ML19224A788
| ML19224A788 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/09/1979 |
| From: | Kammerer C NRC OFFICE OF CONGRESSIONAL AFFAIRS (OCA) |
| To: | Gilinsky V, Hendrie J, Kennedy R NRC COMMISSION (OCM) |
| References | |
| NUDOCS 7905300307 | |
| Download: ML19224A788 (15) | |
Text
t
- s
{
(
(,e w
,r D
- .,o.
.,,}
I 4
F n
F,$k MEMORAIiDUM FOR: Chairman Hendrie Ccnaissioner Gilinsky g~ac Ccmaissioner Kennedy 9,
Cocaissioner Bradford C
Comaissioner Ahearne FRO:l:
Carlton Kammerer, Director Office of Congressional Affairs
SUBJECT:
BACKGROUND MATERIAL FOR HART HEAh:tlG Attached are t'.,o items prepared by flRR in conjunction with the hearing to be held tcmorrow by the Senate Subccaaittee on iluclear Regulation concerning Three Mile Island.
The first is a " Preliminary Cescriptior of Events at the Three
!iile Island 2 Facility Accident". The second is a sum.ary of
" Generic Considerations of TMI-2 Incident".
Attachment:
As stated cc: GPE 000 SECY I;
4]
N_ N f
9.{
168 034
- 1.,l
- m J
se 4
1
'. s
.s
^%~.;.' ' '. $
. l..
o "'a
- i
~-.a.:*
. \\, :. i%
i oar:*
...L.;
,J.,
{
MAC ;oax 3 4 (7 76) Nacx c:44 o..........me....v......c:
..,. i......
[
79053 00 3'C7.d
(
(
~
GENERIC CON 5ICERATIONS 0F T"I-2 If!CIDEig, The following are the significant sequence of events that occurred at
=
TMI-2.
m-gj a.
The turbine tripped due to loss of main feedwater,
~
b.
The reactor tripped, c.
The auxiliary feedsater pumps started but flow was not auxcmatically established, g..
d.
The pressurizer relief valve apparently stuck open,
. =.
i+--
e.
The high pressure injection was turned off, and f.
The reactor coolant pumps were turned off.
E.
I.'.'
Our preliminary evaluation indicates tha the incident may have been
~~
k ccmpounded by misleading indication of primary system water level.
In addition, the consequences of 'he incident were increased by the lack of prompt automatic containment isolation.
g The initiating event, i.e., turbine ' trip and subsequent reactor trip, l.e are anticipated events in that they are expected to occur during the plant lifetime and the system is cesigned to respond safely.
In fact, other B&W designed operating plants have experienced these kinds of transients and have responded safely.
As a result of our preliminary r..
evaluation of the TMI incident, however, we have preliminarily identified several human, design, and techanical failures.
They are all essentially elated to the loss of feedwater (item c above), the turning off of the hign pressure injection, (item e above) ar.d the turning off the reactor 168 035
(
<(
c::iant pumps (item f above).
To en;ure that these potential human, casign, and mechanical failures do not result in a similar incident at other operating facili-ties, we have directed (via IE Bulletins) owners of facilities with 55W reactors to take several steps to ensure that safety margins are maintained.
In addition, we have formed an NRC Task Force to review in detail the causes of the TMI-2 incident and upon completion of these efforts will. take subsequent actions as appropriate.
[iif These The Task Force report will be completed about the end of this month.
NF.C ac ns are also being taken at this time bacause of the preliminary nature of our evaluaticn.
Certain additional information will be
~
developed which will provide additional insights into the actual causes' and ccnsecuences of the various actions during the event.
At this time, however, our preliminary understanding of the event is sufficient to enable us to define the immediate actions required of operating facilities with B&W reacters to prevent such an occurrence at these plant 2, anc provide us with an adequate basis to allow continued
~
Operation of these facilities.
First, the incident at TMI-2 was initiated by a loss of auxiliary feedwatc/ following a turbine trip (item c above).
Since plants are not designed and evaluated for the complete loss of all feedwater, we f
have taken steps to ensure that the emergency feedwater system will be available to inject water uncer this situation. At TMI-2, the block valves in the discharge lines from the auxiliary feedwater pumps stere closed. We have required that operating facilities with B&W 168 036
~
(
(
- reactors ensure hat these valves ~are always open by requesting their specific examination of these valve positions.
