ML20079S214
| ML20079S214 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 12/12/1983 |
| From: | GENERAL PUBLIC UTILITIES CORP. |
| To: | |
| Shared Package | |
| ML20079S178 | List: |
| References | |
| NUDOCS 8402030246 | |
| Download: ML20079S214 (9) | |
Text
. . . .. - -. . - .-- . . -
GPU NUCLEAR CORPORATION December 12, 1983
':NSIDERATIONS IN l.OW POWER OPERATION OF TMI-I initial startup of TML-1 involves operating a plant which has been shut down for approximately five years. To minimize risk to the public and ensure the readiness of the plant and stan lor full power operaLion, GPCN propused n ,
deliberate step by step prel, ram of power escalation in the plant Startup Specification (reference SP 1101-06-008, Rev. 1). That Test Spec.ification
.' identified a period of approximately 30 days at about 48% of power operatfor.
followed by al>out '50 days at li% power. More recent suggebt lunn by t he NRC Staff center around 25% power operation for greater than 34) days.
The reduction in risk associated with reduced power operation centers principally about reduced decay heat at-reactor shutdown and correstnndinnly gicat or I imo la t ake operalor act ion in the event ol' plant a s e i den t : .
BACKGROUND:
In developing t.he plant Startup Test Sacification over a year ago, Ihe following basic criteria were used:
-1. .It would be desirable to limit initial extended power operation to less than flit)% in order lo minimite risk.
- 2. The first extended low power plateau should be at a power level at which all assential plant systems are being exercised and the plant :,t ab le.
- 3. It was highly desirable that the plant equipment be operating in a modo similar, if not identical, to the manner in which it operates at full power (i.e. , similar valve lineup, similar automatic control loops, etc.).
- 4. Plant sof t ware , i.e., procedures, etc.
should also be used in a manner similar to full power operation without excessive number of chance notices or other prouvdural modifications.
En oevelopment of the initial Test Specification, the above criteria were satisfied with som" margin with the selection of a 48': power operating plateau.
The det elmining item shi<h set this level was a dusire to have t he plant in a mode where it was least susceptible to feed system induced transients. To achieve this, the feed water train would have two feed pumps, two condensato, and two conhnsate b:>oster pumps in operation. This feature puts the feed system in its normal full power lineup and makes the plant least sustepIihle to feed svstem induced transients.
RR-ASSESSNI:NT OF l.0W POWER OPERATION:
The question of extended low power operation has been further reviewed given the recent interest in potentially operating at an even lower extended power 8#,02030246 840127 PDR ADOCK 05000289 P pg
=. - , . .-. . .. - --
Duccabar 12,'h8J
}3' O P:go Two l' .
st plateau and at a perceived lower risk. If one examines the various plant v conditions at various power levcis they find the followings low Power Endt At the low powcr un', d certain conditions set technical.llmits on operation.
.The turbine is normally brought on line in the 12-15% power range. [n the-rango of 15r20% power, Lhc once through steam generators are nol fully operating in the once through mode but tend to operate more like recircula-tion steam generators, and over an extended period of timo, would probably concentrate feed impurities. (No on line blowdown capability exists' for the OTSG's,) In the 15-18% range, .the turbine cannot be loaded for extended
. periods because the last stage blading will be operating far from its aero-dynnmic design point with Increased prohnbility of binde buffet inn or flut t er.
At 22-23% power, both trains of Condensate and condensate booster pumps are operating bul only one of the steam driven feed water pumps is in operation.
