants
, The NRC performed an evaluation which compared the leak rate surveillance procedures and results from TMI-2 with two other nuclear power plants for the 1978-79* time frame (Reference 7). The results of the NRC evaluation showed that two other plants l
t (Davis-Besse and Rancho-Seco) calculated j unidentified leak rate based on room temperature conditions for comparision with the one gpm limit. The approach of these other two plants is in contrast with TMI-2, where the leak rate based on average reactor coolant temperature was determined. The NRC evaluation indicated that for Davis-Besse and Rancho Seco:
l
" Appropriate values were used in converting net mass changes to leak rates."
Hence, the NRC evaluation indicates that two other plants operating in 1978-79 used the one gpm unidentified leakage limit based on room temper- ;
ature conditions, and that the NRC concurred that this was the " appropriate" approach. This further demonstrates that a more stringent limit than required was being utilized at TMI-2.
j B.13 1
1 1
l l l l
- d. Evaluation of Safety Significance of One gpm j Unidentified Leakage Limit i i
As identified in the TMI-2 FSAR (Reference 12) i I'
section 5.2.7, the concern with unidentified leakage is that, since its source is not known, it 4 could potentially be leakage from a crack in the reactor coolant pressure boundary. Accordingly, it is necessary to ensure that the one gpm leakage I I
limit corresponds to a possible flaw in the I pressure boundary which is well below the i
" critical" flaw size which might lead to j essentially instantaneous rupture of the reactor coolant pressure boundary. Analyses were l performed for TMI-2 to demonstrate this conclusion, and these analyses are summarized in j t
the TMI-2 FSAR (Reference 12), Section 5.2.7.4. I l
These analyses showed that the crack size which would give a one gpm leak is 15% of the critical crack size in the reactor vessel and 17% of the critical crack size in the cold leg piping. In l
these analyses, tae one gpm leakage was assumed ts l l ;
be at the reactor coolant system cold leg tem-I perature of 557 F (density of 46.4 lb/ft 3).
If 1
the leakage were instead considered to be one gpm '
at room temperature, a larger flaw would be calculated since this leak at room temperature 3 B.14 i i
)
i i
< l
would be greater than one gpm at 557'F. Hence, the fraction of critical crack size vould also be greater. The calculations in the TMI-2 FSAR were reperformed by MPR (Appendix C) assuming the leakage rate is 1.345 gpm at 557*F, which corresponds to 1.000 gpm at room temperature. .
These new results showed that the hypothetical crack would be 17% of critical size in the reactor vessel and 19% of critical size in the cold leg piping. These values are only slightly larger than the original values given in the FSAR.
Consequently,_the FSAR's conclusion that the crack size which gives a one gpm leak is a small fraction of the critical crack size is unchanged, regardless of whether room temperature or 557*F is considered to be the applicable temperature for the one gpm unidentified leakage limit. Hence, there is no safety concern that ties the one gpm unidentified leakage limit to a particular temperature.
In evaluating the significance of a particular leakage value, it is important to consider that the reason given by Regulatory Guide 1.45 for having e limit of " preferably less than one gallon per minute" was "to permit the leakage detection B.15 t
systems to detect positively and rapidly a small increase in flow rate." The Regulatory Guide recognizes that the one gpm limit has no special quantitative significance, i.e., there is no technical consideration which makes 1.00 gpm a sharp " threshold" between safe and unsafe. For 1
example, a conclusion that a plant with an unidentified leakage of 0.99 gpm is safe and one with an unidentified leakage of 1.01 gpm is unsafe is not technically justifiable.
l l
- e. Conclusions Based on a review of the applicable NRC regula-l tions and guidelines, room temperature, i.e.,
l about 70*F, is the temperature which is applicable to the one gpm unidentified leakage limit. This !
result is consistent with the definition of i leakage rate in ISA Standard ISA-S67.03 (Reference 5) which states that 20*C (68'F) is the applicable temperature for determining leakage rate. The NRC participated in the preparation and acceptance for issue of ISA-S67.03. Furthermore, the results of NRC evaluations show that nuclear power plants other than TMI-2 were operating in 1978-79 using room temperature as the applicable l B.16 l
l i
temperature for the one gpm unidentified leakage limit, and the NRC indicated that these' plants were using the " appropriate" method. Finally, there is no safety concern at TMI-2 which ties the one gpm unidentified. leakage limit uniquely to a specific temperature. Although the crack size analyses in the. TMI-2 FSAR were performed assuming 557'F as the. applicable temperature for the one gpm unidentified leakage limit, the conclusion of these analyses remains unchanged when room temperature is taken to be the applicable temperature, i.e, the crack size is still a small fraction of the critical crack size.
When unidentified leakage rate was determined at l TMI-2 using the reactor coolant inventory balance procedure, the value was determined at the average reactor coolant system temperature and not cor-rected to room temperature. Since the calculated leak rate was compared to the one gpm limit, this approach meant that a more stringent limit was being applied at TMI-2 than warranted by applicable regulations, i.e., the limit on unidentified leakage at TMI-2 was, in effect, only 0.72 gpm instead of one gpm.
B.17
OVERSIZE ,
DOCUMENT PAGE PULLED SEE APERTURE CARDS l
NUMBER OF OVERSlZE PAGES FILMED ON APERTURE. CARDS I --
1
? l APERTURE CARD /HARD COPY AVAILABLE FROM RECORD SERVICES BRANCH,TIDC FTS 492-8989 l
i
Appendix C RECALCULATION OF FLAW SIZE ASSOCIATED WITH A LEAK OF ONE GPM AT ROOM TEMPERATURE I. Purpose and Introduction l
l The purpose of this appendix is to summarize the results of repeating the calculation in Section 5.2.7.4 of the TMI-2 FSAR using a leak of one gpm at room temperature rather than at the reactor coolant system cold leg temperature of 557'F. The calculation in the FSAR determines the length of crack which would result in a flow of one gpm and then compares that to the critical crack size.
II. Calculation The method and terminology used will be identical to that in the TMI-2 PSAR Section 5.2.7.4, pages 5.2-21 to 5.2-23 (Reference 12) except the flow rate at 557'F will be 1.345 gpm rather than 1.0 gpm.
The relations are solved by trial and error. If a crack length of 1.69 inches is assumed, "C" is 0.845 inch.
l l
Then, using the FSAR relations and notation, l
1 V = jl'y (2-27-)C = 30400x(2-2x0.3)x0.845 G /2 11.5x106x1.414 j i
V = 0.00221 in A = 2CV = 2x0.845x0.00221 A = 0.00374 in 2 d = 2V = 0.00442 in.
l As a result, j I
U = (2200-15)x2x32.2x144 46.4x ~0.06x2.25 + 1. 5,
. 0.00442 .
U = 116.7 ft/sec.
l I
Then, I Q = UA = 116.7xC.00374x7.48x60 i 144 0 = 1.36 gpm.
The crack length in the cold leg piping is therefore 1.69 inch and the crack length in the reactor vessel would be:
1.625x1.69 = 1.83 inch l 1.5 l
These can be expressed as a percent of the critical flaw length as follows:
C.2 I
l Cold leg -
Critical crack length is 8.7 inches {
Percent of critical flaw = 1.69x100 = 19 percent ,
8.7
{
Reactor Vessel - Critical crack length is 10.9 inches !
Percent of critical flaw = 1.69x100 = 17 percent 10.9 i 1
l l l l l I
1 i l C.3
Appendix D EVALUATION OF THE MAKEUP SYSTEM BATCH CONTROLLER AS A POSSIBLE SOURCE OF OBSERVED DIFFERENCES BETWEEN INDICATED AND LOGGED WATER ADDITIONS TO THE MAKEUP TANK
- 1. Introduction l
The results presented in Section IV are based on the assumption that the makeup system batch controller (MU-12-FT) is accurate and that the makeup tank level instrument (MU-14-LT1 or LT2) is in error. That is, the cause of the difference between the logged amount of the water additions and the rise in the indicated makeup tank level is assumed to be a result of an error ;
in the makeup tank level instrument. It is also possible that the batch controller was in error; i.c.,
that the amount of water added was, in fact, larger than the amount registered by the batch controller. If this were the case, the " slope" method would need no correction, but the " water addition" rethod results would have to be edjusted to reflect the larger weter additions. The two methods would still be in agreement; however, the value of the total leakage would be substantially higher and, consequently, the l apparent unidentified leakage would be very much higher.
_ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . ___ ___ _ ._ _ _ _ _ - - .__ _ a
i
- 2. Estimated Total Leakage if Batch Controller is Incorrect If the batch controller were incorrect by the observed ratio of indicated to logged water additions, then the
" slope" method would need no correction for the observed ratio. Consequently, an estimate of the total leakage if the batch controller were in error, can be l made by using the uncorrected total leakage from the
" slope" method and also by multiplying the " water !
l addition" method results by the ratio of indicated to f
l logged addition. Figure D-1 shows linear least squares fit lines of that data for February and March 1979.
The similar linear least squares fit lines from Figure t IV-11, which represents the results when the makeup tank level instrumentation is assumed in error, are l
shown for comparison.
l
- 3. Estimated Unidentified Leakage if Batch Controller is Incorrect Since total leakage is increased when converted to gpm at reactor coolant system temperature conditions, the unidentified leakage would also be increased if the total D.2
I leakage estimates based on an incorrect batch controller in Figure D-1 were used. Figure D-2 shows the resulting " unidentified" leakage if the identified leakage (RCDT collection rate) from Figure IV-14 is k subtracted from the total leakage of Figure D-1. For comparison purposes the " unidentified" leakage estimate l from Figure IV-17 is also shown on Figure D-2.
l.
- 4. Conclusions As is evident from Figure D-2, if the batch controller were in error, there would have been an excess of unidentified leakage over the RB sump collection rate i i
of about 0.6 gpm throughout February and about 0.9 gpm i during March. This would represent a loss of about 850 gallons of water per day in February and 1300 gallons per day in March. In March the total l
amount of water lost would be about 35,000 gallons.
This water would have to have been lost from the makeup system or some system connected to the reactor coolant system which was outside the reactor building, since it l was not collected in the RB sump, i
Because there is relatively regular personnel access into the areas where this lost fluid would have had to be leaking, it would not be expected that actual D.3
I l
(
l leakage of this magnitude would have gone unnoticed for such an extended period. It would have been bxpected to have resulted in such indications of leakage as visible water drippage, water collection in sumps, or l abnormal activity levels. If the makeup tank level instrument were in error, however, there would be no such large, unnoticed leakage which would have to be 1
assumed. It was for this reason, that it was concluded (
that the more likely source of the error was the makeup tank level instrument.
It should be pointed out, however, that the results of the RCIB test at TMI-2 had no correction for either possible instrument error. Consequently, the value of l
unidentified leakage which the operator would have l obtained from running the RCIB test was most likely to have been the value plotted on Figure D-2 for the case assuming the batch controller was in error, i.e., the makeup tank level instrumentation was correct. As demonstrated by Figure D-2, this calculated unidenti-fled leakage value would have been greater than one gpm j at reactor coolant system average temperature (0.72 gpm at room temperature) essentially the entire period the plant was operating in February ar.d March 1979.
D.4
=
$I N
o1
) _
F o _
1 mI _
8 5 _-
~ _
E )
n R
NI E c .
U R -
T U -
E A T _ -
T R A _.
m1 E R _
A P E .
R P _
M M w .
K E _
T E A n c 1 T .
E E L G M _
A O .
L R O -
A E R a T V aI O A T T A S .
C M ._
\
R P _
\
ol G \.
\
T m
(
A _ 1 i \ nI 1I M .
P mJ G
( _
" i -
oI o _
l
_. . -=
- ~
~ .@ b .
4 4
O -
a -lm
,m
, WIC>I< -
_ ,i *.
l
t j
i i h
i ! i i ! ! l I/
- r t l l
'p !
, p/' !
l BATCH CONTROLLER l p l ERROR ASSUMED g ,
/
r --- M AKEUP TANK LEVEL l p ,
l ,/ ; , INSTR UM EN T j,# l j ERROR ASSUMED y_ .- ! .