The position indication will be further verified by a full-time IRC IE inspector at each of p
these plants.
L...
Another generic aspect of this event and a significant contributor was the apparent sticking open of the electromatic valve on the pressurizer.
Licensees are being requested.to examine their procedures such that operators are aware all valve positions, including the backup
[
ri block valve to the relief valve, and have information available to a
Z:
permit its use.
The sticking open of the relief valve was a significant contributor to this event and w0uld be considered an important event regardless of whether auxiliary feedwater is available or= not.
t The second significant concern, which also has generic considerations, y
concerns the turning off of the nigh pressure injection system.
In general, we have recuested all operators of plants to exercise extreme e
[p caution before turning off any safety system.
Specific;11y, we have p
taken steps to require operators to maintain high pressure injection
[
for a minimum of 20 minutes if it is automatically actuated.
This occurs on low reactor pressure (1600 psi) in the reactor.
We further require that high pressure injection be maintained until stable conditions are obtained. We believe such actions may cause operational inconveniences, but that they are not significant when compared to the gain to be made should a severe transient occur.
We also require 168 037
i
(
(
chat the Lp: be aintained for ' minutis following any low pressure transients, including t5 cese whors a relief valve inadvertently opens anc sticks ope..
ta ensure pump coolant inventory.
Finally, we are recuiring that if the reactor coolant pumps are in p
~
operation when a severe feedwater transient might oce;r, they should be kept in operation if at all.possible.
Furthermore, if possible, one should be i.ept running in each loop.
This requirement provides an L:
p[::
extra l'. vel of safety.to cover a broad variety of transients.
In developing this requirement, it was recognized that operation of the RCps under certain conditions may damage the pump due to cavitation, however, it is believed that such operation is appropriate to ensure adequate response to a wide variety of transients.
[:
The above mentioned staff requirements, in addition to our requirement that all licensees with S&W reactors review their designs, have been
{;
imposed to maintain and possibly increase margins regarding their l {'
j t~
respor.se to feedwater, and other, transients.
Such actions will, we itl':
p..
~
believe, compensate for any remaining generic c.oncerns regarding B&W.
reactors response to such transients.
Because the accident situation appear: ic have been further complicated by the containment not being isolated '.'pon ECC5 actuation, (in this case HPI), we have also taken steps to ensure that the containment is isolated to the extent possible given any particular event.
168 038
(
(.
The above mentioned cor.siderations.have~ been directed tcwards B&W reactors because they app. ear to be the ones most directly affected.
. 2-The described actions are intended primarily to be short term actions and may well be modified as a result of the fGC Task Force review of f
B&W reactor transients.
Certain of these interim actions may also later be shown to be tpplicable to other pressurized water reactors, i.e., those designed by Westinghouse and Combustion Engineering.
These facilities have significant design differences. Two significant
[b"-
differences between the S;W primary system desicn and those of Westinghouse and Combustion Engineering are, first, the primary system water level for Westinghouse and Combustion Engineering are pore directly measured. The operators would, therefore, have had a more direct indication of primary system water level and would have been so influenced before stopping high pressure infection flow.
b Secondly, the steam generator volumes are larger for Westinghouse and Combustion Engineering plants, and therefore are less sensitive to feedwater transients allowing more time to detect and correct any deficiencies in auxiliary feedwater flow.
Therefore, we have not required any actions of licensees with Westinghouse and Combustion Engineering designed plants at this time, although we have been sending them infomation copies of all actions required of owners of B&W reactors.
168 039
(,
I' Q i L
~
i.
s
\\,
P?.ELIMINARY f.
DESCRIPTION OF EVENTS AT T.iE THREE MILE ISLAND 2 L
FACILITY ACCIDENT ihe following is a summary of the significant events that occurred
{::
at the Three Mile Island No. 2 nuclear facility on March 28, 1979, and
{
E thereafter.
Attached is a detailed chronology of these events listed with the times they each occurred.