AL chis power icvel, feed water flow is just being transferred from the startup feed water regulating valves to the main (ced wrter regulating valves anti Ivvel syst e m cont rol rv<juires manual attent ion. At about 23 plant power, Lhc extracL ica Slcam system has marginal capability for normal (cod Waler heating and other systems are well off their design point. For example, at about 25%, only two powdex vessels are in operation in the condensate system, and they would be operating at about 2000 gpm cach vs. a design value of about 3300 ypm cach. in the power range of 28-30%, the
- tenderness" of the plant is largely climinated, the tecd system is controlled on the full flow regulating valve, all of the plant automatic control loops can be utilized for operation
-includinn that for feed pump control, and food water heators and drains would all be operating in a normal modo. -At the 28-30% power level, the steam gancralors should be operating in a once through mode, secondary side chemistry should be maintainable and other auxiliary systems and plant behavior would be expccled to be normal, At the 28-30% power level, however, the plant would still be on one feed water pump, and thus, the lineup of the feed train would be different than for normal extended operation.
Illuh power End:
On the upper and of a band of low power cxtended operations, there are some regions ahich should be avoided. At approximately 607 power, the plant exhibits ils most pronounced control instabilitics with the maximum prcssure
. oscillations within the steam generator secondary side. These pressurc
. oscillations ar e speculated by some to be contributing to secondary side tube tallure probicms vyperivnwd at some plants. Power invuls both above and below 60 show oedu<tinun In Lhese pressure uscillalions. Al ducreasiny power icvelu from 605, the moderator temperature coefficient becomes less negativo
.which is undosirable tar many plant transients.
At 40-42% pm er, the secund seed water pump can be brought on and the [ced system placed in the normal uperating mode. At about 40% power, it is likely that the main steam safeties would be challenged in the event of a turbine trip. At 28-30$ power, the saferies would probably not be challenged by such a trip. Al 45: power and above, the first and/or second bank of safeties U (out of- fiva) would be lifted for short durations. (See Attachment 1 -
Partial- Load Fluid System Operating Characteristics) s
. Decemb r 12, 1983
. P:gs Three Safetv & Transient Considerations for Operation with Reduced Power:
An evaluation was made of the time to RCS saturation and to coro uncovery for a loss of heat sink event with no HPI and with one UPI pump, with initiation at 1200 seconds for 100%, 40% and 25% power operation. The results shown in Attachment 2 indicate that there is an increase in time to core uncovery for a reduccd power condition when no heat sink and no HPl are availabic. With one HPI pump available at 30 minutes, core uncovery may occur for full power operation but not for operation at oither 40% or 25% power.
SB LOCA events of 0.01 and 0.003 ft were also evaluated with no EFW and no HPI and again an improvement in time available prior to core uncovery is found at 40% and 25% power. The timo availabic to regain HPT to provent core uncovery is correspondingly longer for lowcr powcr operation as can bu suun from the attachment.
For a turbir,e trip, the benefits of the reactor scram on turbine trip can be seen for all power levels in terms of the reduced need for steamline roller following the trip of the turbino. For a trip from 100% >ower, i three out at five banks of steamline relief valves will lift. Only one bank will lift from 40% power. The turbine bypass and ADV capacity is sufficient to provant any rollef valvo lifting for 25% power.
Based upon sensitivity studies documented in BAW - 1610, it is concluded that there will be a reduction in peak RCS pressure following a loss of feud water with failure of the RPS to trip the reactor from lower initial power icvels.
The study indicatos a 500 psi reduction in peak pressure for a 15% reduction in initial pwoor. At 40% or lower power, we would expect pressure to stay less than the recommended pressure limit of 2750 psi. Any benefit of 25%
powcr over 40% powar would probably be seen in the longer term response to u.2 ATWS rather than in the initial pressure spike.
Since only one feed water pump is in operation at 25% power and since the ICS reactor runback feature prevents a scram from 100% power, a partial loss of feed water is more severe from 25% than 100% because it results in a loss of all main Ecod water, and the need for EFW initiation. At 405 power, two food water pumps are running, but one pump is more than sufficient and thus, there is no transient on partial loss of feed water from 40% power. Soo Attachment 2 for more quantified information regarding plant transients initiated at different powcr levcis.