W I l i l i i !
E , i i
l I I i .PN'rURE CARD j i >
! I l l Als Available On Aperture Card i!" 9
'5 9 '
( 1 '
6 to 14 18 22 26 kB 4 8 12 16 20 24 28 MARCH l
ESTIM ATED TOTAL LE AK R ATE FOR ASSUMED ERROR IN M AKEUP BATCH i
CONTROLLER COMPARED TO AN ASSUMED ERROR IN MAKEUP TANK LEVEL INSTRUMENTATION (FEBRUARY AND MARCH 1979)
FIG URE D-1 8708140144-D A l
u______
) a _
E E c aJ
- T -
R U
A
) _
E F T R A . .
G
_ R .
K A1 aI c E _
A R 8 P wI! _
_ E E 5 M . - _._
L V ~ E A _
_ E T _
_ D S R wI , ._
M
_ E C U O _.__
I F
RT Q - _. .
A O
_ l _,
T I
T A R R -
E ' _ ,
N
_ l
" T __
E MP A f
! I
_ l D PM E 7.L / !.
I G
( T M -
g l N P .
- 9 U oI G a
(
E T
)
A
)
E E RR O R T U A RN T w H _
R C RO A .
R T EI T K E A NC P _
F O
A BI E M E L E - _
L GI S L S T " _
D N R O S
E I
C M CI EME O - _. _
O
- . l X F UL P R
SL M -
I j ET _
N S OU T -
A ol' E( ART S J t
4 4 .
D NB M _
I P _
N OR r -.
G - _ _
U CR ( m ,
- o : ,
E m~ ~
f~ je V
O n E g I<-
.\ t
!I
l i i
t j
- B ATCH CONTROLLER l ERROR ASSUMED l --- M AKEUP TANK LEVEL ERR R ASSUMED (FIGURE IV-17 )
l
~ l COLLECTION R ATE j
, j ,y. -
' 1 (FIGURE IV-2)
~
--- d
(' ,m. i j
[
~ _._L _ -
- j H
_ I' L_
L~~~ _ j 1 j
1 a
~
. . l i,
i I I Ti ;
) '
! ! hPERTURE
- t j
'~ CARD y
~ J '
_ ] I I s i
- l , Also Available 04 Aperture Card i I i i l l- l 1
2
'5 6 10 14 18 22 26 1 23 24 27 4 9 12 16 20 28 1 l728
- COMPARISON OF ESTIM ATED l UNIDENTIFIED LE AK R ATES TO RB SUMP COLLECTION RATE FOR ASSUMED ERRORS IN MAKEUP BATCH CONTROLLER OR IN MAKEUP TANK LEVEL INSTRUMENTATION (FEBRUARY AND MARCH 1979)
FIGURE D-2 03 q
8 708 W1 l
Appendix E EVALUATION OF METHODS TO AVERAGE LEAKAGE ESTIMATES BY ALTERNATE
(" SLOPE" AND " WATER ADDITION") METHODS
- 1. Introduction This appendix reports the results of evaluating several methods of averaging the results of the alternete methods (" slope" and " water addition") to estimate the actual total and unidentified leakage from the TMI-2 plant records. Specifically, for the operating period from February 1 to March 28, 1979, this appendix shows the results of averaging the total leakege on e daily average basis, a three-day rolling average basis, by linear least scuares fits over approximately one week periods, and by lineer least sauares fits over approximately two week periods es used in Section IV of this report.
- 2. Variability of Observed values of Total Leakaae The " slope" method results are an average of a number of individual slope determinations on the same day. In -
these determinations en estimate of the variability was obtained by evaluating the observed variances of the 1
slope values and the values of the ratio of indicated to logged water additions and combining those variances l
l
to obtain an estimate for the variance of the reaulting daily total leakage estimates. The probable error in the water addition determinations using an adaptation of the instrument error analysis discussed in Appendix A was also estimated. Figure E-1 shcws the " slope" and
" water addition" data points with these bands of variance and probable error plotted with each point.
The bands represent about one standard deviation. As is evident from this figure, there is a substantial variability in the total leakage values. The bars showing the variability or error indicate approximately the magnitude of differences between averaging methods which would be significant.
- 3. Daily Average If the " slope" and " water addition" values for each day are averaged, the plot of Figure E-2 results. It is evident that plotting a daily average is little better i
than plotting the individual points as far as making the results more physically reasonable and showing the overall trend of the results as contrasted to the variabilities inherent in the methods.
) E.2 l
1 l
1
- 4. Three Day Rolling Average i l
l
{ Since the RCIB test results must be obtained at least every 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> (References 6 and 9), that period of 1 time is of some interest in evaluating the leakage results. Accordingly, the average formed from the combined " slope" and " water addition" daily averages l for a particular day and the daily average for the day l l
1 before end the day after the selected day were calculated. These " rolling" three-day (72-hour) averages are also plotted on Figure E-2.
t The three-day rolling average values appear to be well within the band of the probable error and variability. It should be noted, however, that where the individual data points imply a rather rapid change in leakage, for example, around March 16, 1979, the three-day averaging method softens the transition.
- 5. Linear Least Souares Fits Over Approximately One or Two Week Periods In Section IV the least squares fits were made of the
" slope" method daily averages and all individual " water addition" data over about two week periods. The periods were selected so that they did not bridge times when the data appear to have changed the general trend E.3 l
of their slope or their magnitude as determined by visual examination of the data. The same technique was applied over shorter periods, about one week. The specific periods used and the resulting linear fits of the total leakage data are plotted on Figure E-3. For comparison purposes, the linear fits for the longer, two-week period (Section IV) are also shown on this figure.
- 6. Comparison of Unidentified Leakage Estimates Figure E-4 shows a comparison of the estimeted j
unidentified leakage which is obtained by subtracting l j
the identified leakage (reactor coolant drain tank collection, Figure IV-14) from the various estimates of tot.1 leakage. The daily averages and three-day l
rolling averages are from Figure E-2, and the one- and i
two-week lineer least squares fits are from Figure E-3.
{
l-Note that the differences among the various methods (three-day rolling average and the one-and two-week i linear least squares fits) of averaging are geneally I less than the basic errors and uncertainties in the data. Consequently, the specific method used to reduce the scatter in the data is not particularly important as they all result in similar estimates of unidentified l 1eakege. l l
i l
E.4 I l . _ - _
! ** s j
l l MPR ASOCIATES ;
F - r - 9 3 - 17 1
- ' t8 5 )
1 I
l 7
I i
9-1 i
- i f
i t 1
l '
6 --- ,
+
i
- 8- ! )
LL ! i l l l
w !
l l CD I o 7- - r
' f' i
)
W d
/ i j -
l , I l cc r i ! l I j W a ,
C F- '
D 6-
< I I
,,- H c !
l l
~
< W ! i T I 1
, C 1 4 , i W s i
i i
1 '
1 l W i l l l 2 5- H ,
W i
- g s i
TT i r T' 8 i LU O
l j T _
2_y l 7l o C 3j i l
- -~*- --<
< 4- 1 r { .l 1 : -I ' ~. ; , i 1 ;
~
< C ,
p i j
',.. r
, ,i .T % 0 i,u+*-o 7 ,-+D-. T -
'.e J,
- - ,i-
~++1 i i +
r
-) , + . - " ~ &~ , .h
+ ,,N.
i , '
2 -t T j , C '9- .Li
- i i -
~
m 3- 1 l
i i T ; ' '
l
-fr , 7: !l I$ 1l -^-^$, -l
,i lj d ', i o 9 2 '
_ -w9 Aa e'.
c , i _,- ._ w i -
- + , r
[ '. , 3 {L ' ' ' 7 __a 41 + ~ l F-
< 9_
~
~
o .-
1-s-
'-W1i 4
W3-i-
m<.
- ~'ms i a -
O
~
l _
~
1- i l . !
' I l
. 1 0- 0 I i i I i 1
2 5
6 9
10 13 14 17 18 21 22
'*5 26 1
2 3 7 11 15 19 23 27 4 8 12 16 20 24 2 B:
FEBRUARY
i l
l T
l J l l l
T I
l , r i
l j;
i l vl'lI
,11 4 i i I
T .
. i
' ~*-
7 5I I l r r 7 d i,I , I l '!l l l ! ilT-*~('g4!Till1l __
fje . _ i -i '
,l 1
' ^
E STIM ATE D INDIVIDUAL lI _ ' ~
i 5.L! l 70 l *
- PROBABLE (0 WATER ADDIT!ON 7' '
r Tii >
~ 'i i ERROR BAND DETERMINATION i, ; ,i i i _ : : T + _Tg.I: b' ' -.-
.- 2 l' s j! S
- 1 i,
l l! LTiL!Q~ If{ i!
I~
+ ?1 ,[*15 l..g i
~'
I 1' oBsERvEo DAILY AVERAGE 4 os sloeE
['-Ifhl!f[>,^}llI;
(.1 DETERMINATIONS ww.
,.i.-
+
i 1
i o
Il- : _
d h- -
i ;
l l FT
, : .i j ,
i .
w JTI _
%PERTURE l CARD ;
- i i 'I I
i !
Also Avn11able On ;
's c
'9 30 17 '21 ': s g3 '
.A ;>c r t ii r " ' '-- l 7 11 13 ,, 15 33 19 g,23 27 4 8 12 16 20 24 28 g MARCH j!
l ESTIM ATED TOTAL LE AK R ATE FROM
' SLOPE' AND ' WATER ADDITION" METHODS SHOWING VARI ANCE AND ERROR BANDS OF DATA (FEBRUARY AND MARCH 1979) a FIGURE E-1 8 7 0 814 014 g - 0 Y
)
c .
oI c __
)
F _
1 s1 n\
c 8
5
~ )
E n
c1 E _
R R _
U U T T o E A A _
T R R _
ml E A E P
~
R P M M _.
K E E _
A T T ac ai E E M _ _
O L G _
O _
L A .f f ;h ew A
T R
E V
o c
R T -
_ . iY f
.,f.
l' _
b ; ,
1_
O A A an 'i;: l T
S M
{
- ;;b l [,3
,%d4llO,ggi
- f'
,f P 4t f C : ' ,
R al n G
(
- 6tNl
, t s
T A
a M _
P aJ G
(
o1 o _
= _ G _0
~ - a : =
~
=
G a- }-
n@xcMI<
1 i/
i -+-
1
_N., - .-T v l - + + --.e,.. r - +
s . <, n s 1 - - -cr +
~ ~ * ~ ~-+-
l l _>_
....o __ -.
l
~
+
- -+- DAILY AVER AGE l
~
+ ., # g_*c;g , OF SLOPE AND
,-c-+ -o %t ,
i WATER ADDITION p - - -c- - - + DETE RMIN AT!,0N S
._,2* o- - - l l - - . . , . i. ! ,
i
-'r-_,_ ! '
! ---> --3 -D AY R OL LING
(#~ D ;S - - l i l AVERAGE OF l +, j l l ! DAILY AVERAGES l
I l .
1 l ! i I
l i l
! l l l i i l l I
-l i
) t i
. I i y tg
' APERTURE 1 5
,0 ;13 1 7 21
'25 CAHit 2 0 3 7 11 15 19 23 27 4 8 12 16 20 24 2B edeo Available On M ARCH Aperture Card '
ESTIM ATED TOTAL LE AK R ATE FROM COMBINED ' SLOPE" AND ' WATER ADDITION' METHODS BASED ON D AILY AND THREE-D AY AVER AGES OF DATA (FEBRUARY AND MARCH 1979)
FIGURE E-2
/
8 7 0 814 0144 - 0 $
.8!
r
)
c -
oI c
)
F .
NI n
1 c 8 - . _
5 - _ _
~ . .
)
E mI E - _
R R .
U U -
T w T _. _
E A A - .
T R nl R -
o E _
A E P R P _
M M .
K E T
E T a A Al c _
E E M -
L G O .
\
A O _
\
L R R .
A E ai c T l
V A u l -
O A s
/ . i
\a _
I /
S M j 1 .
