ML19224D218
| ML19224D218 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 05/24/1979 |
| From: | Dieckamp H GENERAL PUBLIC UTILITIES CORP. |
| To: | |
| Shared Package | |
| ML19224D215 | List: |
| References | |
| ACRS-SM-0136, ACRS-SM-136, NUDOCS 7907110111 | |
| Download: ML19224D218 (29) | |
Text
'
s \\, '
m g
Testimony Before the Subco=mittee on Energy and Environment of the House Cc=2ittee on Interior and Insular Affairs by Herman Dieckamp, President General Public l'tilities Corp.
.".a y 2 4, 1979 260 1~c
/
7 90711oni
t.
The accident at Three Mile Island on March 28, 1979 has had a profound and shocking impact on the residents of central Pennsylvania, Met-Ed and GPU, our customers and employees, and on the future of nuclear energy. While nuclear power plant systems and procedures have been designed to accoccodate extre=e =alfunctions of both equipment and personnel, the reality of this accident has had a far greater impact than we could hava ever projected.
We pledge our sincere support and cooperation in the efforts of this-co=sittee to scke known and to assess the full =eaning of this accident. At the outset we would like to emphasize that we do not in any way wish to minimize the significance of this accident and we seek no excuse from our responsibilities as plant owners s.id operators. We strongly believe that it is important to understand the factors which contributed to this accident and to the ability of our Co:pany, govern =ent agencies and the affected population to cope with it.
If this accident is viewed si= ply as a =atter of canagement or operator failure, the full significance of this experience will be lost.
The accident was a result of a complex cochination of equiptent =alfunctions and human factors.
The accident departed from the accepted design basis for current nuclear plants. The response of all organizations was influenced by the fact that it was the first accident of this magnitude in the history of the U.S. commercial nuclear power program.
It ir our hope that this testimony and these hearings can contribute to an understanding of this accident, the many complex factors that led to it, and the critical learning that we are obligated to derive froa it.
ACCIDENT CAUSES We would like to focus this portion cf the testimony on our initial is-pression of the primary causes of the accident. We do not propose today to present a detailed description or sequence of events for the accident. We are in general agreement with the NRC testi=ony on this subject as previously
<) D U 1 ~j ^r p
L ll i
1.
presented to the U.S. Senate Subconsittee on :!uclear Regulation. We say, however, differ somewhat on the rela tive importance of the various ingredients of the accident.
While extensive data and information have been made available Met-Ed and GPU have not completed a detailed reconstruction of the ace'. dent or verified the relative importance of the cany ingredients. The follow ng appear to be the major causes of the severity of this accident.
a) Shortly (4 sec.) after the turbine and reactor trip at about 4:00 a.m.
on :! arch 28, a reactor coolant system pressure relief valve opened to relieve the nor=al pressure excursion, but the valve failed to re-close after the pressure decreased. The operator was unaware the valve had not closed. An order for valve closure was signaled in the control room. The operator monitored te=perature near the valve to indicate valve position. However, the temperature did not clearly confirm the continuing coolant flow thru the valve. The loss of reactor coolant and accompanying reactor coolant systes pressure decrease continued for about two hours until the' operator closed the block valve which stopped the uncontrolled loss of reactor coolant.
b) The operator anticipated reactor coolant system behavior and immedi-ately began to add cake-up water to the system. When systen pressure decreased to 1600 psi about 2 minutes into the accident the High Pressure Injection (HPI) safety systen was automatically initiated.
Four i o five minutes into the accident the operator reduced injection i
of er f rom the HPI system when pressurizer level indicated that the system was full.
260 178
=
. c) Operator training and experience had emphasized the retention of a steam vapor space in the pressurizer, huvever, following the rapid depressuri:ation of the system, the pressurizer level indicator inferred a fullness of the reactor coolant system. This level indication led the operators to prematurely reduce HPI flow.
The operator apparently did not anticipate that continued depressuriza-tion could lead to steam void formation in hot regions of the system other than the pressurizer and that under these conditions his level or fullness indication was ambiguous and misleading.
d) Because of the presence of steam voids in the primary system, indi-cated primary coolant flow decreased. The operator turned off the main coolant pumps in ord.er to prevent damage to the pumps. The plant staff expected cooline by natural circulation. Voiding pre-vented natural circulation and prevented reestablishment of pumping.
e) An emergency feed system, designed to provide cooling to the steam generators in case of loss of the norecl feed water system, was blocked because of two closed valves. This system would have been available to provide secondary cooling. The operator discovered this condition and initiated secondary system emergency cooling by opening the closed valves 8 minutes af ter the start of the plant transient. The plant safety system surveillance program had called for the placing of these valves into the closed positi)n six times during the first 3 ranths of 1979 for testing of the operability of the pumps or valves.
The surveillance program required a verification of valve position twelve times during this period. The last test of the emergency feed system was conducted on the morning of March 26, about 42 hours4.861111e-4 days <br />0.0117 hours <br />6.944444e-5 weeks <br />1.5981e-5 months <br /> before the >brch 28 accident.
2b
7.
f) Primary coolant initially vented through the pressurizer relief was pumped into the auxiliary building because the containment design did not require isolation until building pressure reached 4 psi.
Continued plant operation required soco, transfer of fission products to the auxilliary building.
The first five of the above factors led to severe undercooline of the reactor core.
The fuel became. extremely hot and the integrity cf the fuel cladding was lost.
The first indication of fuel cladding damage and fission prcduct release came with high radiation alarms.
An extensive reaction between fuel cladding and primary coolant steam liberated large quantities of hydrogen gas into the pri=ary reactor coolant syste=.
