ML19224A922
| ML19224A922 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/13/1979 |
| From: | Cockrell R WASHINGTON PUBLIC POWER SUPPLY SYSTEM |
| To: | |
| Shared Package | |
| ML19224A921 | List: |
| References | |
| ACRS-M-0101, ACRS-M-101, NUDOCS 7906130220 | |
| Download: ML19224A922 (10) | |
Text
1 THE THREE MILE iSLN1D !llCIDEilT AND ITS IMPACT Oil WPPSS PROJECTS Presented by R. G. Cockrell to the Board of Directors April 13, 1979 We\\S HINGTON PUBLIC w.q -
w c' b
? 9 0 613 0 2ao POWER s 'Ei~.]
SU PPLY k[:lj.Ip>J SYSTEM t:Jf-~
PRESENTNi!ON Pr\\CKr\\GE 174 141
CONTEf1TS 1.
If1TRCDUCTION 2.
TMI-2 SEQUENCE OF EVENTS; MARCH 28, 1979 3.
liRC REPORT ON PRIMARY CAUSES FOR Tril-2 IriCIDE!1T 4.
CCMPARISON OF UPPSS PLANTS JJITH TMI-2 5.
POTEtiTIAL IMPACT ON WPPSS PROJECTS 1.
INTRODUCTION At 4:00 A.M. on Wednesday, March 28, 1979, Three Mile Island Unit rio. 2 (TMI-2), an 845 MWe Babcock and Wilccx pressurized water reactor was operating at approximately 981 power when feedwater flow was interrupted to both steam generators.
The initiating event apparently was loss of punp suction as a result of_ valving problers_in__the_ condensate systen.
This report briefly describes the sequence of events thatdoT4cwed which resulted in the most severe accident in the history of ccmmercial nuclear power.
Also provided is a synopsis of an NRC report on the primary causes for the TMI-2 incident, a comparison of WPPSS plants with Tl11-2, and some educated guesses regarding the potential impact on WPPSS projects.
174 142
2.
TMI-2 SEQUE"CE OF EVENTS; MAP.CH 28, 1979 1.
At about 4:00 A.M. one of two main feedwater pumps tripped auto-matically due to a loss of pump suction.
This loss o pump suction r
was a result of valving probicms in the condensate system which is installed ahead of the fcedwater pumps.
The trip of the first pump was follomd almost immediately by the trip of the second pump (we don't know why). (See Figure 1, area (1))
2.
The loss of all normal feedwater resulted in undercooling of the steam generators and an acccmpanying increase in the reactor coolant system tcmperature and pressure.
The increase in pressure resulted in the opening of the Electromatic relief valve on the pressurizer followed within a few seconds by the automatic shutdown of the reactor at about 10 seconds.
(Note:
This would have happened even if auxiliary feedwater had been obtained.
At this point it is only an enpected occurrence.) (See Figure 1, area (2))
3.
With the reactor shutdown and the core power im ediately reduced ta only a few percent of full power and the relief valve aischarging primary coolant, the reactor coolant pressure began to decrease.
4.
The relief valve should have reclosed when the pressure was reduced to the nortal operating system pressure; however, the relief valve remained cpened and the reactor coolant system pressure continued to decrease and the pressurizer began to empty.
5.
About 20 seconds after the loss of all normal feedwater, the auxiliary feedwater pumps catcmatically started.
However, some valves in the system had been left closed and auxiliary feedwater was not supplied to the steam generators. (See Figure 1, area (3))
6.
Eventually, the system pressure became so low, as a result of the loss of reactor coolant through the stuck cpen valve, that the high oressure injection system was initiated.
This system is designed to provide more flow than can be passed through a stuck open valve (as was the case) and the pressurizer began to refill.
(At this point, the failure to obtain auxiliary feedwater and the stuck open valve is still a manageable event.
Primary system heat removal is still available.)
7.
When the operator became concerned that the pressurizer might become overfilled he turned off the high pressure injection system.
However, auxiliary feedwater was obtained and with forced circula tion of reactor coolant by the reactor coolant pumps, the core remained cavered with water and cooled.
174 143
8.
At 5:40 A.M. the operator stopped the reactor coolant pumps due to potential pump damage that might result front operation at a low reactor coolant systen pressure (cavitation).
This loss of foi ced circulation resulted in boil of f of the reactor coolant, subsequent uncovering of the core and cladding failure (natural circulation was prabably not available because of the p esence of steam in the reactor coolant systcm; it is known that th2 system was at saturation at this time).
Fuel cladding failure resulted in release of firsion products to the reactor coolant.
9.