C P R oi G /
t (
_ / - .
T A -
a M
P a1 _.
G
(
o1 o _.
a
. . a "
~
a- G :-l. G 3"
mg-;oC>n '
( (
I i
I
/
?
l ,
l - i t
l ,
! I i ,
l , .
, , s-l I , - -- LINE AR LE AST i SOUARES FIT
/ ;
OVER ~1 WEEK l
/ , j m
l LINE AR LE AST l SOUARES FIT i OVER ~2 WEEKS (FIG URE IV-11) i r
i
! i i I.
I l ! -
i TI
'APERTUllE t
I
_ l I
5 9 '13 '17 21 25 {jkj{[I 3 7 11 15 19 23 27 4 8 12 16 20 24 28 =
MARCH Also Available On Aperture Card ESTIM ATED TOTAL LE AK R ATE FROM COMBINED ' SLOPE' AND 'WAYER ADDITION" METHODS BASED ON LINEAR LEAST SQUARES FITS OF DATA OVER APPROXIMATE ONE AND TWO WEEK PERIODS (FEBRUARY AND MARCH 1979)
FIGURE E-3 8.7 0 814 014 4 -d b -
i
'4PR ASSOCIATES
- -80-93-20 6/6/85 1
l 6 , ,
i l 8-i l i' l l 1 1 l ' '
7 -l 5- 1 l
- 1 ,
LL ! l I l
I o l >
w i
- l w l-w e 6- i s o e ,
l '
a
< l D 4 E w I
g i i
(
M o 5- w i ;
< e ,
a W C 2 '
l i w i J l w n 3- :
~
a C 4- l W 2 2 !
__ w o !
H o !
E T I
' i
& D ;
3}
z o .
E2 w C a r 2 g
< e j
i D 2 2~ O c.
- +- !
I O~ -+- #,
1-s s o 1 i -+
j g i
+ 0 ,- +
J
~
,g
, T-+- q ~+% s -t;- !
y ._ _ + -
g- ,
-+7+ =6 - ya p N I + " ' ~ ' -
0- 0 --l - we I
1 5 9 13 17 21 25 9 2 6 10 14 18 22 26 1 3 7 11 15 19 23 27 4 8 12 16 20 24 28 FEBRUARY 1
I.
em s%
I l
j i
l
! l I
I i !
i
!. i !
t l l 1 ' -+- BASED ON l
l !
l DAILY AVERAGE I
- ! i l ! ! l 0 BASED ON 3-DAY AVERAGE I !
k I
I --- BASED ON i LINE AR LE AST i l l !
i i ,
SOUA 9ES FIT.
OVER ~1 WEEK i
I .! BASED ON l
! LINE AR LE AST I SQUARES FIT
(
OVER ~2 WEEKS
! 1 i l !
l l I ;
l l l
- i 4~
i -+,
-*] ! j
~~
N? Ch-$_ ?
2I ~ ' APER'fURE
- j- 5 t2l~ l l l CARD 6 10 14 18 22 26 3 7 4 6 12 16 20 24 28 Also Available On MARCH Aperture Card COMPARISON OF ESTIMATED
, UNIDENTIFIED LE AK R ATE BASED ON DIFFERENT AVERAGING METHODS (FEBRUARY AND MARCH 1979)
FIGURE E-4 8,10 814 014 't -0 7
Appendix F l
SAMPLE CALCULATION OF TOTAL REACTOR COOLANT SYSTEM LEAKAGE BASED ON THE I
RATE OF MAKEUP TANK LEVEL CHANGE
(" SLOPE" METHOD)
- 1. Introduction and Purpose This appendix provides a sample of a calculation used to estimate the total reactor coolant system leakage using the '
" slope" method. The method is described in general in Section IV.C.2 and the results are presented. The portion of the makeup tank level strip chart record which was used is reproduced as Figure F-1.
This sample is for March 19, 1979, which is a period when the loss of water through the pressurizer safety valves was l
high. Consequently, there are a large number of individual
" slopes" to evaluate; however, their duration is relatively short -- about an hour. During periods when the amount of water lost f rom the systo.. was low the continuous slope could extend for many hours. This reduces the number of slopes used to determine the total leakage during a day; however, it does not necessarily reduce its accuracy.
- __ - - -__- - - a.- - - - - ~-_--- - - _ -. - - - _ _ . - - . _ . - - _ _ _ _ _ . - - . - _ - - J
- 2. Selection of Slopes for Evaluation The slopes are selected based on ~ their regular linear downward trend. Traces like those starting at 1435, 1525, and 2253 for example, were not used because their distorted shape indicates the plant probably was not in steady state. Short traces were also not used, for example, those l starting at 1730, 2000, and 2035.
- 3. Determination of Correction for' Makeup Tank Level Instrumentation Error As discussed in Section III of the main report, . at some times the makeup tank level instrument was observed to over-indicate changes in level. If no correction were made, the slope method would consistently over estimate the rate of inventory loss and bias the results toward high total leakage values.
The " correction" factor, or ratio, was obtained by comparing the amount of water logged by the operator to the indicated rise in the makeup tank level. The portion of the strip I
chart record reproduced in Figure F-1 has had entered across-the top at the appropriate times the CRO Log entries of water addition. Note that in this period the chart time F.2
_ _ ._ _____.m._ ___ _ -w
appears to be the same as the actual time; however, at some times the chart time was many hours different from the actual time.
The value of rise is obtained by extrapolating the slopes and reading vertically on the chart as shown on Figure F-1.
This should account for the water lost while the addition is in progress.
- 4. Estimetes of Total Leakage Table F-1 shows the starting times for slope and the cases of water addition which were judged to the regular enough to be used. As shown in the table a total of 5 uscable slopes were identified and 6 logged additions could be used. The average correction factor, 1.25 was used to correct the leak rate of 6.51 gpm to a corrected value of 5.21 gpm at room temperature.
F.3
)
4 8 ) 6 7 ) 7 7 )
( 9 2 0 6 2 4 3 3 M ( ( (
. P 6 6 6 6 6 G
E P
D R I H 6 ) 8 0 ) 6 4 )
S / 2 2 3 N 3 ( 1 3 ( 2 2 ( .
I 1 1 1 1 1 n i
o t
i o dt d
OE I T at
' A r DR D ro eh
)
K K OTE ED n t s 1
p a 1
E'I EA IAG TCG 2 5 8 2
)
5
)
1 1
2 1
2 1
2 3
3 g woo
- ML AIO ( ( n dt RDL 1 1 1 1 1 1 p1 e L N g2 ge F "E A E I gp L PT 1 5 oo o
BDO A I I 55 ll ns n
T "S E ' N 2. 6 =
1 u
r isf tn I
O) na iod E A' I4 =
=
1 5 al dl LM T( 9 5 ) ) 2 2 7 0 ) 52 u dl s PI I rg aag MT DL 9
4 8
3 5
(
1
(
8 1
8 1 4 3 0 4
D6 E( 61 oe AS DA G f r s SE AG GE OT=
IAE df eo iht08 D
E
/RT n 3 D A id r T EKR mo es A 2 5 ) ) 9 9 2 0 TA ri tl C 5 1 AEK er fapa IN DI N
6 1
2 1
( ( 5 5 4 1
3 1
CLA I E t e ep .auq I OLL t A d ereai st a ITL eialhdi O
OAT bhuuced tl gnt n I 'I 0 n af vr aeiri d T D 1' coe 1 e A T g RICED onyoni t
g EET iol node N o GRC tib DO E IL L ARE at aedea R0R r ai p e c GTA 0 0 7 t 0 0 2 0 E0R rl osae GIG 0 0 7 o 5 5 6 0 VNO ouel arv OD I D 4 3 2 N 1 1 3 3 AUC NDrSBTA A
) ) )) )
0 1 2 3455 0 (( ((( (
_ 9 1 3 5
E 0 t
u - s 0 0 5 5 6 7 e M 2 3 5 o 3 0 5 4 t I
T 6 7 7 b 0 1 1 2 o 1 1 1 A 2 2 2 2 N
r s r '
g {
e EARLIER
/.t]D tlooIADDQO('JS E AR J 9 bJOT REGUG R AUD U U v
l/ SIRE MOT UiG D. .. q --
e m
C %1 q%
O O (
w O O4 g, n %
O O 4t o ni J
r-5 0
F" '
'; Q f', & VMa -
=
2 L i ==
A c' y
- h --
, --~. A +-
hM ' \ '
N R, 7, \ 'Fh q
%mL- . -
p wIr= wu t -- r .
Eg. Y--
~
'%} W %+~
V
^ - - -
_m.1
~ r g ,.- -
.n, _ ;; -
- J M W -
._ g _m mW M
,M R T n1_ _a' 7 :
C
,na 1[
m b *M@
r --
A m .
l I
i w gf 1
/
/
MP3 ASSOCI ATES g C f n F -8 0 - 0 3 - 21 f ,,C f f 6/19/85 e
e - e o e d E i s
+ .o 3 -. .- v -
.a m o e, c, - w a e e 3 'o=
a o
o e
- g. - -
1e i_ o Jm t, b ie i_ c w i.n m_
o 2 - -
'D O g#
T 2 \ I '- --
~ +
o o v w a ~m o 8 8 o in 7, 'S m e3 m ~m .
N (.a gio 1 ss cmc cwm uns.isa ov m en. w w wo n mo- me w e
R ph en d
- t. ,
b^
y 3 c. m. .. a a rl=% .+ '
a
& I&x. l % i= % -i k l
.w
'M &
w_
F:
- N%Ne.
.,w Y.
6,0 s 50 il s --.x. * - -
W ; .,
- n
~
^7. p s
- TI n- c .
APERTURE yn CARD
- c -
- i
~
^
-r.= 'Also Available On
~9 = ~~ ~~R_ -Aperture Card
[
s av
" ~,
'pg;uyw Mt1 y
- ;n~g,5 :as " ' ' "
1R 7: _
- g. . ,.
TMI-2 M AKEUP TANK LEVEL CH ART RECORD FOR SAMPLE ' SLOPE' LE AK R ATE DETERMIN ATION FIGURE F-1 8.7 0 814 0144 'Og _
I 1
Appendix G -
SAMPLE CALCULATION OF TOTAL REACTOR COOLANT SYSTEM LEAKAGE. BASED ON SUMMATION OF WATER ADDED ]
(: (" WATER ADDITION" METHOD) l 1
- 1. Introduction and Purpose l
l l This appendix provides a sample of a calculation used to estimate the total reactor coolant system leakage using the
" water addition" method. The general features of me...od are !
described in Section IV.C.3. of the main report and the results are presented there.
This sample calculation is for the same time in March as the
" slope" method sample in Appendix F. Some of the same limitations apply. In particular, the' sample time period has a large number of water additions because the leaking pressurizer safety valves resulted in the need for a relatively large number of water additions. During periods when the inventory loss was low, for example, during October and November of 1978, the number of water additions was greatly reduced and sometimes an entire day would pass without a water addition. Under these circumstances the
" water addition" method values are potentially unreliable.
i l
l t
- 2. Selection of Periods for Evaluation l
In order to apply the water addition method it was necessary to select a period of time over which the additions were to be summed. The basic criterion used to select this period j l
was that it be " regular", i.e., have periods of linear level l loss and distinct rises which correspond to the additions.
The portion of the makeup tank level strip chart reproduced in Figure G-1 was used. Note that there are times when the i
plant appears not to be in steady state, for example, prior l 1
to about 1600.
In the sample case the period from 1600 to 2100 was selected for evaluation since it did not appear to involve any plant I
transients. The period starts and stops on an even hour because the plant parameters are available in the computer l generated 24-hour summaries at those times.
l l
- 3. Determination of Correction for Makeup Tank Level Instrumentation Error As discussed in Section III of the main report, at some times the makeup tank level instrument was observed to over-indicate changes in level. Although this was not as important in the " water addition" method as in the " slope" method, it could still distort changes in makeup tank level from the start to the end of the selected period. It was G.2 l
u_________._._____.________._____.______ _ _ . _ _ . . _ _ _ _ _ _ _ _ . _ . . _ _ _ _ _ _ _ . _ _ _ . _ . _ _ . _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . ,
i also important to know the correction factor when an unlogged addition was in the period for which the water additions were summed. In general, unlogged additions were avoided; however, there were. some instances where they were included in order to get a substantial period of time over which to sum the water additions. ,
i The correction factor was .obtained for the additions in the selected period by comparing the amount logged in the addition to the indicate rise in the makeup tank level. See the " slope" method determination in Appendix F. The value in this case was 1.25.