The resulting configur-ation of the reactor core is still the object of analytical atte= pts to reconstruct the accident. At various times during the day of March 23 as the operators worked to reestablish control of system cooling, the core suffered additional overheating and da= age.
Forced cooling of the primary system was reestablished at about 3:00 p.s. on the 28th.
A summary sequence of events is attached as Appendix A.
Performance of the plant operators has been the subject of =uch specu-lation.
Their perfor=ance must be viewed in the context of:
1.
Ambiguous and contradictory information relating to pressurizer level and relief valve closure.
2.
The experience and training underlying the operators' emphasis on maintaining pressuriner level.
3.
The operators' awareness of equipment limitations.
4.
The time and opportunity to assial. ate large quartities data l
with varying degrees of physical and chronological availability.
2 b 0 1 n (t c i ou
The operators on duty at the time of the accident are a qualified and competent group. They performed their functions professionally in a period of extrece stress. Our own investigaticn and the many other governmental investigations will ulti=ately attempt to determine the role of operator performance in this accident.
PLANT STATUS - CURRENT AND FUTURE The plant is stable and in a cold shutdown state.
The fission product decay heat being liberated in the damaged reactor core / fuel is just slightly inexcessof1Mwthermal(0.04{offullpower).
This power level is normal for this time after a reactor trip.
The core is being cooled by the natural circulation of primary reactor coolant. No pri=ary systes pumps are required in this = ode of cooling. The average temperature of the primary coolant is about 175 F.
As a result of local flow restrictions associated with the physical damage to the core, the highest in-core thercoccuple reading is about 300*F..
The heat from the reactor is being rejected through one steam gener-e' ator and the plant condenser.
An i= mediate objective of the activities at the plant has been to establish a redundant heat removal path through the plant's second ateam generator and an intermediate heat exchange loop without using the plant condenser. This will enable the transport of the core heat tt gh the plant's two steas generators for ultimate rejection through r
.ndependent secondary paths.
The objective is to minimize the number of active cocponents that must func-tion in these circuits in order to ensure relicble heat removal.
i The plant has been in the natural circulation mode since April 27, 1979.
The plant's several and original e:ergency cooling capabilities are available to backup this cooling approach. One of these systens, the plant's decay 260 18:
T heat removal system has been the subject of a high priority effort to upgrade the ability of that system to si=inize releases to the environment while operating with high primary coolant radioactivity.
As pa r t of this effort, work has been under way to enable the installction of redundant backup modules in addition to the two that are part of the plant design.
DEVEL0?>2NT OF UNDERSTASDING The accident differed from the popular perception of co==on accidents because of the extended ti=e necessary to achieve a full definition of its scope.
In this case, the ti=e required to develop a reasonably co=plete understanding of the accident and its result was approximately 2-3 days.
It should be stressed that while the full i= pact of the accident was not fully evaluated, there was suf ficient understanding of system conditions to naintain plant cooling stability during this period. There were three key areas in which full evaluation required ti=e :
1.
Asressment of the degree of core damage.
2.
The generation of hydrogen gas during the accident and a)its potential impact on system heat transfer and b) its i=plications relative to core damage.
3.
The impact of continued operations on the potential for re-lease of radioactive =aterial from the plant.
The accident's initiating event was a loss of feedwater flow. During the first few minutes following this event, the plant staf f attempted to. recover fica what t.ey thought was a normal transient.
Beyond this tice, the plant behavior became inceasingly abnor=al.
The loss of coolant via the reactor coolant system relief valve was identified and the valve was isolated around 6 : 20.a.m.. At approximately 6:50 a.s.
several radiation alarms alerted the j D' U 1 n ',
o e O s.
L
. staff to possible reactor core damage.
In the time period of 5:30-7:30 a.m.
the reactor core beca a uncovered and suffered extensive da= age, including significant zirconium-water reaction. During the next 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />, the operators attempted a number of strategies to establish dependable core cooling. This objective was achieved abodt 8:00 p.s. on bbrch 28, at which time the plant sy=ptoss included:
a) Some local reactor coolar t temperatures were above coolant saturation tempe ra ture.
b) High radiation levels existed in the reactor contain=ent and the auxiliary buildings.
At this point in time the high radiation levels indicated that fuel damage had occurred but the extent was not defineable. The complicating presence of hydrogen gas in the primary system had not yet been detected.
A preliminary sequence of events was being extracted from the various plant records by the afternoon of March 23.
The data for the 16-hour accident period became available in suc=ary graphical form on the morning of March 29.
The probable occurrence of a zirconium - water reaction and the presence of hydrogen gas in the reactor containment building was deduced during the the evening of March 29 from containment pressure records that indicated a pressure spike during the accident. The size of the hydrogen gas bubble in the reactor coolant system was first ceasured.froa system data just after cidnight March 30.
The concentration of hydrogen gas in the containment building was determi.ed from analysis of the first containment gas sample taken about 4:00 a.m. on March 31.
The first quantitative data with respect to fission product release and degree of reactor fuel damage becace available via analysis of a primary coolant sample taken at 5:00 p.m. on.'brch 29.
The data P6q t n=
U
$ D.)
s
? on hydrogen and fission product release provided the basis for the next level of core damage evaluation.
The point of the above enuceration is simply to indicate the time neces-sary to gain insight into the scope of the accident and, in turn, to provide the basis for a meaningful analysis.
In any raview of the timeliness of the accident assessment, it must be recembered that, the plant management and staf f faced i sediate, continuing and first priority denands to maintain the damaged plant in a controlled and safe state.