Dkh t;
At about 4:00 am on March 23, 1979, the secondary (nonnuclear) cooling n
i::
system of the Three Mile Island facility suffered a malfunction.
This
])
it=
system normally pumps water through the plant's steam cenerators where
[?
k the water turns to steam which then flows to turn a turbine generator.
[F
~
The water is then condensed back to water, is pumped by a condensate I.
pump through a clean up system, through a feedwater pump, and finally back to the steam generators, and continually flows around this loop.
A malfunction in the main feedwater system caused the feedwater pumps to turn off (trip), which in turn caused' the turbine-generator to turn off and stop generating electricity.
Since the steam generators I
were not removing heat due to the stoppage of feedwater flow, the reactor
[
f-coolant system pressure increased and the pressurizer relief h
r valve opened to reduce reactor pressure.
Immediately, the reactor turned off by the rapid insertion of the plant's control rods (scrammed) as designed and the nuclear chain reaction stopped leaving behind only residual, or decay, heat.
These events all occurred within the first 30 seconds following the event.
168 040 t
f..*.:.'...
Up to this point, this sequence is~ normal and the auxiliary feedwater system should startup and deliver secondary coolant to the plant's two o
steam generators to remove heat.
In addition, the pressurizer relief valve should close as reactor pressure decreases.
=~
v
~
All three of the auxiliary feedwater pumps started but were unable
}.[E to deliver flow because their flow paths were blocked by closed. valves.
In addition, the pressurizer relief valve failed to close and there# ore allowed the reactor coolant system pressure to continue to decret se.
jf(
'iG As the reactor pressure reached a preset value (1000 psi), the plant's j![
- s. i.. ;
Emergency Core Ccoling System (ECCS) started as designed and began to
=. -
i+
I?-
inject cold water into the reactor.
It is at this point that an indication of a rapidly rising pressurizer level apparently led the plant U
.L :.
operators to terminate the ECCS flow. At this point the Three Mile 1..
i:.
4..
Island incident had been underway for 11-12 minutes.
fr e
Between about 1 and 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> into the transient, the operatcrs turned
~
off the four large pumps which circulate the reactor coolant through the reactor.
It is fcilowing this action that we believe the severe damage to the nuclear fuel began.
For the next several hours there
=
was a very large temperature difference across the nuclear core E
indicating little flow of coolant through the core.
166 041
=
1
(
(
f t i.
- _rir.; this several hour period, when severe fuel damage is occurring, primary coolant from the. reactor primary coolant system was being dumped onto the reactor co 'tainment floor from flow out of the pressurizer ei.
relief valve and throucn the drain tank.
This coolant, which contained radioactivity, was partially pumped from the reactor containment building floor to tanks in the auxiliary building.
The tanks overflowed
[y.{
permitting radioactivity to be vented from the auxiliary building.
This g
p situation lasted until about 9:00 am when the reactor contain ont was E.
sealed (isolated).
c k=
F=--
During this time, from about 5:00 am until 8:00 pm, the licensee tried E
to depressurize the reactgr coolant system sufficiently to be able to
- h turn on the residual heat removal system.
Since his attempts failed, it was decided to repressurire the system.
N[
- p N
- t; After re:ressurization, one of the main reactor coolant pumps was j(
restarted and ' flow through the reactor core was re-established.
~
t Since feedwater was being provided to the steam generator, heat was being removed and the reactor system was slowly cooled.
Reactor cooling has essentially been in this mode since that time.
=
go e e 168 042
?RELIMIN'RY CHRONOLO3Y 0.:
TFE l**RCH 28, 1979 ACCIDENT AT THREE MILE ISLAND f
Time (a:oroxicate)
Discussion of Events Before ':00 a=
TMI operator working on Feedwater System.
4:00 am The loss of all (main and auxiliary) 11 feedwater flow occurred while the reactor was operating at 95% power.
The transient Mi" was initiated by a loss cf condensate 15 pumps.
The turbine tripped.
45 3-6 sec later An electromatic relief valve opened to
- E relieve pressure in the RCS* (2255 psi).
t+-
EM 9-12 sec later The Reactor tripped on high RCS pressure jgE (2355 psi) to terminate the nuclear EF reactor and reduce power generation to
][;.
decay heat alone.