- 4. Estimate of Total Leakage Table G-1 summarizes the specific values of the water additions, the beginning and ending conditions, and any corrections used in sample calculation. The specific coefficients used in the compensation for inventory changes were derived from the values used in the RCIB test and 1re based on calculating gallons at room temperature.
G.3 t
_.____m._._..___---A
The . change in the mass, AM, is given by the expression:
{
a AM=+ (Sum of Additions)x(B.33 lb/ gal) 1
-(257-lb/in) x AL mu x (1/R) -
-(7.73 lb/ psi)x AP + (862 lb/F*)x ATave -
_.l
-(102.4 lb/in)x ALpr, '$
where: AL mu is the change in makeup tank level (in),
R is the ratio of indicated.over logged additions, A P is the change in system pressure (psi). ,
OTave is the change in.RCS average temperature (F*), and i
dL pr is the change in pressurizer level (in).
The leak rate in volumetric units at room temperature is then given by:
Leak rate = 6M gpm,
( o t)x(8.331b/ gal) 1 i
where t is the length of the time period (minutes).
l In this sample case, 6M=+ (1812 gal)x(8.33 lb/ gal) - (257 lb/in)x(9 in) 1.25
- (7.73 lb/ psi)x(2 psi) + (862 lb/F')x(0.6 F*)
(102.4 lb/in)x(-3 in) = 14052 lb.
I G.4 l
l l _- __ ____ - _ -
l 1 l
I since the time period is 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> (300 minutes), l l
l Leak rate = 14052 _ = 5.62 gpm.
300x8.33 l
I 1
J l
l I
l l
i G.5
~
.l
~l TABLE G-1 l SAMPLE " WATER ADDITION" METHOD j TO ESTIMATE TOTAL LEAK RATE '
I DATE FROM 24-HR
SUMMARY
PARAMETER . INITIAL VALUE FINAL VALUE CHANGE I Time 1600 2100 300 (min.)
RCS Average Temp (*F) 581.0 581.6 +0.6 l RCS Pressure-(psig) 2151 2153 +2 Pressurizer Level (in) 230 227 -3 j l
Makeup Tank Level (in) 60 69 +9 WATER ADDITIONS TO MAKEUP TANK TIME AMOUNT ADDED SOURCE Gal. at Room Temp.
1620 400 CR0 Log 1730 300 CR0 Log 1750 CR0 Log About 1900 (13.8 in) 277(1) 340 Est. from Chart About 2000 (14.0 in) 345 Est. from Chart 2035 150 CR0 Log Total Added Water = '1812 !
Note (1) Based on R = 1.25, Addition = 13.8x30.8 1.25 -l O
_ - _ _ - - _ _ ~ _ - . - - _ _ ._ _ _ - - - - - . . - . - - _ - _ . . - - . , - - . - - - _ _ - . - _ - . _ - - _ _ - -
o
. _ i
% g i O
q~7 W
- O O O O V roi I O o
-o m O
- F_
e
, - G H 4.9 A A. , _
l O f'n '-
e =
L.
? , -
3
~~~~*
a
,g 1 -
P o
?~h ' (', u J - - ,
,jR ,
_~~ -} -U ws 1,
a~
l - -- y D_
b __ _- m q_ r1
,_u urw __.
1' l ", '<
l__
+ L.-- -
+ _ ,
l~
l~ %'w h_h, _'.b_,J N. :- v_ , M
, Wn -
g _-- ~% m '
Mr- ,_,-_
1
,1 -.\ - ; __'
, - - , ,i '
m_ .1 N N-v'
- W
w-b' 7_A Il
- I' 3
-,g H O'f'
--, ,9 3, Im I w
M gA -m a IV
-m-- garw ,,
FyK: _ 12-3J dtr d
,, 1
' MipM e e mon 7.M I
i l
I l
1
?
MPR ASSOCI ATES F-80-93-22 .-,
6/19/85 )
i F 7 !
O O
O,a C ID W R.
g$.
- e. 2'-
Og $D W i 3
J - aM M ^ ly .
> 2Q 2' O ets
~? D I3 IIO EE cwoc coremrxsv,m_.cs. mwr*n tow vmx 1 m
G i_ .[ ,
' O i]
i l9 3
E-Nn cn W
+_
3 "
r .
% ~1 an _. .. ~j d Ik ; J_( it . __
- k. t- lih n
y IWWWik
- 2 u
mi l '%J Nkvg! ?
s, t--y wa w
- _v.,,
w 21
_ _v- ---
m_m e _-- v mn -c .
_, y tyg
~~
MPERTURE En-- m :i CARD 5 ,/ "
i
~
r Also Available on
~
~
Aperture Card
'~
- b. w , Ai ~
i
-..m- ._ ___
+.
'Ut4fE7D , _ .
=ed 4--
_.- ^
yy r n t a
= 00 3- + , y TMI-2 M AKEUP TANK LEVEL CHART RECORD FOR SAMPLE ' WATER ADDITION' LEAK R ATE DETERMIN ATION FIGURE G-1 8.10.8140144 -0#[
Appendix H
SUMMARY
OF SPECIFIC ALLEGATIONS OF WATER ADDITIONS NOT PROPERLY INCLUDED IN RCIB TEST CALCULATIONS AND THE RESULTS OF MPR REVIEW OF THOSE ALLEGATIONS This appendix summarizes in tabular form (Table H-1) specific allegations that water was added to the makeup tank in the course of an RCIB test and was not. properly included in the leakage calculation. This included those potential instances identified in the Faegre and Benson Report (Reference 1) and those i identified by the NRC in Exhibits 1 and 2. 1 The MPR conclusions are based on examination of the makeup tank level strip chart and other appropriate plant records. Those cases where the addition can be confirmed are identified and are discussed in further detail in Section VI of the main report.
For the allegations which MPR could not confirm, the factors involved are briefly summarized.
H.1
~
nbl o m d0 a to em4 o i e V hal w t un 3 y bi56 E tl ioaar ud rri a2 a n t04 i t t ul cr t
R s pnd on tf t y iorh at hsr e 0d o of t erh0 .
3 oeeo idbvlf R wki a d cno i ec P oot hoid dt ae ivde d ir r
vt s aai h h t l c a )h5t e t1 at d daef a
f ldy e,
M stde sem nk e nt e c eo6 r dok ah F t0 dd Ann. rfbr an oo l f i c 0da e nedc ef i ae ti n .
e p at et geac 5 O r0 eoi t oof gtd t oh d t y ra .gs gst as2 a1 a t n o a rc c t eg ri5 S h0 .t t we o l rr a a dl ot loi rpth0 T c dsa t sc ner u e les0 n4 e e cb f n pl s oum u t L t eel l att u air nae ek o pf o5 p ei1 a oant e neee U e r c r eckh nn siem e.t a dt o gf ao e aaToa S vdael af po eps wea r t eeh, c t t n eie wrm E eet r a Dl gsoc t t 3 nbrio ie t r t t sh R Dda D ef e Do cl c Das f D e f .ec E g f E o Fbop E n af5 E ar e odf F bdomm0 S
Mk ot en R nl sbi N sht R wc M
R ee O M R s Td0 r e2 N ca PMt nfde R
M n ei3 Ri e E ac t a y,ab R eael t3 a e I oil s ce e N I t nt sd F Fhh an IF nu .r fF c r1 f p IFi p . Fl .e I F nc . uor sb 0
N o ae N swco N eqme Nd .e , N nay Nf t al t N el are O pi e w d O Orcet O thh Og a l
sdae l r a c ph h TIO Cut r eio7l u Ce.ii cnni tt Ceel u fbbp Ce et c c so0 Cied sc e
Cre at s gl p
t C ege f bno0 t c A Tk df5c Oade1 n Taied O r dd Tf Oier o om Taf Orf o3 2 T ae Tn Oifh0 o i0m Tf8 o 5 O orh L
U NN ab0i NT4i a NDl pc NT ot0 NNTt NNot4 c N D1Isif O i2 t o0 1
C L Y A B C E DC4 o
- T EI8 DT 9 t l P
S I S n SE VU1 OJ o .
NT l
R 5)2 i t a O
I B PF1 O T n Ti Dd g
n TI S Y1 o Cd 0 TTR 8 R a 5 io AC i
t NNA H t t GR EEU M MR E X l a
u ss aa .
h c ht s
l a
u E2 UTBf c dt k 0
3 tcce c L-LI CRE OAF hhl cca h
e eon r na 2
0 h 5f
- f h ca hl 89 c 5 e ,t / c 6 e c0 c l AM DP E N 54 0t n 5
t nep 0
6 0
5 4
3 e 04 34 n
- T CD O 0Oi 0 egu 0 n 40 i C
l R // / ne r O /n 0 /
H I N F N 000 000 0
2 0 ak 0h a e
v
/
0 0 e 0t
/0 8 052 I I 5cm o E 222 1 6 4* 611 L CD B EE A P D .
T SU T R 3 j
a F L O P 8 t O R
OCN E9 l
o a
C R 1 I nC n YI T i RT ) 6I G 2 B id dn d d
e d
e d
e l
e w
d e
AO N U Y H t
larea e r
d u
d u
d u
d u
d u
MN H A x totere e l c
l l lc l
MS C M E
(
f t nt t
n n c
n c
n n c
n U N C F ,eup e. i i i i i SO I R
N 00 m 00 o t g 0 o o, N
t o
t o
N t
o t
N o
t N
o 42 c 9 NI N T
I D
D T d O .
A R n a
R Cdi0 n p0 d T.
D O nr6 e C g R P dn. . at r M n E E) eo sft i y Rt gi )0 o o db T gt r43 s l rt r A NE oa e42 al a . e on OC ll t 1 eamnt e co u at rt roag ei W S N N E t c ul wit eooiwn baht nt a rt i
F ER ba di w cth rd B E ,c eh0 ,ameuc ed O F dn dn0 rE30 erl ns ol p da n
& E ei o( 5 vf oia ur e.