RADIOACTIVE SLATERIAL RELEASES A release of radioactive caterials to the containment building occurred during the first forty-five cinutes of the accident when water was released from the primary reactor coolant syste= through the pressurizer relief valve.
This water was first contained within the reautor coolant drain tank in the reactor containment building.
Shortly after the initiation of the accident, pressure buildup in this tank resulted in the release of coolant to the containment building floor.
This coolant collected in the contain=ent building sump and was puaped into the auxiliary building sump.
The auxiliary building sump overflowed and resulted in ;everal inches of water on the floor of the auxiliary building. Operator action turned off the contain=ent sump pumps approxicately 40 minutes into the event.
Co nta inmen t isolation automatically occurs in the TMI 2 plaat upon a 4 psi pressure increase in the reactor building.
In the accident that occurred this pressure buildu, did not exist until 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> into the accident and thus contain=ent was not isolated until 3:00 a.m.
High fuel cladding temperatures produced by inadequate core cooling during the accident resulted in the breach of most of the fuel cladding in the 260 18(
. core beginning about #1 minutes into the accident. This failure of the first level of fission prod.
contair=ent resulted in the release into the pri=ary system of the gaseous fission products from the fuel-cladding gap and extended periods at high temperatures released a fraction of the fission products normally contained within the fuel pellets.
Af ter extensive fuel damage occurred, highly contaminated prisary coolant and gases =ay have entered the auxiliary building through a number of routes including reactor coolant pu=p seal leakage, instru=ent sample lines, and the primary coolant =ake up and let-down systems.
'w'e are conP.inuing to evaluate in detail the importance and contribution of these possible release paths.
Our analysis is impeded by the inability to physically exacine specific systems due to high radiation levels.
Continued operation of the primary reactor coolant letdown and makeup systems to recove gas from pri=ary coolant circuit resulted in a buildup of hydrogen, iodine, and r.oble gases in the reactor =ake-up and let-down systems and in the waste gas decay tank in the auxiliary building.
Steps necessary to restrict tank pressure levels, the taking of gas samples, and ef forts to discharge these gases back into primary reactor containment building resulted in a series of radioactive gas releases.
The largest of these occurred on Friday, March 30 at 7 :00 a.m.
The iodine releases from contaminated water in the auxiliary building and f roa gaseous sources passed through iodine filters in the auxiliary building with the result that net iodine releases off site were limited.
In recogni-tion of the inventory of iodine in the auxiliary building and the de teriora-tion of existing filters, charcoal filters were replaced and an additional charcoal fil ter vstem was installed in series with the existing plant filter a260 1n:
os
7 -
system.
This existing iodine inventory is being reduced by a factor of 2 every R days by radioactive decay.
NRC has calculated the highest integrated whole body dose possible to an unprotected individual continuously positioned outdoors at the plant boundary and thus totally exposed throughout the cccident.
This was 85 millirem and is consistent with the highest offsite dose ceasured by Met-Ed.
This exposure is equivalent to 2-3 chest x-rays.
In addition to the maximus integrated t.5cle body dose measured from the accident, the total dose to the population within 50 ailes has also been e valua t ed. The results of this analysis indicate that the aggregate whole body dose to the population within 50 miles (about 2 =111 ion people) was about 3550 person-ress largely from noble gases released through April 7, 1979. A report of the Ad Hoc interagency Dot e Assess =ent Group consisting of representatives of SRC, HEW, EPA indicates that the total potential additional cancer deaths associated with this whole body dose are less than 1, in addi-tion to the 300,000 cancer fatalities which would be normally expected to develop in the population of about 2,000,000 persons.
Low levels of iodine-131 have been detected in air and nilk sa= pled near the ite.
To date, measurements indicate the maximum level of iodin 2-131 in milk to be about 40 picocuries per liter. (pico = 1x10-12).
This level is below the 10CFR 20 naximum per:1ssible concentration of 300 picocuries per liter, and is below the levels of iodine in milk detected following the 1976 Chinese wenpons test.
Low levels of liquid releases occurred to the Susquehanna River through the industrial waste water treatment system.
The available data indicate 2()
. cuculative releases of about 0.23 curies to the river, well below the level of 10 curies per quarter allowed under cur license.
The releases have been below maxi =us permissable concentration (MPC) and below technical specifi-cations release races.
EMERCENCY PLAN Both Three Mile Island and the Cc==onwealth of Pennsylvania had formal written e=ergency plans in place before TMI 2 r aceived its operating license.
Under the e=ergency plans, there is a clear division of responsibilitp between Met-Ed and the state authorities.
In ter=s of the division of func-tions, it is Metropolitan Edison's duty to make an initial assess =ent of the accident, to do whatever it can to terminate or investigate the event, to read the plant instru=ents and monitoring devices which give an indication of the level of releases fro: the plant, to read the instruments telling wind direc-tion and speed, to dispatch tea =s of technical personnel to areas outside the plant with handcarried =onitoring devices to record ceasure=ents in the path of the plu=e and report these back to the plant e=ergency control center by radio and to keep the Bureau of Radiological Protection infor ea on all these =atters.
Plant personnel have been trained in these functions and perform periodic drills for various si=ulated accidents.
So f ar as state agencies are concerned, it is the responsibility of che Sureau of Radiological Protection to make the decision as to what measures of protection, including evacuation, should be undertaken.
If evacuation is called for, it is the responsibility of the state and local e=ergency centers to carry out the evacua tion.
E=ergencies which could have consecuences of' site are classified as either a Site E=ergency of a General E=ergency.