Eventually, high pre::sure injection was restored and water thus adoed to cover and cool the core.
A single reactor coolant pump was put back into operation and with the auxiliary feeduater having been previously restored, the system was put into a stable condition a t 280 F and 1000 psi.
10.
The continued release of steam and wa:_r through the stuck open relief valve had eventually passed se much steam to the quench tank that the rupture disc on the tank burst and water began to accumulate on the reactor building floor.
Also, fission products that had accumulated in the reactor coolant as a result of the failed fuel were released to the containment through the failed rupture disc.
(The quench tank is aesigned to rcmain intact for the worst expected transient, i.e., turbine trip because of reactor over power, but cannot tolerate the continued blowdown through a stuck open valve.)
11.
As the level of water increased in the reactor building, water was pumped to the auxiliary building.
A defective seal resulted in the leakage of some water and the release of fission products.
Additional actions nay have resulted in uncovering the core a total of three times.
In summary:
Length Estimated Peak Time
, Uncovered Claddino Temn.
(or~~12' total) 5:45-6:45 A.M.
5' 2000 F 7:45-9:45 A.M.
5' 1725 F 1:00- 3:30 P.M.
74
>2000 f (15' may have remained uncovered between 9:45 A.M. and 1:00 P.M.)
B&W has estimated that 17-31C of the fuel cladding surface may have become involved in Zr/H O reaction.
2 174 144
o t
s a
y reS er nunk eS on G Si ak tT n erc a
my, nP ehTjc i
bh nnp
- 3_
G r 9I er ui uu C ps T
T' OS i
~
i f.
TP'.
Gt QS I P i p A D. ~ s
Dc
' r r
u e
,3)
I'o P
t 1M
=
a P
f t r a
mi H
/
nord t
r,E p..-
t at ee l at e
)i
^y, rrrG ea roraF rH dr <
I e c e ;.
1l zC n i /-
C f\\
i eer S ps o"o fi t St n l
cS ccni d
aoa e i <r n
eret ta e
RPRSl A p
M o
c.
RPPGWU l
I CSFF
'g U
R X
W W
31 A
F y0 F
f\\
X 7
S U
y O g t' A
N i
=
)
'l j,,
l t
n j
e f
r
. f'A m
n s
i iI J
a t
o
(
n n
P G
C S
R g,t 1
1 (9%
H M
r u@!
)
-a h
m
(
t P
{
,4
~
T r._
>/
///-
i, R
rn,
^
^
^
[
^{n-
^
/
/
/
7
/
~
q
_[(
ytn
/7
3.
f1RC REPORT 0:1 PRI:'ARY CAUSES FOR TiiI-2 INCIDEtiT On April 4,1979, the f RC Staff (D. Eisenhut and E. Case) briefed the Commissioners on the TMI-2 Incident providing the following information:
Primary Causes o
1)
Cesign Errors 2)
Equipment Malfunctior 3)
Operator Errors Six Specific Items that Contributed to Seriousness o
1)
Redundant auxiliary feed. vater valves were closed. - Operator error (Violation of technical specifications.)
2)
Electrcmatic relief valve on the pressuriaer stuck open. - Equipment malfunction 3)
Pressuriaer level indicators appear to have given erroneous readings. - Equipment malfunction (l i 4)
Containment did not isolate on actuation of Emergency Core Cooling System.
This is considered a design error in that the design did i
not require containment isolation until containnent pressure reached q }k 4 psi.
Isolation did occur as designed.
5)
ECCS was turned off prematurely. - Operator error 6)
Reactor ccolant pump tric caused fuel damage.
The core was uncovered three times.
The first time was when the operator tripped the RCP. -
Operator error.
e NRC Plans 1)
Shcrt tern - Require operating plants to review operating procedures to insure that the high pressure injection system (part of ECCS) can-not be turned of f unless two low pressure safety injection systens are operating, and require that contairmnt isolation take place immediately upon ECCS actuation.
2) long term - NRC Staff will conduct an evaluation over approximately the next three months to determine how the Eil-2. failures may inpact other plants.
1/4 146
4 COMPARISON OF WPPSS PLANTS WITH TMI-2 4.1 WNP-1/4 WNP-1 and 4 are the rost like TMI-2 since they have Babcock and Wilcox Nuclear Steam Supply Systems.
However, WNP-1/4 is a more recent design and has features that probably would have prevented the accident, and had a release of radioactive ruterials into the reactor coolant system occurred, recovery would have been easier Specifically, Annunciators in the control rocm indicate any valves that nay a
be closed and thus blocking a systen required for safe operation of the reactor.