E R d d c s2 o c d d d en rn e e leta lo dd e enw1h e w R( ae g roo scf oo g g g G g nnti g g g E ru d
j o
- i ht rt sat aC at j o j o
j o ibw sa g al A et ei riR ri F
t e ddg0 ar wd s an 0 nd 0haf sd 0 0 0 of5 Ni 6 uaI o. 2 cVoOa 4 6 6 P o 8 T
E RD 88 00 55 00 22 I1 00 M AN 55 55 00 00 33 55 22 T! TE 0 00 45 00 23 11 23 00 12 00 34 00 34 00 S
Y 9 5 e 1 2 0 5 E AR 9 9 l9 9 I9 9 09 T DA 1/7 1/7 /7 l9 1/7 /7 39 1/7 /7 39 39 39 39 39 A /E 01 01 U1 01 01 01 01 D OY W
)
. T ) ) ) ) ) ) )
O S 7 0 1 3 4 5 9 N E 0 28 31 41 51 61 71 T I0 ( 0 0 0 0 0 0
( ( ( ( ( ( (
l
v e i nl erdl ip2 al E dl e tk n epo.
d yai vtOe t
t e t ec a aa rl y2 t s R i al vcp ida dt ives st lhd t d s
.n .
gf ci naap eta s c t
t pl R eim a n y a yi ry a t aua P
pa synh. p ekat e . ea rmdd n l
t c
e i
rdTt un e .i t y p s e ,k e er h
W i t ec 5
du.
ee gk s lat ec
,h0 a ei o pt h pi r
r da
.b e tst se t ean l Me F hor 4 gam pi c a t o et e O hnt0 eet gp nm . g ict,f1 o2 omel co ir0 4 a ,t ri a d c
n prso ooet t
e b
i ahf S
T h
wn ,d l
nt o ob iht watd e te rbed oe l a i lff sef s gro not r xco u e0 L oyn e r gf n ival e bor .ay f . d 3l U
,i a a eep ,eaa te an m l iresy r1 a S
gtd wu gl cwtf a i eenea ubra n o0c inagth tdn aoo t v p a pd d ad of E
R Dneac Dbot e o Dninldh enc a0 r hnds na t Dl eo p f at Derpt se Dt i
,), ipy E eth0 Mtit3 E
M s t*a E ee Mthe0 0 oeda tt n d r
a E
Mk lso th a E pupa M
E c M e s1 B t S Rt s n0 9 R eet R t t r0 asase .h R nl ast ilooah t
h t Rt e/e N I a o Dre egdc
- o. l s h I csn I a I aage tt Flf F nie Flff t t etdsge F t rl t n F 2 cn F ei(m2lt OI Nf o O eh N erm O r u O Nf oeu bo np R ogerll a eio ed N O pn ea go Nl Oad3 0 o Nd t O h s T Cesrc Ceer Cee b I nf bh Cueons Cro3 s Ce rce cl e0 f ct ctha F ei he egnie ei0e c e0t A Taeh3 Oret6 Tf as Tast N mrdt vd8 Oounoa n6 Tk Oaoeui rm Tnr Oeeti m Tah3 Or t3 n L NTbo0 Oirn Oraot NDti NTbba CSdabh aF NNtldt Ncpat NT o2i U
C L Y A B C E DC4 T EI8 DT 9 S IS1 V
SE OJ U
NT R 5.3 2 O
I B PFt O T
)
d TI S YI e TTRB g 5 AC NNA H t
h c
g o
3 GR EEUX M MR E 0 j 1
1/
E2 UTB(
3 7
d e
(
0 d
e L-LI CRE OAF 7 0
d u
0 h
c h5 5
c1 d
l u h h l
c c I
AM DP E N 7 0 c n
0 2
0 g 3 o c
n 0 0 C T CDO 0 i 0 1L i 3
3 2
2 H I N R
N 0
/
t
/
0
/
0O t 0
/
0
/
F 2 o 2 5R o 0 0 E I I 1 N 1 1C N 9 6 L CD B EE A PD T SU T R 3 .
F L O P 8 g
o OC N E 9 3
L YI R1 1 O T R
)6I RT G 2 B d e
d e
d e
C d e
d e
d e
AO N UY I
H d u
d u
d u
n d u d d MN HA E X l c
l l c
i l
c l
u l
u MS CM( n c
n n d
e n c c
(
r n n UN i i i i F i i C e SO I R
N t
o t
o t
o 0t 5n t
o t
o t
o N N N 1E N N N T
I D
D T A R 5 d
n O 05 e R P n r E E) iy t R1 db T e r
on e l
A NE d co d a .
mn OC ro.
W S N N E l
e t
ei rt i
i s
t oin nt w F ER u rd u co B E o ed o med O F t da n t ol rld d
& E n . ur n . f oe e ee ee cp d R e d d d u E(
R rc an et . rc an ee gcTu m e g e l G pe l aa bwn pe gnDp g g
g c
pu pu o n E o oaC o i A al isat al jiRT D
j j F f s t f r t on of e on 5 afC 0 0 o Ni Pog Ni 7VoR 6 5 N T
E RD 66 99 8e 77 65 uAN t
144 00 33 00 33 89 66 33' 11 00 T TE 2 90 01 82 12 0t S 00 12 00 11 00 00 0O Y 2 8 6 3 1 9 9 E AR 09 29 29 29 29 9 19 T DA /7 39
/7 29
/7 29
/7 29
/7 29 1/7 29
/7 A /E 01 29 01 01 01 01 01 01 D OY M
) ) ) ) ) }
. T ) )
3 5 8 0 g 3 4 O S 2 02 1 2 2 3
3 3 4 53 N E 9 0
1 0
1 0
1 0
1 0 130 1 T ( ( ( ( 0
( ( ( (
5i m hai dll rst pg E e 4 tt yu r ctt ot t Iti s hapa4 c mc awn poae0 I
V c a
3i t 2de t
. ld r
toicr dea ude idaio pll m2 E nlt t ndt hexpt alPi1 e r ds s nod nil s ui a wgey . o R ry t . t a e ooe o o om B g T2k Cf t(
ol y a o t ps sech b dI hor a R .s R mev a nv P
f a od .ds l g ,
e g n
l ge anr eaTi BhDh c a5eC 1 gR y g 5cjo.rp f
ee Nnnin oihi M fif ags i 3a,ha- hi t F
ot ost a na goo r gic k e ddR Cw broe 5 2 o y5 yt c o ct4 4 ulh i1t O dll oh oeait 0n r ti d
u 0rd 0 o nl aul7 9 t ont chu t t a1 a-d7dk ihd wdt At u
irct 0U 2 wl oa/ e g as a 8
T.
r i cn eapeo 3 ,.d nl e t nv e ohm 1 rsr1 r1 en r hi ee ettdip 1
se e au.b srbea o
o t p yl f .i as uee g o/ o yT r cuhnh i eat .
u htf e er oes5 ga e b o pl een2 cb u abt at) rt ael S at al n rI aft re E
i ,De ni 0r a ,T Fbm n it .d t : n,ne. l wosev gwe h or t5c rD ao o oCt tohot a we R Dn p u1 e yeC Dft r DonRo Dft c a D uol y Ei so.to itt o d l vR ne to e
ft Et
,Nn Eoan0sg tdt edsl S N rrld R uasof ieb o dv atl ow s
e R2he e s N R esys N e0 n R g yk 5 o 0 s a e
R rseeon vank N I de F
io Dde E al sne e D Enhr o.
FI4Tc nt . I meba I nlf1i2 I gelti a O N o ht r nel M aoe Ro e Nt a e8 Fit Nt ud w
s nn o,t1 Fit s a F
N sun a t stt I
Ot2apa Rd .hil Ri sd Oi .ih0 O neh Oeeehuo O binup T C t c I en t I tt r Cbrrt8 Cti mc Ct rrctt C'e i it u A t nsei Tsanmp F gatiT N e sdD Fi ro Ndac l oa Thrv r 1
Ta rmu6 s3 t au0c 7apc0 u0 crp ce T at ot uk L Oehoiy NTt ctt Oo2 Ct 1 t aR edC Odhe CAcr Oar Nteaf a ot Oh0s3 0l pc4l6 s
O rt rol a U NC4 a0 a Fao1f( N Tadnf m C
L Y A B C E DC4 T EI8 DT S iS9 SE vU1 OJ 5)
NT R 2 O
IB PF1 O f l
d )
TI S Yl e d TTRB g e AC NNA H t
h 5
2 g g GR EEUN MMRE 5
c 8 1 jd o
( e g
o E2 UT8(
d e
5 2
/
0 d
e t j
(
L-LI CRE OAF l d
u 2
/
0 h0 e3 5
d l
u h
6 c
ha cc 5i h
C l
AM DP E N S
n 5
1 2 g 9 o c
n 3
3 3d 5 n 5
1
- T CDO 1L 0 1i 2 HC i i R r / / / 0 e 0O FIN N t t 0 0* /
J v 0R o 0 07 0 E I I N o 3C N 2 1* 6 L CD B EE A PD T SU T
R .
F L O 3 g OC P 8 o.
N E9) t YI R1 t 0 R
) 64 T RT G 2 B d
e d
e C d e
d e
d e
d e
AO N UY H I
d u
d n d u
d u
d u
d u
MN HA X l c
lu c
i l l l l M CM(E n n d
e c
n c
n c
n c
n U S
( i i r i i i i F
N CO e SO I R
N t
N o
t N
o 0t 0 n 3E t
o N N t
o t
N o
t o
N T
I D
D T A R d n
O e R P g
. r E E) t Rl o T l l a . e e e A NE o. g mn d d d OC R o. ro W SN NE Cn w
L n w
oi nt i
s t
i t
s l
t e
F ER no oo c u u u B E id Rd me o o o O F dd C
d ol rl t t t
& E ee ne f o n . n n .
ER d rp i p c ee ee. e e R( e em m e rc rc r c g t u np du cT an pe an pe a n G g ep nD p e E o e r aC pu pu pu A
j T eT iR al al al 0tD 0tO r f f f F 0 5 oC 0nC af on on on 5 1 NR 3ER vo Ni Ni NI T
E W
RD 33 00 88 11 00 22 55 l AN 55 23 22 12 00 89 44 23 22 23 55 34 22 T TE 01 00 22 11 00 00 11 00 S
l Y 6 2 1 1 9 7 6 19 19 19 19 09 09 09 E AR /7 /7 /7 /7 /7 /7 /7 T DA 29 29 29 29 29 29 29 A /E 01 01 01 01 01 01 01 D OY M
) )
T ) ) ) ) ) )
7 3 4 0 5 20 OS 7 3
8 14 9
14 24 17 24 25 32 25 N ET 1 0 0 0
( ( (
0 0 0 0
( ( ( ( (
v l e th d viet - n e - neea stibd erkb or l oit .
adt soae t roal rh nd t
_ E nbnd !C hduet g uot c l edtl s nt oe R e sR apt - sa pf - see euasee oawh s R
edeao bnv d n uN o
m t m odorah a k ek m k e gh l ot ahet e vq eenje uvt i r t eo T
necoece P aodd ue cneck v l sa l e ipt epnaye dO M e hee vdsoe ae gl nh ut h sch h
et ct a f pnl i
d hm cua iell lb n
gomenT l h d
a eso rt
- F O h eeob yy oat n 0e t r 0 ak incrvi i e .t f cl nb a dbied1t itd1dbn k fl p. ndut aep a wma wsa eoa or t a d a ren S sarrn d hf suhtl snhtt rt ot ok rh gmrat T
L u
ms t ea h
we og m 0 co r ott oc uu o
i uap t ol p wscnd/
e a n weooeu ol cf po U altd h e 3 et snbl . tl nbusu dt ot ei hl nweoe s! 2 s neaad eeacie e t h saht a W o.
_ S o v e ! n 0idim ci dvm l k hr pT2 cn E ri t e re uf eeuf asa coeu t st es a t reef R
ii ol m h d yk S b oe Do E ns n a Ii t
Dt E aft et N
tl oavsnt n roeal hv Dll rocim Ei NagsneTl t h Dtfl Ei Nvraarl bk e.s o D En o a wu N hsr im S
(i R s Rn a Reeemha R sc N _ Rdl n I at a .o F ct nt I eh F oc E d
,da I oe Fit r f
eno,i lii tt y Rf nnoh I
F rk iiit t .i c I Fruuf ns t s / I e9osni F o4 t ne O N .e Nd0 Nd1 i
- o Nt a o gd a ll N orda ,dp Nop oia I Ol r pi r O 0 8" d e
Ocrt nauub Ot oauti y Ot mt w. O Ce 1d el T Car ut a Ce1 Ce aibt sa Cawbtllt Caooe2 d coeie c0 g l erk ceb r cua - rcnp ti c0l et ca A T0 ckdp Ta g Tf mereuro Teeeusvn Te ossl Ta2i vsu L
U O) nadp NriMaa -
Ort NTa [l o Oeapohl NDsOwTf ap r O ph h l eno NOtTf rin Oposl aea Noti swt v Or1 aenl NT1flif a C
- L Y A
- t t
C D 4 T E D
8 )
_ S I V
9 5 SE O 1
3 2
N T R 5 -
O IB T I P
S 1 2 YI R B T
0 0 l
a AC T N"AH I
0 1
0 0
m r
h c
h c
GR
_ E2 E W RE UX 0
/
0 1
/
7 o
n 0
4 0
0 1 4 U B( h0 h
- L c1 LI-h s 1 0 CNEF 0 c3 h c 5 1 c * - -
O 3 g 0 0 g 0
3 3 0 2
0 5
A M D N 2 o. 01 01 Io.