Site emergencies are those 260 187
+
7.
which have a potential for off-site consequences and General Emergencies are those with definite off-site consequences.
The emergency plans specify precisely the conditions in the plant which trigger the declaration of a Site of a General Emergency and which initiate implementatics of notification and intensified radiological conitoring procedures.
Both levt is of e=ergencies require notification of off-site authorities.
In the initial stages of the accident at D1I 2, the plant operators thought they were experiencing a normal plant transient involving loss of feedwater, which resulted in an auto =atic trip of the electric turbine generator and an auto =atic reactor trip.
About a half hour af ter the initial reactor trip, a radiation alarm on the intermediate cooling system was received.
In light of the operator's knowledge of the position of this detector in an area of generally high background radiation and its low setpoint, this was not viewed as an indicator of an emergency and it is not a criterica for declaring a Site or General Energency.
Thro ughout the next several hours there were no additional radiological alaras or other indications of the potential for off-site releases. At about 6:40 a.m. a radiation monitor iccated near primary coolant sampling lines alarmed and chemistry / health physics tech-nicians surveying with portable monitors in areas of the plant detected radiation levels.
It was not until 6 :50 a.m. almost three hours after the accident was initiated and the reactor tripped, that radiation monitoring devices in the unit alerted operators to the real potential for off-site releases.
At this time, the first criterion for declaring a Site Ener;ency was ce t, when a reactor building high range gacca monitor alert alar: was received.
In accordance with the emergency plan procedures, a Site Eme rgency was declared and notifications to authorities were initiated.
"'anaylvania's b
i
~
. Energency Management Agency was notified at 7:02 a.m., Dauphin County's Energency Center was notified at about the same time.
These organizations in tura commenced their notifications to state and local authorities.
The State's Bureau of Rac.iological Protection (BRP) duty of ficer was notified at 7:04 a.a.
by the State Emergency Manage ent Agency duty officer. The 3RP duty officer, thereaf ter, contacted the control room at Three 5 tile Island to gain technical knowledge about the event.
A call was placed at 7:04 a.m. to SRC's regional off'.ce in King of Prussia, Pennsylvania. The answering service which received this call was alerted to the reactor trip, the possibility of primary to secondary leakage through a steam generator, to the declaration of a Site E:ergency at TMI 2, and to the fact that no releases were known to have occ..rred at that time.
- !otification followed wi;hin minutes to others on the prescribed list of organizations to be notitied. About 7:24 a.s.,
the reactor building high range ga =a monitor high alara was received, which by the plan triggered escalation of the energency classification to the level of a General Emergency. Notifications of this new change in status were initi-ated.
During the period from 7:30 to 8:30 a.m. the e=ergency plans were fully initiated.
Cccounications both on site and off site were established.
Radiation monitoring teams were dispatched of f site to detect and verify releases.
Throughout the day of yarch 23, 1979, on-site and off-site radiological conitoring teams were providing a full flow of data to the Emergency Control Center at Three : Ele island. Constant communication existed through open lines I
from Unit 2's Control Roca to the Sta te 's Bureau o f Radia tion Pro tec tion and to i:RC's offices it Regica I in Kin; of Prussia.
As data was received at the site from radioicgical =cni: ring teans o f f site, it aas inmediately
'40 1 O h'
.UU iO/
7 i relayed to both NRC and to the State through the open-line channels estab-lished in the e=ergency plan and implemented on this occasion.
From shortly after 10:00 a.a.,
NRC had personnel in the control roon itself.
Fron our vantage point, the Three Pdle Island radiation energency plans and procedures were effectively i=plemented. The decisions to declare the Site and General E=ergencies were made by the individuals in charge when the specific criteria required these decisions to be ande.
Energency stations were canned and off-site notifications were cade and in accordance with the plan.
Open lines and a flow of eccaunications with the critical off-site agencies were established.
Radiation scnitoring results and plant status information was available and cocaunicated to both NRC and to the Pennsylvania 3creau of Radiation Protection. We must expect that further review of this experience will identify opportunities for improve =ent.
CRG.CIZAT!CNAI. RESPONSE In response to the initial perception that the plant was experiencing a severe transient, by approximately 7:30 a.m., Wednesday, :Mrch 23, available senior plant operations and technical support personnel were on site.
By ths; af ternoon two Me:-Ed and four GPU Service Corporation personnel arrived at the T:1I site to provide technicel assistance to the plant staff.
By evening pricary cooling was restored and the technical staff began the process of atte=pting to identify and understand the cause of the accident on the assumption that while there was sc=e fuel cladding damage, conditicas could be kept stable.
On Thur4 day morning, :brch 29, a seven-nan teac was dispatched to the si:e to initiate a detailed investigation in to the accident.
Jhen :he team gained a :irst hanc awareness o f the condition of the piant late Thursday a f ternoon, they immedia:21) turned their full attentian :o operating cecisions : hat had to be =ade, and identify in; continency ' lins in 260 i?0
_5 -
order to keep the plant in a safe condition.
This activity was a de=anding one and absorbed the approximately 30-100 personnel, about half fro GPU member cc=panies and half fro other utility industry companies, brought to the site over the next few days.
The GPU vice president who is responsible for generation plant design and construction, and who previously had been the Met-Ed vice president responsible for IMI, arrived at the site early Friday corning, Shrch 30, with plans for organizing and canning the ongoing effort.
Later Friday morning when a burst of radioactive gas was released fro the auxiliary building, awareness of the agnitude of the potential public i= pac: was sb rply increased and the probability of further gas releases was ide ntified.