Qualitative Risk:
Data is not availabic to make a realistic quantitative assessment of public risk as a function of power operation in the 25-42% range. Qualitatively our judgement is that risk, defined as the frequency of occurence times consequence, is cicariv substantially reduced as one moves down from 100% power to the 40% range. At 40% power, the frequency of a transient or accident is not perceived to change (compared to full power) but the plant and accident conse-quence should be less given the lower decay heat and greater time for opera-tional response. Most of the FSAR accidents would be judged qualitatively to
Uvuumbui 12, 19H1 ,
, Pag 2 Four be of lessor consequence. Over the range of 40% down to about 28% power, the continuing reduction of decay heat le judged to further reduce risk as well
.as further increase the time for operational response to an accident condition.
Qualitatively most accident scenarios would seem to be less severe but these have not been examined in any significant depth. However, we have reviewed the dominant risk sequence from other B&W NSS plant PRA's and conclude that some of the sequences would be less ficqucnt (due to additional timo l'o r operator action) in the 25-42% range compared to 100% power. In the 28-40%
range, there is an offsetting factos in that the plant is running on the one feed water pump, and it is judged that the probability of feed system initiated trips is increased. Increased frequency of feed system upsets will lead to greater frequency of challenge to reaccor protection systems and hence provide some increase in overall risk. The relative trend of total risk due tcw these competing factors is not abic to be quantified at this time. At approximately 2H% power and below, the greater degree of manual control and "lundertuws" of thu plant is judged to OYYsul any risk reduction which might occur from lessor decay heat and, therefore, leads to no risk advantage below 28-30% power.
SUMMARY
In the 20-30% power range, a number of plant systems and control loops are operating at the end of their performance range or require manual intervention.
At the 40-42% powcr range, the ability to place the second feed water pump into operation places the plant in its normal operational lineup with all controls in automatic. In addition, operation in the 40-42% range and above allows virtually all plant operating procedures to be used without modificaliun whercas extended operation below the 30% level will require a number of tempos ary procedural ch.inytw. Operation at 40-42% is judged qualiLalIvely la c1carly be a reduction in risk as compared to 100% power and would fully meet all of the test criteria initially iden'tified. Operation in the range between 301 and 40-42% involves running with a single feed water pump but t he rest of the plant is in its normal modo, and other than increased frequency of feed watcr upsets, there is lictic basis on which to judge variability of risk.
Extended operation (greater than one ar two wocks) at below 30% power is not rccommended. We recommend further that the original test objective be maintained. Adding a 10 day or so operating period at 28-30% power would be acceptabic.
h e
GPU NUCLEAR WRFORNFION December 12, 1983 ATTACHENT 1 THREE MILE ISLAND NUCLE AR GENERATING STATION, UNIT 1 Partial Load Fluid System Operating Characteristics PARAETER OR 0 25% to 40%
ISSUE OF CONCERN LOAD CAPACIT LOAD CAPACITY
- 1. OPEPATION OF CONDENSATE SYSTEM
- a. Pump (s) operating Yes Yes at minimum or
-gre ater flow,
- b. Number of Condensate Two. Second pump Two Pumps operati na, bmught on line just prior to going above 25% load,
- c. Number of Condensate ditto above Two Booster Pumps op era t i ng . .
More polisher vessels
~
- d. Condansate Polisher. Fewer pol.isher vessels Operation. can be in' service; in service providing more effective polisher perfonnanc e.
2.
OPERATION OF f%IN F EEDWA TER SY ST EM .
- a. Pumps opera ting at Yes Yes minimum mouired (But a transient causing ( flo rec i rc . cycli ng fIow or areater, reduction in required - problems f oreseen) flow could cause recin:.
valve to cycle open &
closed.)
ATTACHWNT 1 Page 2.
PARAETER Op (a 2 5% ld40%
LOAD CAPACITY LOAD CAPACITY ISSUE OF CONCERN
- b. Feedpump turbine-drive operating at most desirable operating point.
- one (1) pump Yes Yes operation.