17
!E~
12-15 sec later The RCS pressure decayed to the point
+--
(2205 psi) whera the relief valve should have reclosed.
The RCS continued to
~~
depressurize for about the next two hours.
15 sec later The temperature in the RCS hot. leg peaks at about 6100F with a pressure of about 2150 psi.
j.
30 sec later The auxiliary feedwater pumps in both safety trains (1 turbine driven pump and 2 electrically driven pumps) were j._
started and were running at pressure i;
ready to inject water into the steam 31 generators and remove the residual heat produced in the reactor core.
No water i=
was injected since the discharge valves l-'
were closed.
-Throughout, RCS denotes " reactor coolant system."
O 168 043
(
(
p '
i, ir.e (accroxir.a te)
Discussion cf Events 4:01a$
The pressurizer level indication began to rise rapidly.
The steam generators, A and 5, had icw levels of water and were drying out.
p-.
4:C2 am The ECCS was initiated as the RCS j.:
pressure decreased to 1600 psi.
~
4:04-4:11 am The pressurizer level indication went offscale high and the operator manually tripped the first HpI pumps at about 4:04:30 and the second at about 4:10:30.
id 4:05 am Water in the RCS flashed to steam as the pressure bottoms out at 1350 psi. The hog leg temperature was about 5850F.
4:07-4:08 am The Reactor building sump pump came on.
inE-r:
4:08 am The operator opened the valves at the il E
discharge of the auxiliary feedwater pump allowing water to be injected into the steam generators.
p 4:11-4:12 am The operator restarted the ECCS to inject F
water into the RCS to control pressurizer i
level.
j..-
,j..
4:11 am The pressurizer level indication comes
[
back on scale.
4: 15 am The RC Drain (Quench) tank rupture disk blew at 190 psig cue to continued discharge i
of the relief valve that had failed to open.
4:20-5:00 am The RCS parameters stabilized at a satur-ated condition of about 1015 psi and
[-
5500F.
g..
w..
5:15 am The operator tripped both RC pumps in y..
Loop 3.
5:40 am The operator tripped both RC pumps in Loop A.
168 044
(
(
Time (accroximate)
Discussion of Events 5:45-6 am The reactor core began a heatup transient.
The RCS hot leg temperature went of fscale at 620 degrees F within 14 minutes and the cold leg temperature dropped to near the temperature of high pressure injec-tion water (150 degrees F).
b
}s 6:20 am The failed open relief valve was isolated M
^^
by the operator by closing a block valve.
The operator also isolated steam generator 3 to prevent leakage of radioactive E
secondary water from leaking S.S. tubes.
b E
7:00 am The RCS pressure had increased to 2150 psi and the relief valve was opened to relieve RCS pressure.
7:15 am A pressure spike of 5 psig occurred in the RC drain tank due to steam from the relief valve.
7:45 am A pressure spike of 11 psig occurred in thc 2C drain tank and the pressure in the i CS was at 1750 psi.
9:00 am The pressure in containment peaked at 4.5 psig.
9:00-11:00 am The RCS pressure increased from 1250 psi to 2100 psi.
11:30 am The operator opened the pressurizer relief valve to depressurize the RCS in an attemot to initiate RHR coolino at 400 psi.'
~
b:
h 12:00 am - 1:00 pm The RCS pressure decreased to about 500 l
psi and the core flooding tanks partially 3
discharged.
The relief capacity was not sufficient to vent enough to reach 400 psi, 2:00 pm The pressure in the containment spikes at 23 psig causing containment sprays to be initiated.
The cperator stopped the spray pumps after about 2 ninutes of operation.
168 045
(
I Time (acoroximate)
Discussion of Events 5:30 pm The pressurizer relief valve was closed in order to repressurize the reactor coolant system.
5:30 - 8 pm The RCS pressure increased from 650 psi to 2300 psi.
8 pm RC pump in Loop A was started at which time the hot leg temperature decreased to about 560 degrees F and the cold leg p~
temperature increased to 400 degrees F, indicating flow through the steam p
generator.