During :he next 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br /> we were in phone contact with the nuclear industry.
We asked for support at the site in the form of senior experienced nuclear sc.entists, engineers, and technicians and found everyone eager to help.
By late Saturday af ternoon, thrch 31, about 30 people from 10 organizations arrived at the site to form the nucleus of what became known as the Industry Acvisory Group.
I et with the group early in the evening of Saturday, Shrch 31, and asked the group to organize itself to evaluate four price areas:
- 1) What problems do ae face in waste management to minimize offsite exposure?
- 2) What is the state of the dc: aged core?
3)
';ha t proble=s exist in the then current primary cooling code (with a bubble)?
- 4) What are the options available for progression toward cold shutdown?
Over the nex: three weeks, the Advisory Group utilized the skills and e::perience of about 100 nuclear s pr cialists.
Their participation has been these topics and began to work with the 35W staff in Lynchb
'la } (p'pto
7 access the capabilities of the other nuclear steam supply vendors.
'a'e we re atte:ptin; to deal with current and prospective problems that bore limited relationship to the design basis of the plant.
Eespite GPU's seventeen years of nuc'3ar involvement, our thirteen power reactor years of experience and a co=plement of over 1000 employees devoted to nuclear activitics, our resources and the lack of prior experience wich this kind of situation limited the rate with which we were able to in=ediately determine the plant status, to establish a plan of action, to deter =ine priorities and to supply =anagement leadership.
Curing the first few days af ter the accident the prioritics were iden-tified to be:
a) :-hintain the plant in a safe operating code with e=phasis on contin;2ncy plans including anticipation of component failures due to the high radia-rion levels and radiation inhibition to caintenance.
b) :Enini:e the fission product ac:ivity releases and the off-site exposures to the public.
The initial problen areas included waste water managecent, suppression of icdine release frca liquid spills, replacement of iodine filters, and filter additions, c) Devise and implement a safe transition from the post accident cooling mode to cold shutdown with provision for backup strategies to ensure continued safe removal of the core's residual hcat.
d) Rainforce the plant's anergency systems to assure safety in the cold shutdown code with its unique demands.
A critical acti-itt, was to imprave the integri:; of the deca;. heat renc'zal syster and ta enable :he installation of radundan: 32ckup cys: ens if re pired.
By Tuesday, April 3, the cc:bined ef for:s of the ':et-Ed/GPU staff, 160 1921
e
- L7 -
35U, and the Industry Advisory Group resulted in a Base Plan for transitioni.g the reactor from its post accident status to cold shutdown.
Subsequently, minor adjustments were made to the plan as a result of further independent review by the Advisory Group and NRC and as a result of the added 'nfarcation and experience gained by our staf f as a function of tice.
Cn Wednesday, April 4, an organizational structure for the TM1-2 recovery effort was put in place.
The organitation gave recognition to the continuing control of plant conditions, the need for significant engineering and analysis support, special emphasis on waste canage=ent, and leadership to the various plant :odification tasks. This overall r ganization waa placed under the direction of Mr. R. C. Arnold, Vice President-Engineering & Construction, of the CPU Service Corporation. A: the same tice the organizar'
- s bolstered by the infusion of a number of senior executives fro Duke Power Co. and Ccenonwealth Edison Co.
The organization was further strengthened by health physics and plant operations people from a number of utilities as well as numerous engineers from the nuclear industry. We wish to publicly express our gra:itude for the outpouring of support we were given.
COMPANY - NRC INTERFACE The role of the NRC and the relationship between the Company and the NRC has beea the cource of nuch speculation in the press. The Company's view of the relationship is o.e of mutual respec: and cooperation. The popular percep:1on of the relationship cay have been significan 1) colored by the Cc par v's 21ec tion to reserve con =ent on plant status and plans.
The SRC spokesmen adequately covered this aspect o f cc =unication.
It has been our judgmen: after the first few days and up to this time, tna; the public inter-es: was best served by mini.uiring the opportunity for media emphasis 0{3 ginar
')
T t, J
\\ /
Lv
?
- 13 -
nuances of expression.
A serious side effect of this policy has been to create the public impression that the Co=pany was not contributing to the canagement of the post accident efforts. We believe that Met-Ed and GPU f ulfilled their Itcensees responsibilities and have effectively responded to this accident.
The manage ent and resources made available by the Company for accident control cust be evaluated in light of the unexpected and first of a kind nature of this accident. As a result of this accident all parties should be
= ore aw.re of the de: ands of this kind of situation and better prepared to cope in teres of leadership, canpower and caterial reso ces.
In retrospec t,
it is our i=pression that the Company and the :iRC both experienced similar and so ewhat concurrent phases in coming to grips with the si t ua tion.
The question of ::ho was in charge was not critical factor. The Company ha; frem the outset recogn'. ed the role -f the '2C in this accident situa-tion.
The NRC's access to the control roon provided direct and innediate actest to plant status froa mid-corning of March 23 on.
The need for "RC approval of "of f normal" actions and procedures has occurred with limited bureaucracy.
The Company encouraged a cederation of the normal regulator /
regulatee relationship and invited the 'TRC to participate directly in the twice daily technical and pro;ress review meetin;s at the site.
There were tense so ents, but we cust emphasice that it is the Company's view that the relationship with the "RC is constructive and effective. We have been able to close canks so as to effectively employ our joint ri utrees.
rn,
.,L,bi-w v
7,c
-A.
a
--e i
oi t's respect to the langer term outicc.< for repair and return to service of T11 2, it is too ea'ly to be able to provide even a rough schedu.2 or cost 7
q L U' U
}n/ j.,
~ estimate.
