ML19309A426
| ML19309A426 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 03/14/1980 |
| From: | Poindexter T NRC - TMI-2 OPERATIONS/SUPPORT TASK FORCE |
| To: | Jay Collins NRC - TMI-2 OPERATIONS/SUPPORT TASK FORCE |
| References | |
| NUDOCS 8003280369 | |
| Download: ML19309A426 (3) | |
Text
DISTRIBUTION:
Central Files
~
tt.AR 141380141380~ NRR RF ASB RF T. Poindexter RHVollmer flRC POR 26~al PDR c
T!il RF SEP RF
!!E!:0RANDUT: FOR:
J. T. Collins, Deputy Cirector i!RC/Till Technical Support Staff FRO:1:
T. C. Poindexter 0hh CDOf gP Wl]g@[}kg TitI Technical Support Staff l
SUBJECT:
REACTOR BUILDING SUliP LEVEL REFER'El:CE:
l'et-Ed Letter TLL-051, coactor Building !!ater Level:
llilson to Collins, 2/20/80 (copy enclosed)
As the sump level rises, it becomes more critical to be able to monitor and predict future icycl increases in the rcactor building.
The referenced l'et-Ed letter proposes some values and techniques which by the technical staff's oun calculations have bean shown incorrect.
The rast critical inconsistency is the anticipated rate of rise of the sump l evel. l'et-Ed predicted that the sump would reach the red-line" value of 290.5 feet by Gctober,1900. However, the technical support staff's pre-diction.uas l' arch 7,1980
.ictual level c.aasure:. cats stated that the level uas reached briefly on l' arch 5,1930 Through discussions with Ron Greenwood of Gilbert Associates on site, it has been learned that the same method the liRC has been using has since been adopted by f*.et-Ed and their advisors, which has greatly improved their prediction carability.
However, our own equivalent "in-house" capabilities will enable us to anticipate potential troublesome conditions independently.
A.
_ Discussion of i:ethods The following is a brief description of our method for reaching the linear regression predictions.
1.
All data has been assigned a sequence number beginriing with the first day that data was recorded (June 1,1979) example:
5-1-79 = day 1 ; 7-1-79 = day 31; 8-1-79 = day 61 etc. Also the hour of the measurement is incorporated by using decinal fractions of the days.
Exampl e:
1200 hours0.0139 days <br />0.333 hours <br />0.00198 weeks <br />4.566e-4 months <br /> on day 5 = day 5.5; 0400 hours0.00463 days <br />0.111 hours <br />6.613757e-4 weeks <br />1.522e-4 months <br /> on day 10 = day 10.16 etc.
2 This data is input into a linear regression computer progr,am (TI-53 calculator w/PC-100c printer) with the days forming the omu>.....
s-4
.s.o.o m. 3 (, 9
- >i.
MAR 141980 J. T. Collins -
2 x axis the the sump level values forming the y axis. At the end of the couputing sequence, values for slope, correlation factor, and future predictions may be made. The definitions for the chove are straight forward with the prsible exception of the correlation factor.
forrelation Factor (CF)
It 31Vays lies Sctueen -1 (for negative slope) and +1 (for positive slope).
The C.F. describes numerically how close to the straight line linear regression plot the data is. Also by squaring the C.F. you get how many (*.)
of your present and predicted values will lie on this same regression straight line.
The correlation factor is our absolute method for judging the accuracy and applicability of our predictions.
In addition to the correlation factor, a standard deviation of +.1 is also used as a guideline.
3 A conthly prediction is done, giving us " rough results." Then a 1:eekly prediction is done for the present week, using the two previous weeks' data points. This gives us core of a fine tuned prediction. This fine tune prediction occasionally yields results contrary to the conthly prediction because it accounts for unforcseen dips and rises' in the level.
4 Anticipated Sucp Levels The follouing is a listing of the " rough" and fine tuned anticipated sump levels. The assumption is that no significant, induced changes have occurred in the leakage rate. The average is based on' 5 to 6 readings / day.
Dy f*ean Predicted Level (ft)
Actual Average (ft) 2/28 290.32 290.31 2/29 290.32 290.29 3/1 290.33 290.29 3/2 290.34 290.39 3/3 290.35 290.38 3/4 290.36 290.32 3/5 290.36 290.41 3/6 290.37 290.36 i'
omcc).
.. g'.
s u m.w c h.........
out >!.
D"P]D
" F T Y ;3
~
can slf S $M J. T. Collins laAR 14 380
, Approximate Anticipated Date
- Elevation 291 '0" Day 351 = lty 27,1930 291 '11" (Pcn. 401)
Day 4C0 = September 23, 1930 292'2" Day 514 = October 27, 1980
- See Itens of Concern Below B.
Items of Concern The following is a list of critical sump clevations that should be addressed by !!ct-Ed.
Eleva tion Equip. ment Effect a.
291 '0 "
Pressurizer Heater Power Lose Pressurizer Heaters (291.0 )
b.
2 91 '0 "
Core Flooding Tank P,inor Concern, but for extra (291.0)
Pressure protection, lowering the tank pressure to insure that 14DHR will not be overpressurized should be reviewed.
c.
292'2" CH-V171 Decay Heat isolation valve (292.16) inoperability 292'5" CH-V1 (292.41)
Gefore elevation 292'2" 11 reached a policy decision c.ust be r.ade involving the use of the DHR and the f:DHR systems.
If no action is taken and the level is allowed to rise above DH-V171, V1, these valves may become inoperable.
If the decision is made to open D3V171, V1 then portions of the decay beat system will be isolated from the RCS pressure boundary by one valve (D3V-3).
C.
The installation of the suap level canonater assembly during the week of
!* arch 9,1930 should provide rore consistent tsasurenents.
Therefore, our confidence level in the predictions should increase. The technical support staff will continue to provide you with updates on predictions and bring to your attention possible areas of concern.
/s i
t,.)Poinoexter omcc)... O.S B s [.
. TliI.. Technical.. Sum;.r.t.S ta.f f,,,,,
l svenwckTPo.tn exter:b il..
om>. 3/N/80....
_,,, _