ML19317G895
| ML19317G895 | |
| Person / Time | |
|---|---|
| Site: | Rancho Seco |
| Issue date: | 12/30/1977 |
| From: | Walbridge W SACRAMENTO MUNICIPAL UTILITY DISTRICT |
| To: | Engelken R |
| References | |
| NUDOCS 8004020553 | |
| Download: ML19317G895 (5) | |
Text
-. _ _ _ _
DISTRJ.3UTION AFTER ISSUANCE C PERATING LICENSE P N;c P62419s.
occlxhnumesh/JL u.s.nuct. man Rerus.Arony commission Hi-
)
NRC DISTRIBUTION PoR PART 50 DOCKET MATERIAL FROM:
cats or occuusar f' TQ:
Sacramento Municipal Utility Dist.
12/30/77 Mr. R. H. Engelken Sacramento, California o,7,,,,,, y,,
Wm. C. Walbridge 1/10/78
,Mn ONoToRiZED PROP INPUT PoRM NUuSER op CDPtsS RsCEIVED Ic i: enAs.
(unct. Ass:Piso I
slowen a
i n
" :::sce:PTioN gNc;,c3y mg Consists of follow-up report concerning j
an error in the software for the plant computer.....
- I II (2-P)
(2-P) i
! PLLNT NAME: Rancho Seco RJL 1/11/78 I
c3 Ense.t.
R.e e.'d I SAFETY
_ FOR ACTION /INFORMATION I IBRANCH CHIEF: (7) i k' & tC1 I
I I II l
1 i
l
, 't I
iAT INTERNAL DISTRIBUTION l
f ' REG FTT e J l
l i
11 nrwr' tmo I
If C 7 /Si l
i i
i l l
f err n in WATTru l l
'r a. i F i
i i
i i
W i - w o-i )
i c I l
i I strar3
'R A rt
'uTm en
! WW M R l
l l
U. COT T.TNS i
l i I, I Lt. MCCOUCH I
i i I
I l
l i i I
t I I
i
(
I I
I I j
I l
i I I
i l l
I l
l I I
l l
l 6 I
I I i 1 t
EXTERNAL OfSTRIBUTICN 4
CONTROL NUMBER I
- L2DR:.% Af r a n4 eaMeg, C g.
I i
~~
g-i IIC l '
I I
I m
E Q NSIC i
Ea516 CT5 SENT CAIT.GRT D
{
c6ujogof53 hy
. R e. L u
! I I
r
,m g
t SMUD D E C DPY SACRAMENTO MUNICIPAL UTILITY DISTRICT O 6201 S street, Box 15830, sacramento, california 95813; (916) 452-3211 December 30, 1977
.~.
Director of Regulatory Operations b
ATTN: Mr. R. H. Engelken NRC Operations Office, Region V Nuclear Regulatory Commission 1990 North California Boulevard Walnut Creek, California 94595 r-Re:
Rancho Seco Muclear Generating Station f , 7 Operating License DPR-54, Docket No. 50-312
..<6a
~
o.
10 CFR 21 Report dated December 27, 1977 f,,
w 5-
.a l
Dear Mr. Engelken:
s,. g' q
%%t In conformance to the requirements of 10 CFR 21, we are sub mitting this report as the required follow-up to a tGephone notification made to your office by !ir. R. J. Rodriguez, Manager of Nuclear Operations on December 27, 1977.
This notification was made after an evaluation was rendered by the Plant Review Committee (PRC) and their findings were-concurred uith by Mr. Rodriguez. Mr. William C. Walbridge, General Manager was informed that the 10 CFR 21 regulations were applicable to the situation.
The discovery of the " defect" as defined by 10 CFR 21.3(d) consists of an error in c5e software for the plant computer.
The com-ruter, 3 Bailey 855 model supplied by Babcock
,d Wilcox Company, was calcul: ting values for the Departure from Nucleate Boiling Ratio (DNBR, extra;>olated to worst case conditions) that were higher than plant conditions warranted.
This non-conservative error resulted from the method by which "best guess" reactor coolant flow data was extracted from a " spurious input" subroutine whenever actual data is out of range.
The output data from whis subroutine was stored using a multiplication factor different from th t assumed in the DNBR calculation.
This led to the flow data being increasea oy a factor of 10, after which the first significant figure was truncated.
For the particular case of Rancho Seco, these incorrect flow rates were approximately 136 percent of actual flow.
With higher reactor coolant flow rates than actual as data input, the DNBR calculation produced higher values to reflect the improved heat transfer.
A recalculation of data shoucd that at no time did the DNBR drop below the limiting vaue of 1.3.