Experience with the clean up and recovery of other reactor inci-dents suggests that the problem is technically =anageable.
It will, however, be significantly influenced by the availability of financial resources, regulatory requirements, and public acceptance.
The loss of IMI-2 generating capacity is co= pounded in impact by the continued shutdown of TMI-1.
Although technically able to be returned to service TMI-l will re=ain in the shutdown state until construction and cleanup activities associated wi' unit 2 have been reduced to a canageable level and until all regulatory require =ents developed as a result of the TMI-2 experience have been met.
De s pite the loss of this =ajor block of capacity in the CPU systen, there is adequate capacity in PJM to insure reliability of the supply of electricity.
Under current planning and current capacity schedules interconnection will be able to supply energy into the mid 80's.
While an adequate sapply of energy for Met-Ed, Penelec and Jersey Cet
.1 customers is likely to be available froa PJM, cost will be substantial.
Current estimates are that the cost of replacement energy for both units will average about 24 million dollars per conth.
This will Le reduced to 10
=illion doilars per conth when TMI-1 is returned to service. These lat e cost o
increases are a result of the low operating costs of nuclear unit vs the high operating cost of oil fired units.
By investing in nuclear energy CPU positioned its customers to take advantage of the econocies of nuclear power.
In fact CPU nuclear units have provided massive benefits to their customers.
Since its initial operation in 1969 the JCPSL Oyster Creek unit has proviced net savings to customers of over 3300 million dollars and is currently generating savings at a rate of about
') ' q 1 nr LOU i /J
e 7 70.ollars per year over the oil alternative.
The TMI-l unit since 1974 has provided in excess of $400 million benefits to 'St-Ed and JCP&L customers if operating would produce benefits at a rate of about 90 million per year.
In the short ti=e that it operated IMI-2 provided between 15 and 20 =illion dollars in savings. These benefits have been passed directly to GPU customers as they should have been.
Because of these past benefits, it is our view that it is appropriate for our custocers share in the cost of this accident. The elimination of all dividends to stockholders would produce only half of our currently esticated replacement power costs.
It is clear that ia the limit the shareholdar cannot bear -he entire impact of this accident.
GPU ha s theref ore proposed a sharing of the costs of this accident between cur stockholders, our exployees and cur customers, w.,ile the Ccapany canno t and does not seek to disassociate itsalf frc: the causes of the accident, we do believe thr t the accideat involved the entire technolo,tcal, and regulatory infrastructure of nuclear power.
The public is protected by Price Anderson.
The Ccapany ha s the benefit of property insurance for plant damage.
Beycad these, thers are the sig-nificant uninsured and uainsurable costs associated with replacement power and a large investmenc that may not be productive for some time.. If this unan-ticipated cost could be distributed over the 00 reactor years of concercial nuclear power to date, it would no t significantly detract from the econocios and value of this energy resource.
dowever, the cost of this accicent whan concentrated on the 1.3 nillion customers and the 170,0Cu 3 ccxnalders an_
260 196
s
. the other investors in T:!I 2's parent and subsidiaries is extreme. The traditional constraints of the utility regulatory process impose significant impedir.ents to :ne easy discussion of the raalfications of an accident of this type and a ready resolation of the proper sharing of costs between the cus-tocers and the investors. To date the industry brs 2nferestimated the in-portance of diversifying this financial rish and thus spreading the cost of the development of the technology over the total beneficiaries of nuclear power.
The institutions charged with the responsibility to supply a secure, abundant, and economic source of electrical energy must be able to withstand the impact of an event like the accident at TM! 2.
Tac system cast retain the ability to balance the social and economic coste of rnergy supply and energy availability.
I 260 197
Appendix A s
Preliminarv Descriotion of Three Mile Island Unit 2 Accident 1.
lornal operation The Th ee Mile Island 42 nuclear unit shown sche =atically in Figure I is a pressurized water teactor. The system normally operates with primary sys-tem te perature of about 5300 Farenheit and pressure of about 2150 pounds per square inch.
The reactor core (1) is the heat source in a nuclear power plant.
It is in this region of the system where the nuclear reaction takes place.
The rate at which heat is produc2d in the core is regulated by the control rocs. This is the system that shuts down the nuclear reaction when requir-ed.
In nor:a1 full power operation 2772 MW o f hea t is produced in the reacter core.
It is important te note that even af ter the nuclear reaction is stopped, heat continues to be produced by the fission products within :he core.
Inmediately af ter shutdown f rom full pcuer this heat is about 100 : ',',
a week after shutdown o MU, and a conth after shu:down decreases to 3 :ni.
The heat which is produced in the reactor core is transferred to the primary reactor coolant (purple) and circulated by the reactor coolant pu:ps (2) within this closed systes through the steam generator (3).
The hea t which has been produced in the primary system is transf erred to the secondary sys-tem (3reen) through the stean generators where steam is produced (light green).
This steam is then circulated to the turbine. The steam turns tha turbine (4) which turns the generator (not shoun) producing electricity.
This steam is then condensed (5) and is recirculated back to the seca: gen-crator by =eans of several pu=ps and heaters.
In-?.e schematic, only the condensate pu=ps (6) and feedwater pumps (7) are shown.
Tuo thirds of the heat produced in the primary s:ste: nust be discharged as uaste heat and la renoved from the secondary cooling systes by eans of cooling towers.
This bes: rejection system is shown in blue.
A key piece of equip ent in the accident was the pressurizer (3).