2 1 1 3 O 01 0 +
C T C 0 1 0
/ / / T / /
R 0o / ? o 0 U 0 0 FIN I I N 0a 1c 0
6 a
c 0
1 (
N 1
0 0 1
CD EE PD SU T R 3 g g
F L O 8 o o P
OC E9I 3 r t
- N R g o a
o a
Y I ) 6l*i c d c d d d d R T G N
2 B n d e
n e
d e e e A O UY I
H i u i d
u u d
u d
u
_ M N H AX E d l
c n
d l
c l
c l
c l
c M CM( e e n n n n U S
_ ( r i r i i i i F e e N CO 0t t 7 t t t t SO t R 0 n o 1 n o o o o I N 3E N 1 E N N N N T
I D
D T l A R -
O T.
R P g 1
E E) o h Rl L t T e e .
e e e i A NE d d g o a
d d d w OC i l
- o. i l l e c W SN N E t
s u
t u
t o n s
t u
t e
u e
t m
e o u l F ER o a i o o o b BE c o
_ 0 F t
n .
t n . n d t tn .
t r
& E ee e n . n .p ee i r ee ee ee R rc E(
R G
E rc an pe pu an pe pu d
e r
e t
e n
rc an pe pu rc an pe pu rce anl peb pul
_ al al e al al al e
_ A f f 7 t 0t f f f s F on on 1 n 5 o on on ono Ni Ni 1 E 1 N Ni Ni nip T
- E RO 55 44 77 8s 11 99 77 MAN I
55 01 22 01 31 90 44 12 22 34 22 01 44 23
- T TE 00 00 01 00 00 11 00 S
2 5 3 0 6 6 6 Y 09 19 1 9 19 09 09 09 E AR /7 29
/7 9
/7 /7 /? /7 /7 T DA 01 101 19 01 19 19 19 19 A /E 01 , 01 01 01 D OY M
) )
- ) ) ) ) ) )
- T 41 53 6 4 87 90 O S 76 0 3 26 26 26 26 26 27 37 N E 0 0 0 0 0 0 T ( ( ( 0 ( ( (
( (
ll
)I!IIll
, h s . _
r t rc a l _
g o l eie e e _
n of u thh , v i t ks f wtdyS k aa h e .
w tae. e nlI t l _
oodnerd nhabC e y t .
lt eep i noc aR psl r a, eCsr l t m al o ttb r abo r d nbaie t ort uo f aart sv a ieinom tl wero hoo,b cft af e e ri el e
o N
ti s
r dth
. e s eft - an iesmpr t don. cs l ca tt ysr e ,i a t e _
u litk nroe ur b t W ep sk sprna ,i da dbora f
awnaapsg data d uvdc beu aes te t el t E seuee , uraefue _
I V
i a oi e d y r t t al pnl ab a
o endneu ttl s s e csee.
i mp pou sea nirh an nppsm.
a pt uin ssbtrd nr eaege aamdegnr n povcooh mfna bt _
E aorea a d e i gl yt rai c rau ti reo p ei p oeirea g
R nev aa . t t ee t ot m t p r o pl r cr tc pe sb ael ey nfn s s e o mlilk vd d s d o gh l t ot pft al un R isOlahrhcp aea i eenat at e gsonr0 ga edi ehll eoi P e nlih v net a2 . htbeu r _
eef vell k n ter t avql u M ho.g e hr el ea oi taye it sh C itul osep sbeeed cd ntt s erl vv i altuell k tiCtR k s pb r a s o hrolsv _
uN F 0 ein i tho de tf enolb rdr retar ar cer ned O 3t pk eo att gelt c veb at onat o otuseoed t c pk ol e .
eI ormt e h ant o uremadbn wahi .by w d ou rr iu nc r S l l ou g a
tdriakn l ol u eoa ot c,onab eece wdea rr _
.swrt nekl n ft rft rt a gt o a p onatmsn ssltaeu T_ t t uk dapruay p tsltied tal oi io n sc aelosoa e l rns eentbe _
gr Ooctl t l ntt t o a e p yi c _
U cltnbe wlapa b aeinnusu dtvei g dtt ol rgul ro _
S naa orrr ial ge n pki o ci e rOr il k e emdoe hfl udtl esrotned hept r ve taebu E ov.ecub r kk ash iteedfd l osh eo o odtasa co ra l cth et R D t vtaoe D it ek r D c.waacim Dt E igcsnr et sa D t ctha .,l a egaabsc Dlnemoei syo buteTl E ceoub t E rsh hap h Eif iCn l Elil rrh p na N wnnnaas S N eh N oat otM N wol ik e nppw N
Rlt emeol If F entcTt c hohiu Rt w I a F rtt rt .a e
t e
r W Fh watech od hut rTl .i a
c R siii I nS F rn dcpo tp ea R I
F s
et gvn rneen0t y
.mnr a 1 N ak Mf:di oe sns O Ndi al N esueld ste f ,dp N O tg oi.bt aear e
N oil mo2 l F rwteiie N o as h I
O th ua O peft uit N ct at ti y O t n ui pb O t elihec T C epcc.t c C ot eusl rpeapu ehu.lllt pteua-C aed erz rbimdo C anertt ua C ahurcgn A ciett c T adfisue Tee mrvt T a O rf oeeeno msvsvn T e O ptaroro l o rt ricti ueb T e0nsdok o rtc T e l e2a 5ni L O r f welh Ohhoo nu NToctl ri n O pei n d b a r O poah3ht NTaastft NTtt caio NOavften NobsiaaMp U NotcT0cI C
L Y 3
7 A
C B D 4 E
t f
c e _
T EM8 f S D 9S e t SE IV O E 1. ) o n
h c
t e
NT R 5 2 t i 0 f s
O IB P&1 T t
y 30 0
9 f
e TI S YI r 4 0
- TTRB AC NNA H I
0t 1 n 6
0 ,
2 1
GR EEUX M RE 3e 1
/e l
a h
cd 5 E2 B( 0t m
r 5e 4t h 0
L-LI UmE C h h2a c3l o
n 5c a c h0 3
OAF c 5 3i 0
1 c/
l AM DUN 0 3
4 g.
3 o. n
- p 8d 0n 9
0 5n 2 e C T 3 g
- CDO 8 1ti
/ m 4
+ r
/i / 0o R 1 7 H FIN H /
0 0O 8R5 Tg U
0*
07 7
0 0
/r 0d 5 y E II 6 2C( M 1- 1 1H L
B EE CD A PD T SU T R 3 FL O P8 OC N E9 )I R
YI )6I 1. T RT G 2B d e
d e
d e
d e
d e
d e
AO N UY H I
d u
d u
d u
d u
d u
d MN HA X E l
c l
c l
c l
c l l u
MS CM(
( n n n n c n
c n
UN SO CO R
F i
t i
t i
t i
t i
t i
t o o o o o o I N N N N N N N T
I D dh ec .
D T ti t ahr A R d n ts wa O e h R P r ei n c E E)
RI t s i* i p
T e e l
- e. e e whr A NE d d un d e rct e s W OC SN l
e i
s ro ci i
s d
i d
i h" tye s s NE t u
t u
at c
t u
t t obh F ER o o m. o u u sd t
BE o o O o! se n lewo F t t r! t t
&E n .
ee n.
ee fo c
- n. n.
t
- n. nl n o
ER rc ee ee ee ul w rc e rc rc rc R( an pe an pe cT an an an efh oo G pu pu nD pe pe pe n s E eC pu pu pu os A al al ia al al al f f r f f l ss F on on af on on on f lal ea Ni Ni vo Ni Ni Ni gn s T
n u me ili E RD MAN 33 22 55 22 55 11 00 6s 99 euat I
11 11 44 rt e T TE 7e 23 67 89 90 23 acr seeya a S 11 11 00 00 01 00 ire Y 5 1 0 7 6 6 t ah E AR 09 3s 3e 2s 2s 2e i t T DA /7 /7 /7 /7 /7 /7 ts 19 29 29 29 nee A /E 01 11 11 11 29 11 29 ams DOY 11 uia QTc M
) 12
. T ) ) ) ) ) )
OS 1 2 2 6 3 8 4 9 0 e N E 37 37 37 7 5 8 6 3 8 1 s 0 0 0 3
0 3 e T ( ( 0 0 t
( ( ( ( ( o N
aeb tdehee ereit kpc pe m o.dow k r 1
% ew a0 0 cih thl t ehh ra t ant E ptb3 3 evt s ,T ul lhnl sttTp o eomes R - pe o8 7 l er kseb acal u u wtcn hn abr11 f t e nik o c0 o np, e . sa f wi R ss p ersr ,a ar i2df imd .kt oereo t t donodttmp p8 e ondar tttl ee P ea a u uw ra tifi y0es . ecaima alh M o l a el ppen t fip b l h rdt esbt l
t n e argrauth o t o ht at c enncua F _
at ao c ceneveuti eaaer s
ocsvu t aeat nr O vee rial i pih pOktta k o t sd tt e oe u mreoe S _ lt pi et e odd k p y t o ,n omrot r
e f a n
i nfi rioordi rw ittth asrni ont pt utVst msr oet ,
af ea T. ve - .wd ntte d e o ut ehea oo lt h nd n yi u l m oo oal puu i
pitt rha dat r ep pofypc pt l rh t
ctsk dnie na n d
. r orn S kt Tt .t ni t r tasoet bemsr adp r a ey a tbra eu piaa
. E na .ey dnaem hrd - t e R D a h s gl Dnpceor oehsbicrs ycl teebt c aoipe n a rl ne Col nout Ot obtdpr Ett n l E tt r eei ebiv Rniurpo tit - dao S M pmio0a 0 u Met vs vg i plhh attl s t asi w ab t N
Rsft atpmn o
uw Rstha ek a l e f
N I urt F eiiac 2 t I oe enec Fht e Do wona t e aeco I nt e neb .
O Nkfd a Ns t r ili yyv pbl e Eh Msde pit re rol rornl d udoi e Fre pev eee d
I T Oandd Cmoaeh gc O
Cen ndlt na gout y bl I ehoa o t ctallt e ths Nome. tldh Ot us Carun c oc A c o gi ci u wbk Fc ct ra nreu u rtai ekit Te Terdrsnoon Na t e y a e v s yl Tesc t nri L Ohowoh Oruioeuhra Or at py ol peeh a Opne3 eaew NTttl w NTddf raspt CTbiob spOl rwv U leOiB1dt ps C
L Y A B C E DC4 T EI8 DT S tS9 t SE VU1 OJ
- )
2 h
c NT R 5 3
2 - 0 O
IB PF1 O Yt i ,
4 6
m h
GR EEUx MMR d r
od ne d c
0d E2 UTB(
E e t ee t
3 e 6t L- CRE d u
- c a h c lba 0 a LI OAF l 0 ac /c 1
AM DP E N c
n
- 5. di 0n 2
8 ti sd nn 0i 0d
- T CDO i +i 0 1 n H C T / ui 1i R t U" /
IN F N o 0
0 0*
to"7 E II N (N? " 2 57 L CD 1* N" B EE A PD T SU T R .
