ML20058B710
| ML20058B710 | |
| Person / Time | |
|---|---|
| Site: | Peach Bottom  |
| Issue date: | 07/14/1982 |
| From: | Cooney M PECO ENERGY CO., (FORMERLY PHILADELPHIA ELECTRIC |
| To: | Haynes R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| References | |
| NUDOCS 8207260119 | |
| Download: ML20058B710 (6) | |
Text
.
PHILADELPHIA ELECTRIC COMPANY 2301 M ARKET STREET P.O. BOX 8699 PHILADELPHI A. PA.19101 (2156041 4000 July 14, 1982 Docket Non. 50-277 50-278 Mr.
R.
C.
Haynen, Administrator Reqion I U.S.
Nuclear Regulatory Commincion 631 Park Avenue Kina of Prursia, PA 19406
Dear Mr. Hayner:
On June 19, 1982, durinq a scheduled loss of power tent on Peach Bottom Aton ic Power Station Unit 2, an inadvertent ctartun of Emergency Core Coolina Systemn (ECCS) war initiated on Unit 3 while at full power.
Cold water injection by HPCI and RCIC cauned Unit 3 power level to spike to approximately 114 percent.
The trannient was terminated by the onorator who manually tripped HPCI and RCIC.
This renort provides a complete descrintion of the event, its cauce, and corrective actions.
ECCS initiation on Unit 3 war caused by a defective ELMA DC power nunply which feedc a 24 volt DC logic circuit.
The logic circuit is normallv powered from the Unit 3 125 volt DC cyntem throuch a " TOPAZ" inver ter and the ELMA power sunpl y.
Backup or al ternate nower to thic loqic circuit ir cuoplied from a common (Un i ts 2 and 3) 120 volt AC dintribution panel throuqh a separate ELMA power cupply.
(See Attachment I).
The connon dintribution panel in fed from a Unit 24 KV emergency bur.
Al thouah the defective ELMA powor supply was not canable of supnlying power, it could run at full voltace under no load conditionn.
As a result, the power supply voltage monitor did not annunciate the dearaded condition and the operator was unaware that the ECC3 logic circuit was powered from the backun 120 volt AC distribution panel.
When this distribution panel was interrupted as recuired for the Unit 2 lors of rower test, the Unit 3 ECCS loqic lost its alternate cource.
Outnut voltage from the defective DC nupplied ELMA power runnly decreased causina the 8207260119 820714 CIC'sts PDR ADOCK 05000277 S
Mr'.
R. C. Haynes Page 2 1
B&D instrument signals to decrease and indicate low water level: however, there was still sufficient voltage to operate the trip units.
In addition to HPCI and RCIC injection, the low pressure coolant injection (LPCI) and core spray systems initiated automatically.
The LPCI pumps started but isolation valves did not open due to high pressure interleck.
The core spray pumps did not start because of load sequencing time delays.
Our core analysis to determine the effect of the HPCI/RCIC coJd water injection indicates that core thermal limits were not exceeded.
Core conditions and the results of our analysis are summarized in Attachment II.
As a result of this incident, the followina actions have 4.
been taken to prevent future occurrences of this nature.
- i 1
(1) Power sunplies will be monitored on a monthly basis pending final design changes.
Degraded operation of the ELMA power supplies is characterized by a hiah ripple voltage imposed on the normal 24 volt DC output.
Ripple voltage will be checked monthly to verify operation within acceptable limits.
l (2) An investigation has been initiated to determine i f continuous monitoring of the ripple voltage is appropriate to alert the operator to a degraded condition.
Voltage monitors already exist on the output of each power supply to alarm for loss of voltane.
4 (3) A study has been completed which addressed plant design and possible looic cross-connections.
Our review has determined that there are no design errors causing logic crons-connections and that this problem was a result of the unicue configuration for the loss of power test.
i l
(4) The ELMA ECCS power supplies will be replaced as soon as possible.
An alternate supplier has been identified, however, the power supply does not yet meet the recuired seismic cualifications.
(5) Initial review indicates that additional loading on the i
ELMA power supplies may extend their life.
The l
possibility of better loading the system will be investigated to determine feasibility and merit of this approach.
i i
!t
Mr*.
R. C. IIaynec Page 3 (6) An inventication han been initiated to evaluate the poccibility of separating the AC feeds to the power supplies from the common distribution panel.
It nhould be emphanized that this incident was the result of a unique test configuration which does not exist under normal operating conditions.