The pur-pose of the pressurizer is (a) to maintain the hi h pressure in the pricary d
sys:ca and to assure that primary coolant is maintained in a liquid or non boiling state and (b) to absorb changes in volume as the primary sys:en heats up and cools down.
In normal operation, the only steam present in the primary system is in the pressurizer (light purple).
The water level in the pressurizer (9) is in-dicated in the control room by the pressurizer 1: vel indicator.
If pras-sure in the primary s"1:en gets too high it is relieved by the automatic opening of the pressurizer reli2f valve (11).
2.
Accident Secuence en March 23. 1979 Approximate Time 4 a.:.
A calfunctica in the secondary system (3reen) caused a condeasate pump (6) to turn off.
This resultad ta (t =0) the Automatic tripping of both secondary feed puaps (7),
wn ch in turneu caused the turbine (4) to trip.
The tripping of the feed-water pumps caused a redue: ion of nua r2: oval to the staan generator.
lob
')b'Od
/u c
When neat re= oval froa the steam generator (9) was reduceu, it began to heat up and, in turn, the pri=ary systea began to heat up.
( = 2 sec. )
The loss of normal secondary feedwater flow caused the actuation of the energency feedwater pumps (10).
(t =4 sec.)
As the primary system heated up, pressure increased.
When it reached 2255 psi, the pressurizer relief valve (11) opened to relieve pressure.
In opening to vent excess pressure this valve was operating as expectad.
9 sec.)
The nuclear reaction occurring in the reactor was auto-(t
=
natically shut off as pressure reached 2355 psi.
At this point in the accident everything has occurred as would be expected and as designed.
12-15 sec.)
By venting thru the relief valve, reactor pressure was (t
=
reduced to 2205 psi; at this point the valve (11) should have closed but it didn't.
This was the first abnormal occurrence in the accident (NRC itec #2).
It should be noted that the operator was unaware that the valve had not closed. An order for valve closure was signaled in.the control room.
As tire passed the op-erator conitored temperature near the valve to indi-cate valve position.
However, the temperature did not cle ly confir= the continuing coolant flow through the valve.
The primary reactor coolant continued to vent through the open valve into the drain tank (12) and pressure continued to drop.
In response to the anticipated nor a1 transient be-havior the operators began to inject water into the systen through the =ake-up system (13).
39-4l sec.)
II.e e=ergency feed pumps (10) actuated at 2 (t
=
sec. achieved discharge pressure at 15 sec. were ca'. led upon to provide cooling to the steaa generator (3).
However, the block valves (14) on that emergency feedwater systen had been inadvertent-ly left closed (NRC ites 41) and the system was unable to function as intended. The relief valve (11) had already opened at this point so that the availability of the emergency cooling system on the secondary side could not have prevented the actuation of this valve which ultimately f ailed to close.
(
2 min.)
Pressure decreasud to 1600 psi.
A: this point :he
=
high pressure injection (liPI) emergenc;. core cooling system (ECCS) is automatically initiated (13).
2 min. 12 sec.)
The Drain Tank pressure increased to the point where snail (t
=
scount of coolant is released through the drain tank valve (not shewn) and begins to collect in reactor buil:ing su=p (15).
2,60
!99
s At t - 14 minutes 50 see. the drain tank rupture disc blew due to continued release of reactor coolant thru the f ailed pressuri er relief valve (11).
Large quantities of water begin to spill from :he drain tank.
(; - 4 min. 40 sec.)
Pressurizer level indication reached 90%.
Operator turned off one ECCS pump.
Pri ary pressure was now down to 1400 psi.
Following the rapid and continued depressurisation of the system the instrutent which measures the water level (9) in the pressurire.' inferred a high level throughout the reactor coolant syste This was due to the production of steas voids elscwhere in the pri=ary reactor systen.
Operator training and experience had emphasized the retention of a steam rapor space in the pressuriser.
The indication of a cajcr decrease in that vapor space led the operators to prenaturely reduce ECCS flow.
(NRC itea 13).
The operator apparently did not anticipate that continued depressurication could lead to steam void formation in hot regions of the systen other than the pressurizer and that under these conditions his level indication was ambiguous and risleading.
7 ain. 30 sec.)
Reactor building sump pump begins to pump water froa (t
=
reactor building to auxiliary building into the radio-active waste s:orage systen (16) (SRC item '4).
It shculd be noted that the operators turr.ed off the sump pump 30 minutes later at : = 37 minutes.
This.eas prior to any cajor fuel damage occurring.
3 min.)
The operators discovered the closed block valve (14)
(t
=
in the emergency f ed systea, opened it and initiated emergency secondary system cooling.
At this point no cajor f uel damage had occurred.
Sote at this point that :he relief valve (11) is still open and primary cool-an: was still being vented.
Steaa volds had been generated in the priaary sys:em preventing the normal flow of coolant.
Due to the ambiguous pru3sur-1:er level readings, the operators were unable to deter.2ine the state of :.
systec.
The operators made a number of attempts to verify system conditicas during the next 30 minutes During this time, indicated flow began to decrease and operators began to note reac:or coolant pump (2) vibration
-hich indicated cavitation.
(; - do-60 ein.)
Syste parameters in saturated conditions at 5200F and 10!5 p:n.
Indica:ed : low Jecreasing and ribratian in-
- reasing.
1 hr. 13 min.)
'he operator :urned off :wo of the faur primary can::m.
(t
=
- olan
- pumps (NKC itea
.,).
This action was :.s.u c by I' 3 oNrator in ord2r to p r > 'f e a t daca ;e :o :he pu.,s.
o'(J
,n r, p
.oa
1 hr. 40 min.)