TMI-2 dia not have this feature and consequently the operator in the control room may not have.<nown that the auxiliary feedwater system had been isolated for surveillance testing and the isolation valves were still closed.
WNP-1 and 4 contairments have hydrogen recombiners.
TMI-2 did o
not have a reccnbiner and a hydrogen explosion occurred in the containment that increased the pressure to 30 psig causing the containment spray to actuate.
At TMI-2 it was necessary to proceed with extreme caution because of the potential for a hydrogen explosion.
WNP-1 and 4 have reactor coolant bleed degasifier which o
provides a direct method cf getting rid of gas in the reactor coolant.
This system bleeds reactor coolant at the rate of 50 gpm and degasifies it.
TMI-2 dces not have such a systen.
4.2 WNP-3/5 Although designed by a different NSSS vendor, namely Combustion Engineering, WNP-3/5 is a Pressurized Water Reactor and is much like WNP-1/4 Consequently, any regulatory changes that result from the TMI-2 incident will very likely be similar for WNP-1/4 and WNP-3/5.
4.3 WNP-2 WNP-2 has a Boiling Water Reactor which has a cooling systen design that is very different from a Pressurized Water Reactor as can be seen by comparing Figures 1 and 2.
The sequence of events that occurred at TMI-2 could not have occurred in a BWR.
However, WNP-2 will very likely be impacted by many of the regulatory requirements that result from THI-2, principally those changes thought to mini-mize operator errors.
174 147
FIGtBE 2 Secondary Containment v
/*
RCIC y,, Primary Contain ent l
E'f 'R I
=
,e,Y l
~i
\\
F X
\\ \\,
x z
~.
~
s
.I t i s
TG i
$RV( I l
RHR h 9
~
~
M r/
s m
o pH Legend l
l R Reactor TG Turbine Generator h
~
^
^
~
RCP Reactor Coolant Pu-'p a
g-RCIC Reactor Core Isolation Cooling y
$ st HPCS High Pressure Core Spray l
LPCS Lcu Pressure Core Spray E'
LPCI Low Pressure Coolant Injection
'k SRV Steam Relief Valve l
SP Suppression Pool
(
_, o CC
__ ~
4 5.
FOTENTIAL IMPACT ON WPPSS PROJECTS The following corrents are educated guesses.
Tt eate no new requirements stemming frcn the TMI-2 incident have been randated by the NRC for plants under construction.
- License Review
- Of the five units only WNP-2 is currently involved in an active licenso review.
Since it is a BWR, we do not expect licensing of it to be significantly inpacted.
The most likely impact will be the unavailability of key NRC review staff who are now tied up on the TMI-2 problem.
Hopefully, the NRC report on TMI-2 will be out well in advance of our application for Operating Licenses for WNP-1, 3, 4 and 5, so that we tay address any ch:nges in the regu-lations before filing the FSARs.
Operator Training
- Expect more rigorous training and more freouent e
requalification of cperatcrs.
The WNP-l/4 sinulator will help us in this regard.
Maintenance Staff It has been reported that the maintenance personnel worked 40 consecutive 10-hour days prior to the TMI-2 incident.
If this is determined to be a principal contributc-to equipment malfunction, then there may resclt ]
n stricter regulations in the area of plant 1
(
maintenance.
Potential ircact:
larger main-tenance force and increased QA/QC.
o Environmental
- The regulation that sets forth requircr.ents Qualification of in this area requires documented evidence Electrical : quip't that electrical ccmponents that are supposed to operate durina an accident have been tested in a similar environment.
This regulation is already difficult and costly to satisfy.
It will probably be extended to more components and may become more stringent.
Design 1)
Containment isolation system must be modified o
to cause isolation when the ECCS starts up.
2)
Options allcwing operator action will be
- limited, i.e., control will become more automated.
Potential inpact - reduced plant availability.
174 149
~
3)
The resolution of ATWS will be delayed because key NRC personnel are tied up on Tf11-2.
When ATWS is resolved, the regu-lation is c: ore likely to emphasize mitiga-tien than prevention and is not likely to give much, if any, credit for operator action.
Potential impact - signi ficant design changes on all projects with the greatest change on WNP-3/5 and the least on WNP-l/4 Emergency Planning
- The TMI-2 incident was the first real test of emergency planning for a ccmmercial nuclear power plant.
Undoubtedly, increased attention will be focused on this area at the local, State and Federal levels.
Potential impact - increased health physics staff, increased requirer:ents for an emergency coordination center, an emergency plan that provides for a larger evacuation area, establishment of better communications among local and State agencies, consideration of sig-nificant NRC involvement.
O 174 150