Our review has determined that the existing Peach Bottom design does not present a safety concern, but may present a problem from a reliabil ity viewpoint.
We will pursue the optionn outlined above in order to encure reliability and to preclude future occurrences.
If you have any questions or recuire additional information, please do not hesitate to call.
Very truly yours, Y
%sf M.
' Cooney &
S erintendent Generation Division / Nuclear Attachments cc:
C.
J.
Cowqill Site Inspector
ATPACIEENT I ECCS POWER SUPPLIES EIFA TOPAZ Power Inverter Supply
_ l (+)
To Transmitters 12$ Volts DC and Trip Units
()
Channels B & D s
s
_(+)
s 120 Volts AC 4
I-T
\\
3 1
~
m!
t-Power y'
(
f
.^
- i.
s.
' Supply A
s 3
(
+
1
=
r
)
j s.
s
- g <x
~
-7 i.
, s j
- t.
8 s,
s N
,,(
1 -d g
s s
it 0
e i
g g
A w.
J y
,f s
y_o_
_ N s.-
s
(
/,
?Q', [l w '
e l
l w
l 1
. f
, s,/
,a
(<:'.',/>
"/
ATTACHMENT II
' ?'z
' -CORE ANALYSIS FOR PEACH BOTTOM UNIT 3 HPCI/ECIC INJECTION JUNE 19, 1982 During the HPCI/RCIC injection, APR:4 power increased froy 100t to between 112% and 114%, based on strip chart r e c o r d,e r s.
Alarm typer printout showed 4 LPRMs increasing above their alarm set point between 10% to 20%.
,At the time of the transient, reactor conditions from the P 1 at 1200 (transient occurred at 1325) were as follows:
- CTP 3280 MWT
%'PWR' 99.6%
~ Core Flow 101.96 s.CPRRAT 0.940 at 33-46 MFLPD 0.929 at 33-46-12
'MAPPAT 0.845 at 09-32-5 (from P-1 at 0800)
MLUGR 12.5 KW/FT MFLPD ir; Bottom "of-Cote
.866 at 11-42-5 MLHGR in Bottom of Core 11.6 KW/PT Tm'3'khermallimits (Critical Power Ratio, MAPLHGR, and LUGR) and,PCIOMR are considered.
CPR:
The Inadvertent HPCI Injection transient is considered in the performance of plant safety analyses and has been determined to be bounded by the similar but more severe Loss of Feedwater Heating (100 F) transient.
For this type of transient, the CPR safety limit (1.07) will not be exceeded if the transient l/
initiates with a CPRRAT = 1.0 (CPRRAT =
l critical power operating limit actual critical power ratio) and scram occurs at 120%.
Since the actual CPRRAT was < l.0 (0.940), the APRM scram setpoints were set at 117% to 118% (verified on 6/20/82), and power j "~ -
only increased to about 114%, the transient did act cause the CPR safety limit to be exceeded.
.~.'
.~
MAPLHGR:
MAPLHGR protects against exceeding peak clad temperature of 22000F during LOCA.
Since MAPRAT was 0.845 and a 14% overpower event would increase MAPRAT to a maximum estimated value of 0.96 - 0.98, MAPRAT did not exceed 1.0 during the transient.
There is no MAPLHGR Concern.
MFLPD:
The MFLPD thermal limit is designed to prevent the fuel clad from exceeding 1% plastic strain assuming APRM scram occurs at 120%.
Since power increased to approximately 114% and MFLPD was < l.0 (0.929), the transient should not have caused the clad to exceed 1% plastic strain.
PCIOMR:
The effect of the transient is analyzed for two locations in the core.
Because the plant had been operatina at steady state full power for more than 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />, the power level just prior to the transient is existing power level.
1.
At the MFLPD (location 33-46-12) LHCR =
12.5 KW/FT.
Assuming an increase to 114%, KW/FT = 14.25 - 12.5 = 1.75 KW/FT.
2.
At the largest MFLPD in the bottom of the core (location 11-42-5) LHGR = 11.6 KW/FT.
Assuming an increase to 120%
based on increase in LPRM readings, KW/FT
= 13.9 - 11.6 = 2.3 KW/FT.
Based on General Electric's review of test data and analysis, overpowers of up to 120 percent for ten minutes are not expected to cause throuch wall cracks with the release of fission products from the fuel rods, hence the risk of fuel failures due to pellet-clad-interaction is judged not to be significant.
i
.