The reactor operator turned ott the re=aining two (t
=
reac tor coolant punps.
- this point, primary pressure had reached 920 psi and there was no forced flow in the primary system.
The operators atte:pted to establish natur:1 circulation in the system. The unknown presence of the very large steam voids in the primary s/ stem prevented the operators from acco:plishing the natural circulation state.
1: was at this point that a heat up transient began to occut in the system and in the next hour, the major portion of the fuel damage occurred.
The lack of adequate ccoling caused fuel tesperature to increase to the point where the zircaloy fuel cladding reached a temperature where it reacted witn the hot steam and produced hydrogen.
This hydrogen gas was released to the pricary coolant systes.
Some of the gas ::as ultimately vented through the failed relief valve to the contain:ent building. Two points should be noted here:
1.
There was never a possibility of an explosion in the reactor pressure vessel due to the presence of hydrogen in the primary system.
2, Despite the high temperatures experienced in the fuel, coolant sample data indicate no fuel =elting.
2 hrs. 22 ein.)
The operator discovered the presence of the (t
=
failed pressurizer relief valve (11) and closed the reliaf valve block valve (not shown). This cut off further release of steam and water froa the system ar.d closed the primary systen for the first tire in over two hours.
(t = 2 hrs. 45 min.)
Reactor containment building radiation conitor indicates potential for off-site releases.
2 hrs. 50 min.)
Site emergency declared.
(t
=
(c = 10 hrs.)
28 psi pressure spike occurs in containment builting.
This was later deduced (evening of 3/29) to have re-suited from the explosion of a locally high concentra-tion of hydrogen vented f rom the primary systen, (t = 2 hrs 22 min.
The operators endeavored to restore prirary coaling to to t = 16 hts.)
the system.
However, the presence of large ar.ounts of hydrogen and steaa voids prevented this.
After attempt-ing a number of approaches to restore adequa:e cooling to the prinary system, the ope rators were finally successful in restarting a primary reactor cooling pu:p at about 3
p.m.,
16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> after the initiation o t' the accidenc.
The plant was cooled in this mode for several ;eee k s until the primary reactor coolant pump was shut down and the system.as brought to a natural circulation state an 4/27/79.
' fA
')
, 4 OU LJi
This information taken from US NRC D: 3ullet;a 79-05A, April 5, 1979 - Enclosure 1.
- UCLEAR I CIDE ;T AT THREE MILE ISL/J:D - SUP?LEST Description of Circumstances:
Preliminary information received by the NRC since issuance of IE Bulletin 79-05 on April 1, 1979 has identified si:< potential human, design and mechanical f ailures which resulted in the core damage and radiation releasas at the Thre2 Mile Island Unit 2 nuclear plant.
1.
A: the time of the initiating event, loss of feeduater, both of the auxiliary feedwater trains were valved out of service.
2.
The pressuri:er electrocatic relief valve, which opened during the initial pressure surge, failed to close when the pressure decreased below the actuation level.
3.
Following rapid depressurization of the pressurl:er, the pressuricer level indication =ay have lead to erroneous inferences of high level in the reactor coolant system. The pressutizer level indication apparently led the operators to prematurely terminate high pressure injection flow, even though substantial voids exisced in the reac:or coolant system.
Eccause the containten: does not isolate on high pr?ssure injec: ion 4
(UPI) initiation, the highly radioactive water fro = the relief valve discharge was put;ed out of the contain:2nt by the au:ccatic initiation of a transfer pucp.
This water entered the radioactive vaste traatecn:
system in the auxiliary building where some of it overflowed to the tloor.
Cu:;2ssing froa : hts water and discharge through the au::iliary buildin; ventilation systea and filters uas the principal source of :he offsite release of rcdioactive noble gases.
5.
Subsequently, the high pressure injection system 22s inte& 11:en1y operated attempting to control primary coolant inventory losses through the electrocatic relief valve, apparenti-based cn pressuriner level indication. Due to the presence of steam and/or noncondensible voids elacwhere in the reactor coolant system, this led to a further re-duction in primary coolant inventory.
6.
Tripping of reactor coolant pu ps during the course of the transient, to protect against pump danage due to pump vibration, led to :ue.
damage since voids in the raactor coolan: system preven:cd natural circulation.
260 202
a TMI-2 Schematic A<
.. s
/(11) Reliel Valve Containment Building
/
(9) Pressunzer
{
Rupture Disc Level
[
Turbine Building
\\
\\
l i
= - --
1 )I--)
o-o
,i f.,
3 G
,, j
, A c uu c,
f 7 (4) Turhine
.,A Y
(1 ) Orain Taak j
f_ y,p -
g[ _
_ Cooling Auxiliary Building
==-9ML - [ Y= - + --- /
Tower
,b
'-- T I,unipN T(13) Make Up!ECCS Lined (
b)
.(5)Contienser g[-
--S u
h 1-
_d
^
(7)F dwater Ph=L b)
(8) /
Circulating Water Pump (16)Radioatt Waste r-Borateil Storage Systein l
Pmssurizer n
U-
'h
,_%d'* - ~, Condensate Pump
. _.T L_ _..
(
] h, 3
% " ~ ~~= =
~ ~ = = = =
Water 3
)
3, Storage
_3_
Puinp Reactor (10) Ernergency eed Pump t
'b (14) Block Valve I
=.
U(
I Condensate Storage Tank
^
(16) Sump \\x__
(3) Stearn Generator N
u l?) ReaClor Coolant Pump N
CD ta