F L O 3 g o
OC P 8 3 L E9 N R1 B O YI ) 6I T R
RT C d d
_ G 2B e e d d e
AO N UY H t
i n d u
d u d d e
MN HA X d l
c l
l u
l u
M CM(E e n c
n c c U S
( r i i n n F e N C i i SO I R
N 0t 0 n 2E t
N o
t N
o t o
t o
. T N N I
D D T A R t
a O e R P dl e
E E) dt RI dr T aa e e A NE h d d OC 0 c W
l i SN 0 2 p. t e s N E ie u t
u F ER erv o o O BE vti d d F ost t t e
& E h i n . n . e st n ee ee d d ER u u
_ ui rc rc
_ R( gbf an an l c
l c
- s G
E In,de pe pu pe pu i n
i n
0 al al A
_ O3 t f f t F on' on t
- R8 o o o C1 n Ni toi N
_ N
- T
_ E RO 66 8e 00 MAN I
33 7e 44 01 11 09 99 55 80 00 T TE 1n 22 00 23 23 S 00 00 Y 4 2s 3
26 2
2e 1 5 E AR /7 /7 /7 28 1 e
_ T DA 2, 29 29 /7 29
/7 A /E 11 11 11 29
_ 11 D OY 11
_ )
M
) ) ) ) )
T OS N E 7
38 4 8 38 5 9 38 7
0 8 4
8 1
4 6
9 T 0 0 0 0 0
_ ( ( ( ( ( (
_ t hl ec E
V I
yct t t neek iel rn a p r
.snt oai vo.e yp r yctbhg t t not n ier f .t u ck ol c
t s oa r
dkou oa ntr t
rrs eaot
_ E swm a c,ycd hl o ewmprr i
at a t e p .ttds en t e upt n R nsueet nsu ou l e c rp nepri ppn o r gh edl0 ed o apr R i ot at p hnb0t e t otl refd c- e , ea il ue bepe mae P t t sk u t aa3 ae t d o
_ M a naye b1 el tts o neoe nl iott pt ul uta5ak oh psb1saot ct u srt udielk ht o( pb u di g s v y- pu cnr me - umc eeti F
O t e ct g ut ndb l aieon lba am itm ep0rd 0
e n
t e aae cl gk el uinar tk ms aoe eecr otf0 c
d o1tf ee oolt r ht s t
dik vuo4 ea l aie k rp
_ S ffk t ri ert1l ffkl e n a ea . enedn bf dn rap hace saa eade T.
t l rol e t
oseoae ld l rodt r h a t
.h snt si d t at oen h U aowuog aowe t trat uod ctb -
tingtf ef e. ni al dntl l sn t n
_ S E
mt rael orh ash ct n a d a
k0d t mt raeal yp u
o i4 c h ,av ee hmeiiVe lt aaa lwmo eft e ga0 nf a orhlb e t tdn cusc re . ou R Dnet E p cic h Dtilne9 Et 0 aots Dnetp ua E cb k
a Daoini Dt Eit rus aok , .
rd roe D re S
N N eOoeTl R v I
Fh o.r th t
a c .
.it N ak Rf rdt eoy I oeusid F r wt oira sd NsOol o R w I
Fh o.r cpn t ar m
Eut va Nt R c3 Fl u,t ea I
e onb tl y
I Mwsc R sneteoeec anafh i B n pi pb t Edt cen Masape Rh nr I it
_ O Nsye ,dps No abree N sye ,e i
Nf y mb Fr et p i F r I
T O
Cedull t t at ti ye Ot el a Cahu e pt s
O Cedul attl e g
O aduo Nog Ot n . moiaew utd ds Nogeie Ot ns c pua - Cederr Caidr dt o Caiu5 e A
L U
Taemsvne Orhoenoh NTt crint Te r t cmca-O poarreo Not cFtbn l oa n Teemsi c
Orhoeh NTtcrT pus h
c n
a c
Taeis Orhano NTtfit l t p rkitk aeie Terl sa Opoaneneee NOwsipal pw vrr rk a1 Terc -
Opoe0 o NCwB1t C
L A f B
C D E
4 h
T EM8
_ SE S D IS9 V
O E 1. )
)
0 0
5 c
_ NT O
R 5 2 2
1 )
" h )
I B PMI h, 2 c ?
_ T ,
- S d TI S YI l c 5 e a
_ AC TMRB N AH I
m 0
2 l, 0 2
g g
_ GR EEUX MMR E r
od 4
1 a
m
/
0 j o
E2 tJTB(
ne r o
6 (
t .
L-LI CRE OAF *gca h c
n h
, h c
h c
i 0 3 l
AM T DUN CDO
- l. d 0 n 1
t 2
1 cl 0d e
g 2
3 0
1 5
2 H C R 4 i T
t
/
+
T 0
2 g 2
/
1
/
I N F N UM "r 0
U N
/o 0 0 I I
_EL 0j 0 0 I- 1 (
CD 6i 1 1
_ B EE A PD
_T SU T R
_ F L O 3 OC P 8 3 E9 YI N R1 g y
) 6g
_ RT G2g d d d d d d
_ e e e e e AO N UYg i
d d d d d d e
_ MN HAxE l u
c l
u c
l c
u l
u l
u l
u M CM( n c c c U S
( n n n n n i i i F
N CO i i i SO I R
N t
N o
t N
o t
o N
t o
t o
t o
_ T N N N I
_ D D
T
_ A R
_ O
_ R P
_ E E)
RI T N E
_ A OC W
SN NE
_ O F ER B E d d
_ .F e e d
e d d d
_ t E e e e
_ d d d d d d ER u u u u u u
_ l l l
_ R( c c c l
c l
c l
c
_ G n n n n n n
_ E i i i i i i A
t t t t t F t o o o o o o N N N N N N
_ T E RO
_ M AN I
T TE 55 55 23 22 00 1
011 00 55 67 77 11 67 55 00 90 88 22 67
_ S 11 00 11 12 00
_ E AR T OA Y 2 1/7 e
2 1a
/7 1
18
/7 0
0a
/7 6
08
/7 4
08
/7
_ A /E 29 11 29 11 29 11 29 11 29 11 129
_ D OY 1
_ M
_ )
. T ) ) ) ) ) )
O S N E T
2 4 9 9
O 3
40 0
1 4
40 2
1 5
41 8
0 6 4 1 9
0 7 4l 2
I
_ - .( ( ( ( ( ( (
r ms et.i es t ieest sc i0 g ed e oe t t e o6
. li r
W t ua Sn of c e ar t e am ti n
d t
imtuttea et p s h4 t1 e av ve ed hn h e
E p poa sh s 2 ,
st toeg g e e t a t rn.o iut t
V ms .s yo t
yr nddhy ign . tl qo en o n
E R oadeeti c i rk
,l e d a t ph a i et d
af eo denCa ah oRh etiNc i rntd vaa ad fe a
hih crc
$ u0 aeen.
t v senst vo ai ht r n
a y
R sntot 0d4 l ois P hnvnst r t t ot er s 0 9 y i e ie m tane e sl s iel tt ad 2 s1 tdh W i mnf p a fdhe hehon. e drtnt odt f nc aeco nao cnyro uwoo o ht g cii F ii p
inodt a ve n r t m, eih eseiS ttl tI of y O t eltdtd iebs ne t y t a l
.b yl iy olepoet l
t c srfiC don l
a w *g aneee ot oc1 r n 0 uedR a . e S _
srbi n ni4de as v n or0 nsdd sl my T .ml s pJes da ee i e7 i s ag na a lpy .
t o wlh e L odrs . ebv lyi yc gt n t e th1 ct ee n ocfd ya U thpugod l t0er o rn t rc.i i o eoe rw. eo. itp opal r td S
d t o.mna sl e t n no .
t ns1l a 0l u d t l e n rau c ylte ne E ei y es dt* f yl n T gd R SI ne gi om e Ddt ei e udin2 erp yi let ctah oh h d a D e o Dg pesa Ddam Dl i t EIitoeTe El cd nhsr . 0t Offi S Nak R 'f snt rees e nwe R ol e EEvoTaPt M
R ei s E gmttl6 No ooup y
E Ne mes Ebe Nagt
.5 o n m d g eopg yn Dt E cg N en N I ot e .f I oe fh FI so.hc .te t
Rl l cno,b R cf e I ao Rbnsd I
oaeee R
g es eti IRli O
I Frw No ,fo tese F
Ntd t
g N i m y0 t f n No raood
,snn F r Ntl o r F rht gl Npc e r cieu r f r F eu Ot et gh e O O a1 ng Oi ewiiett N arrd Ndd T O af O A
Cahl sncb rt unin Ctncein Ct ed6 en nc 1di Ct v esiin g Crc oi, glh Ct .nl c C t p O r a s O Ces L
T e Opoeiu sgriy a
Taa Ol r r
u Taae Oirht vu ir Tdi widda Tnpy nnia0 Tair 3 T eaem rye Tam ce NPTd Od ohddh Oiya Ol uo3 O nhhi U 1 NOt rsd7 m NFTt aed NAh Taac NMtd Np?dt1 N 7 st t Ori NTt C
L Y A B C E DC4 E
T O U8 S tS9 v
SE OWt.
NT R S) )
O PF OI2 i
3 Y -
TIB I S Yt )
l 1
d e e g
GR EEUX a m
1 2
g g
M g o
E2 MW U B(
R E h
r o
0
/
0 j
o
(
j
(
L - CRE c hn c h1 h3 t
hhh L OAF 0 0 : c c3 h hi cc c h 1
AMI DP E N 1 1 2*
15 5j 1
1 g
e 5 0 c 055 o c
- T CDO 2 1o 2 0 350 t H C 2 + 223 I N R 0
/ /t 16W
/
0L
/ 2 0 3 111
//
0 E
L F
I CD I
N 5 1
0u 0N 2(
0; 0
1t 0O 3 R 3C
/
0 6
3V 1
/
0 5(
00 00 11 1/
0 0
1 B EE P D A
T SU T R
F L O 3 OC P 8 N E9)
R 1 I Y I
) 6' (T RT G 2B d e
d e d e d d d d AO N UYH I
d u
d u da d e
d e e e d e
MN HAX l c
l !
t u u d
u d
u d u
E MS c l CM(
( n n c
n c n
l c
l c
l c l c
U N F i i i
i n n n n SO CO R t t t i i i i
o o o t t t t I
N N N N o o o o t T N N N N o I N .
D dh D ec T ti t A R ahr t w O s R P E ei n
E R )I s T irct w "h A te E 0 C e*
W 3 N H E h
t ye F ER obh O B E d d sd
& E F e e d e
d e
d e
d d d se ew d d d d e e e l o u u d E R l u u u d
u d
u d nl R (
l c c l l l u ul G n c n c c l
c l
c l c efh o
E i in n n n n n i i i n oe A t t i i i
F o o t t t l s o o t t o t lal e M N N N oo o o r N N N gn T il n um E RD 22 euat 44 MAN 33 67 55 00 22 22 33 77 00 99 11 88 rte 7t TE 11 90 12 56 12 01 33 22 23 101 1
ac S a1 00 23 e a 22 31 11 11 eey Y 2 9 9 t re 08 9 6 t ah E AR /7 08
/7 28
/7 28 28 1
28 8
18 8 i T DA 29 !9 /7 /7 /7 /7 18 t s A /E 09 09 09 /7 nee D OY 11 !1 11 11 11 09 11 09 11 09 ems 11 uia M QTc
)
T S
) ) ) ) ) )
12.
8 0 ) )
O S 8I 4
9 4l 09 13 2 3 3 9 4 3 4 s
e N E 52 5 531 5 T
I
(
l 1 531 54 e
( (
8
( ( (
1 541 t o
( ( ( N
,l
h e .
- E t ne r.dn I
- V o i e E y ea v
ar t g
- R n ed ra a we e
- R m oe ct
- P hr . no
- M t e sas t un
- F l n e eo O e odt md S
T.
e y.
tpa ie t ms iri doh i
tt r t c
t c
t
- t a c yd df t rh e e e ul t e areg ac. y r r r S h r r r E
R Dt Eeg nh ot rpn e i t or aeu Cdd Dae E
na t
DnEi c
o o_t s
o c
a Dn Ei o
c S R en Rti wld e fe f e ss N Re t
n e s
N Rs N It r D n I ut I m I m I e
- O I
T A
F eu Rdd O
Ces Tem ce reo e vi u at s ha t
r r Fl o ft 0T C
t n en Nt t
Fi O
Ct r
Pi O
Mt Ct r
Fi Nt O
Ct r
N ye Ts a Ta ta Ta L Ori NTt Oap Oel Oh oh Oh U C mo NTP NC NC NC C
L A Y B .
C
_ T DC4 EI8 DT E d l
i e I 5
- S tS9 f -
_ SE OJv U 1. ) t .