Thereaf ter, the reactor was t
being cooled by reestablishing ~ condenser vacuum and steaming to the condenser by steam generator A with the RCS cooled
- r to about 280 degrees F and 1000 psi.
y March-29
$p The ;I5 temperature and pressure was stablized at about 250 degrees F and 52C to 1020 psi.
The maximum reading on the incore thermocouples was 612cF, but several were not with range for computer readouts (printing "?")
which was subsequently found to indicate greater than 700 degrees F.
P.
Iu Narch 30 The RCS tenperature and pressure was stable at nearly 230 degrees F and between about 1000 to 1050 psi.
Several incore thermocouples were beyond the range for computer readout, the maximum indicated reading was 659 degrees F.
The f;RR staff estimated the bubble size in RCS to be about 1200 ft3 and recuested the licensee to refine their calculation of the bubble size.
March 31 k
0 The RCS temperature and pressure remained stable at about 280 F and 1000 psi.
Slight drop in pressurizer level 251-191". Temperatures in the core as measured from the incore ther occuples were gradually decreasing (maximum indicated about 5000F). The hydrogen recombiner was in an operable status but additional shielding was needed and was being cbtained.
Two samples of containment atmosphere were analyzed wnich showed a hydrogen concentration of 1.7% and 1.0%.
Licensee calculated bubble size to be about 620 ft3 g 875 psig.
168 046
(~
(
Acril 1 No substantial change in RCS temperature and pressure
-[
Incore thermocouples continue to show decreased u'end.
Licensee continued hookup of hydrogen recombiners and addition o.-
shielding.
Licensee calculated valves of bubble size varied.
Containment air samples indicate 2.3% hydrogen.
g7-k5 Aoril 2 g
Reactor pressure stable at about 1000 psi.
Incore thermocouples continued
~
to show a decrease with all measurements below 4750F.
Inlet and outlet temperatures were still about 2800F.
One hydrogen recombiner was put in operation.
Analysis indicated that the oxygen generation rate in reactor less than p
originally estimated. Measurt ents indicated that the bubble was E7~
being significantly reduced.
~~
Acril 3 0
Reactor pressure and temperature stable at 1000 psi and 280 F, respectively.
Thermocouple readings analyzed-maximum 4770F, only 3 thermocouples were 0
above 400 F.
Gas bubble size much reduced.
Conta'inment about 1.9%
g hydrogen.
One pressurizer level indicator failed.
- F x-Acril 4 4_
Peactor pressure and temperature stable at 1000 psi and 2500F, respectively.
Thermocouple maximum temperature was 4660F.
Gas bubble size decreasing.
Vent valve on pressurizer intermittently opened and degassin.1 continues throuch letdown system.
- c..
Acril 5
-I J.
4 0
Reactor pressure and temperature stable at 1000 psi and 280 F, respectively.
1 Maximum thermoccupie reading is 4620F.
Pressurizer level responding normally to pressure changes indicating a completely full system.
p
$.+
Containment atmosphere indicates 2% hydrogen.
One recombiner operating, one in standby.
Pressurizer vented to containment about 15 minutes every
~
6-8 hours.
168 047
~
(
(
Acril 6 Reactor pressure stable.at about 1000 psi and temperature about 285cy, At approximately 1:25 pm, reactor coolant pump 1 A tripped and reactor coolant pump 2A was started within about 2 minutes.
Shift in thermo-m.
couple readings.
The three thermoccuoles previously reading about 4000F are presently reading between 2550F and 3150F.
Central thermo-0 couole increased from 3750F to 425 F and is the only one reading about 0
i@:
400 F.
t-Centainment neasurements indicate about 2% hydrogen.
Pump-back system for pumping waste gas decay tank volume to contair. ment began.
Acril 7 0
Reactor cressure and temperature stable at about 1000 psi and 250 F, respectively.
i,,_.
i= P At about 8 pm. the licensee began to slowly lower reactor system pressure.
The slow decrease will end wnen reactor pressure reaches 500 psi.
This W
is a step toward cold shutdown and ir.cludes degasification to prevent
!1 E
bubble formation as pressure and temperature decreases.
Hydrogen concentration in the containment is about i. 7%.
m-e p
e p..f;L 168 048