For this reason, the filing of a reporc-i able occurrence report was not necessary.
However, this may not be the case for other units which have different reactor coolant flow rates.
l The results may be extremely conservative ar they may be non-conservative AN ELECTRIC 3YSTEM S E RVIN G MORE THAN 600,000 IN THE HEART OF C A LIFO R N I A
g7 Q
t' -
Mr. R. H. Engelken December 30, 1977 to the point where the DNBR limit is violated.
For this reason, Babcock and Wilcox was informed of the problem so they could pursue having this problem corrected at other plants with similar software if a similar situation exists at other sites.
The computer programming has been changed to correct the e
scaling problem, allowing valid DNBRs to be calculated.
Ver tr'uly your,
k L.
y, Wm. C. Walbridge General Manager i
y
\\
j SACRAMENTO MUNICIPAL UTILITY DISTRICT OFFICC MCMOf tANDUM L
DATE:
December 28, 1977 CUOJtCY:
10 CFR 21 Response on High DNBR/ Flow' During the startup testing for Cycle 2 at Rancho Seco Nuc1 car Generating Station, it vas noticed that the primary heat balance calculated by the Plant process computer was showing appror.inately 120% full power when the reactor was operated at 100% full power.
During Cycle 1 with similar conditions, it was usual to observe the primary systco heat balance to indicate approximately 90% full power when the secondary heat balance was controlling the plant at 100% full power.
This dif ference was not considered to be an operational problem as the primary systen heat balance is'not used 'above 15% full power.
In attempting to determine what had caused the apparant increase in the primary heat balance, it was found that the reactor coolant system flow used in the heat balance calculation was abnorually high, approxinately 136% above actual.
Normally, flow is obtained independently f rom each of the four reactor protection system channels, then ave.rnged and placed in the computer nemory for further use and reference.
In this ense the stored value was unusually high, while the indicated flows from cach of the RPS channels were essentially the same as Cycle 1 values.
In researching this difference, it was found that o flag had been set in the scf t-ware, telling the computer that one of the input flous was bad.
This directed the software to calculate a substitute average flow and store it for further use.
In the normal flow averaging routine, the values obtained carry a zero preceding the two significant figures before the decimal point.
Since the averaging routine always expects this leading zero, it automatically truncates and stores only the significant digits.
In the case at hand, the substitution routine simply calculated the two significant digits and passed them back to the averaging routine which was expecting the 3-digit input which it immediately 6
truncated. As an example, typical loop flow is 69.5x10 pounds per hour.
The 6
effect of truncating then provided the value of 95.0x10 pounds per hour to be stored.
Since the primary system heat balance relics on stored flow, the high power icvol resulted. What was not known was that it is the stored flow which is used in the determination of DMBR in the subsequent thermal-hydraulic calcula-tions. Hence, when the flag was reset to return the flow to normal, a dramatic drop in calculated DNBR was chuerved.
This drop atounted to approxinately 1.5 DNBR units when operating at. full power with nominal power distributions.
Of major concern was the effect this had on the results of startup testing done at the beginning of cycle in which abnormal power distributions were imposed to sinulate design rather t,han norcal configurations. We were abic to correct the recults of the startup test program by correlating them with the before and after flew fix DMBC's.
The b'estinghouse U-3 DSSR correlation used in the software incorporates flow as an additive and lincsl parameter; thus the corrections were straight fo rward.
In correcting all of our test results, we foundfthat, where extrapolated to the design overpower at 112% full power, the
.dinicun DSDW dropped. from 2.99 to 1.84.
No safety viointion or consideration-
e o
- p f ',
y,.
Pega 2 Dec:mber 28, 1977 W
cxists at Rancho Seco rince the minimum acceptable design'DN3R is 1.30 and this value was never equaled or exceeded.
~
1hc implications of this probica are apparant to other plants using the same or cimilar sof tware.
If the design margins for a particular reactor or cycle were not as great as those at Rancho Seco, they < tuld be deluded into believing that their minimum D"BR',s were greater than accep6ance criteria, when in fact, this would not be the car.c uhen the correction to actual flou was made.
The Rancho Seco computer has been reprogramacd so that it will not truncate when using the cubstitution routine.
At this point in time, we have not been abic to establish khat originally set the flag in the software which caused the substitution routine to be.used.
The purpose of this report is to make other users of this cof tpare aware of this idiosyncrasy and to suggest that they revisc their software prior to observing the probicm.
e e
e e
e 9
e S
e e
e S
o 6
g t
e t
I e
g
,I D
D e
O a
4