_ NT R S n o l O PFI 2 a m
IB O Yt i e r
e o
n GR EEU N MMR t
E2 UTB(
E e c
2
+ d e
d e
d e
d e
L-LI CRE OAF he n %l ce d u
d u
d u
d u
el 0 v
_ 1 AM DP E N 7 I 3!
4 e l
c l
c n
l c
l c
CDO 9l n n n CT R 9 e 0v 1
/t i i i i
_H FIN N /r 0 u 0u 5 n t
o t
o t
o t
o I I E 6s N N CD 1i N N
_L B EE PD 0 0 g g
.A o T SU T R 3 L o
L o
L L o
_ FL O P 8 O o 0 O OC N E9) R C
R C
R C
R C
YI R1 I a n n n
)6I T RT l i i i C 2 B d e
d e
AO N UY H t
d d d
e d
e d
e d
e MN HA X E l
u l
u r e
r e
r e
r e
M CM( c c t t t t U S ( n n n n n n N F i i e e e e C O SO I R
N t
N o
N t
o 0t 5 o 0t 2 o 0t 0 o 0t 5 o T 1 N 1 N 6N 4N I
D D T A R O
R P E E)
R1 T
A NE OC W S N NE F ER O BE d d d d d d
& FE d e
d e e e e e d d d d ER u u u u u u l l R( c c l
c l
c l
c l
c G n n n n n n E i i i i i i A
F t t t t t t o o o o o o N N N M N N 1
T 3 E RD 99 0e 0
MAN I 55 22 44 00 22 11 55/
000 44 33 T TE 85 90 58 56 3o1 79 S 00 52 11 00 201 11 Y 0 5 4 4 2 E AR 1a
/7 18
/7 18 0s 04 08 T DA 89
/7 09
/7 0s
/7 09 37
/9 A /E 109 1 11 1i 11 91 D OY 11 M
)
'.T OS
)
5
)
0
)
2
)
6
) )
sT i
6 7 S 9 1 5
N E 51 4 5 1 S1 5
51 5 eS 4gl T (
I
( ( ( ( ( (
Appendix I .
SUMMARY
OF SPECIFIC ALLEGATIONS OF HYDROGEN ADDITIONS TO THE MAKEUP TANK DURING RCIB TESTS AND THE RESULTS OF MPR REVIEW OF THOSE ALLEGATIONS This appendix summarizes in tabular form (Table I-1) specific allegations that hydrogen was added to the makeup tank in the course of an RCIB test and af fected the leakage calculation.
This included those potential instances identified in the Faegre and Benson Report (Reference 1) and those identified by the NRC in Exhibits 1 and 2.
The MPR conclusions are based on examination of the makeup tank level strip chart and other appropriate plant records. Those cases where the hydrogen addition can be confirmed are identified and are discussed in further detail in Section VII of the main report. For the allegations which MPR could not confirm, the factors involved are briefly summarized.
I.1
V d I. et y rpd a rlft
_ E R
t nl tf at 1
9ia dd f
n a l ,
s t e ,e t o
t t rn imn sI n ut nu t=dae
. R P
n s h uedao nrt e ns!!it s at h l t
uno
.g e
lai o.
t m pl ao mu ost a b
- M i sf mi t e sir ti e cec
- e o gf e d odl ref F a noo O
t .
ct ec f
otwl ngy ued J. lda r r3 y 1h ndb ao d
eae t o S f e ly t enrn isdl ya y
t a le np r h cs n ds T.
f f ef o n e pI ee so vg . n i hc lbu aoo ngt e t aoo i
fl e o1 ,ti ioon l
u t e. iodp ey bbt oa re w t
- S un t
u t
u f sh sr4 o
e tid 7 t
u Dt mt Dr t Dtdm sd ea n u
_ E bi b bt e
y- b E cr E pf o E cyu dalbtk.
a
_ R ,2 ,. .*
nhf eh2 at op 2 ..
4 eis plf r M
p e on M eh r t
aca t eue D Dt Df D ad- Dt ]l na l snd lpl ds n sl p F feen E e E e Ef E eneh e9 E e F eoe Fi on e uf -
Ml N e M e MlI mss 1 me Ndcp N ria Ndmi " om o
- Rb IRt P e e .i e tpf O Cera p O t O r u tst r oa -
l rn t el I i t I1 I f Cecl Cein F s F e F1 F c* - F o ce cue t nhk
- N s Oo N
O N
- O
- N eieanI Oh .h rom N Tahd Tadv Ta n cf a s
nis61 ep a
CP Cs C* Ct4 t t cT O
Cm Ort n NToa Oree NTrl Or oe NTcb I S Y
- T
_ B S 0C4 E
E SB T
EI8 0
tS vU1 T
OJ 5) 9
_ NI O C R
PF4 2
I R O Yi S
t T TR 8 A2 T NNA Fi t s t
G- EEUX r E I L M M MR E UT B(
CRE 5 c h
a i 5 L T
_ 1 4 0r OAF 1 1 n 4 t ) 8
_ I AG DP N 0 2 o 6 c t l 1
- N CDO E
5 98 3
0 5d 0
0 e f
c e
l a l/a I CI
- E FI R R 0 2
10 1
0 5
1 e 0s 3
1
.f 5f f
4 g
g N 5 e 0 0O 0 2 r 2 0 e 00 I U / /n / /e / 1 6 o 3* 90 0
-- BLA ECD
- e *
- m
- 0 n 02 01 0 2 3t 2 4I 3 /I /1 /1 1 PS
_ T SN O
T R
FI O OT I P
E
)
s R'l Y D e
- *i s h s R D
)
G l
e c d e d d d d d B
A A
_ n e N
h h e e o e e t
H c i d c d d e, d d
- U a u n u u u u u MN H X l 5 l i l l l l l C"(E
_ ME (
2 3 c
n 4 c
n c
n c
n c
n c
n
_ UG C 4 0 i
5 i i i i i SO R R 1
1 4
0 t
o 2
0 t
o t
o t
o t
o t
o N
D Y
0 0 H 2 N N N N M
_ H T F R O
O P
_ E)
RI N E 5 0 0
- OC 0 3 0 0 0 5 SN 2 1 1 1 3 4
_ N E 0 0 5 2 1 8
_ ER t t 0
t 2 1
_ B E a a t a t d d d t
& E F a e e e e a d d d d= d d
- ER e g
e g d e d t u u d
( e g e l l l e R g g g g c
_ C o o g o g
g c
n n c
n g
g E l l o l o i i i o l l A O O O t t t l
F P R O R O o o o O C C A C A N N M A
- EJ T
pD le4 11 11 66 66 55 99 55
' eN l21 44 11 22 44 4 4 00 Yt TE 01 45 01 01 1671 23 $9 S O0 00 00 22 22 00 00 00
- E aR r
19 2 10.
7 9
5 19 4
9 3
3s 6
28 0
28 5
/7 /7 1/7 /7 1/7 /7 /7 /7
- T cA A /E 39 0i 39 01 29 01 29 01 29 01 29 11 29 11 129 1 2
1 0 oY
_ u
)
T
) ) ) )
l
) ) ) )
6 4 5 8 9 8 9 O S 10 22 33 43 S 3 67 78 1
e8 N ET #0( R0( H0( M0 ( M(0 N0( 40 H0( n
( (
rl
t V o tdat b E n ra s of R uef n ot oi rn p
o R t cs en e P ynk e ss t ai M c rit e pe wr F e vt ob p. .
O f
f o enny ont e wa ris e
S l yom d T . 0t n eld l1 L cb e e e U t i
aea r eeh S u 2 t rb e t t sht E b t mTooh re t b u
b u !t g
R ,
ot g
.de , . pt a g .. , .kf noi n
t ce . Ot s n Dt Dt a r t cas Ec E ctt u n el t efl n uewa itne Df
- m. P e Nf de Rf pu ld tre u
f ae ei d f t e leenrib el lf f e le usl reoee N aedo M u lrv ooEv Oo h nr O ooa ae C nT e CcFT ip Cn CnMal S Y T B S DC4 E
E EI8 T DT iS9 V U SB NI O J 8. )
5 R 2 OR I
C PFt O Yti T S TRs A2 T NNA H i
G- EEUx E M I MMR E
)
y L T UTB( t CRE c ) 4t i
L OAF e 7 t 1 t l
AG DP E N f f
1 c 7 e 2 e 0 e
- CIN CDO 5 1
e 1f
/f l/f I R 4 0 e f
l e ee E FIR I U N 0
/
3 m 2 o e
po p
o L
B E CD
- 2 2 s
/( 2I n 0 n 2(
A PS T SN T R
FI O O 3 OT P 8 I E 9)
Y D R1 t
) 6I T R D C B d d o d e
A A e e e N t d d d d U H u u u MN H x E
l c
l lu l e
ME C
( ( n c
n c
n e UG F i i i f
SO R C
R N
t o
t o
t o
t o
n n N N D
_ Y H T F R O
O P
- E)
Rt NE 5 0 7 OC 1 3 1 d e
S N 4 2 7 g NE 0 2 1 g
ER t t o BE a a t
a l
& FE d d d O e e e A ER g g g R( g g g d0 G o o o a3 s2 E l l l 0
A O 0 O O F R R R Rt C C C Ce T
E RD 99 55 77 33 MAN I
00 45 123 1 22 67 11 23 T TE 00 22 00 S 11 Y 9 5 5 0 09 08 28 28 E AR /7 /7 /7 /7 T DA 19 I9 09 09 A /E 11 I1 11 11 D OY M
) )
. T ) ) )
O S 9 22 34 40 N E ll l 12 83 1 4
T Rl
(
31 31 3 1
( ( ( (
,ll ,j
L l l l l I
! Appendix J i r
CORR 2CTION OF RECORDED TOTAL i
! LEAKAGE VALUES IN INDIVIDUAL RCIB TESTS I
- 1. Introduction and Purpose i Since the RCIB test value of totel leakage includes the effects of the procedure errorr and instrument errors discussed in Section III, it is necessary to correct the reported total leakage value from the RCIB test to 1
compare it to the results of the " slope" and "weter l addition" methods presented in Section IV. This !
appendix describes the method used to make this correction.
- 2. Correction for Water Additions i The water additions made during an RCIB test and :
included in the calculation were not evaluated et the same temperature as that at which the total leakage was calculated. Accordingly, the water additions were undervalued and the total leakage calculated in the RCIB test was lower than the correct value. The ratio of the densities at room temperature to RCS average temperature-(about 581*F) was assumed to be 1.396.
Since the test period was alweys one hour, the correction for a logged addition of A gallons would be:
J
\'
I Correction *
+A (1.396-1) = 0.0066A gpm ]
60 In this expression A is in gallons at room temperature, but the corrrection is in gpm evaluated at RCS average temperature (about 581*F).
l i
- 3. Correction for Makeup Tank Level Inekrument Error l Where there is a difference between the makeup tank >
level at the start and the end of the RCIB test and the makeup tank level instrument is in error, the makeup tank level instrument appears to overindicate the actual change in level. The amount of this error is determined from the observed ratio between the amount of the addition as indicated by the increase in th'e level of the makeup tank and the amount of the addition which was recorded in the CRO Log. Except in a few isolated time periods and during February and March l 1979, this ratio was approximately 1.0 and the I correction for makeup tank level instrument error was l small.
When the ratio was greater than 1.0, a makeup tank level reduction during an RCIB test (the usual case) I J.2 l
\-
would result in a correction which would reduce the total leakage calculated in the RCIB test. In a few cases the makeup tank level increased in the course of the test and the correction increased the RCIB test '
calculated leakage.
If the indicated change in makeup tank level in the course of the test is dL, i.e.,
dL= (Final level) - (Initial level) inchea a
and R is the average observed ratio between the indicatdd and the observed amount of water )
I additions,* the amount of the correction expressed as j in the RCIB test result at RCS average temperature is as follows:
l l
Correction = + 1.396x30.8 x (1 - 1/R) x dL gpm 60 For this expression the test period is one hour and the volume in the makeup tank is 30.8 gallons per inch of ,
level.
l For those days on which no average ratio could be determined the closest available data was used or if midway between data, an average value was used. See Section IV.
J.3 l
'