- two (2) pump Some potential for operation increased blade Yes erro sion. Less flexibility in speed contml rang e.
- c. Tra nsf er from Transf er occurs at Transf er to main f eed-st art-up f eedwater 22-23% power. Feed water regulati oq val ve regulatina valve to system control completed well oetore main feedwater requires manual 40% load.
reaulatim v lve. attentjon,
- d. Number of main One (1) (Loss of Two (2) (Loss of one f eedwater pumps pump causes loss pump is compensated oy operati ng per of feedwater) other pump picki ng up nomal opera ting loaa.)
omcedu re s.
PERFOR fM NCE .
- a. Onc e-th m uch OTSG in transition OT SG i s i n o nc e-th m ugh v s. recirculation fmm recin ulation mode; no impurity t o o nc e th m uqh i n c onc en t ra t ion, mode.
range of 0 - 20%
power; impurities will tend to concentrate in OTSG.
- b. OTSG nutlet steam In superheat region S I ight ly 'gre ater quality, degree of steam superheat than at 25%.
j.
f* .-'
ATTACHtfNT 1
Page 3.
PARAETER OR @ 25% 0 40%
ISSUE OF CONCERN LOAD CAPACITY LOAD CAPACITY
- 4. l%IN TI.RSINE-G ENERA TO R -
TERTDMiCE
- a. full auto control Yes Yes (Transfer to Auto 9 15% power)
- b. Light load operation Increased probability No concerns expressed by affects on last stage of blade buffeting or turbine manufacturer, blada loadina . flutter.
- 5. EXTRACTION-STE AM FEE 0 WATER HEATING SYSTEM PERFORMANCE
- a. All extraction Yes Yes
. steam block valves open
- b. Need for eiohth Transition from No aux. steam necessary;
- stage feedwater aux, steam to extraction steam flow he a t i ng' f ro m a ux i l- extraction steam is suf ficient to ensure iary steam vs. is estimated to proper feedwater heating turbine extraction occur at ~ approx, in low pressure f eed-steam. 20% power, water heaters.
- 6. FEECWATE4 HEATER OPAIN SYSTEM Eighth staae Transition to Transition to nomal drains nomal 8th stage feedwater heater j
dra ining normally at about 20% load, heater drains well before iO% power.
'(i.e. aux, steam isolated & acr.
steam drains no lamer baina returned to aux.
boiler).
\
F
_ ,~ , . . . - -- - - , , < _ - - - . _ - _ - . _
ATTACHMENT 2 GPUN December 14 1983 EFFECT OF POWER LEVEL ON PLANT BEHAVIOR Page F.
DUR]NG SELECTED TRANSIENTS AND ACCIDENTS ,
Analysis Method Or Event Reference Parameter 100% Power 40% Power 25% Power Conunents Loss of Heat Sink CSMP & l. SG dryout time 110 sec 130 sec 130 sec RCS saturation results in rapid insurge with no HPl RETRAN 2. PORV actuation time 130 sec 160 sec 160 sec into pressurizer with subsequent opes.ing
SBLOCA (0.01 ft2 ) CSMP 1. Minimum pressure prior 1550 psia 1430 psia 1400 psia RC pump heat not included. Pump trip on eith no EFW & no to repressurization loss of SC margin would occur from 500 to HPI 2. Core uncovery time 2400 sec 8700 sec 12000 sec 700 seconds for all thr e power levels. HPl
- 3. Time req'd for HPI Core uncovery 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> initiation implies one HPI pump only, initiation to prevent dependent upon Boiler-condenser needed from full poter to core uncovery EFW initiation depressurize system and get higher HPI flow.
(estimated) EFW not needed from 43% or 251 power because
Initiation to prevent core uncovery (estimated)
SBLOCA (0.004 f 2t ) CSMP 1. Minimum pressure prior 1600 psia 1550 psia 1500 psia RC pump heat not included. Pump trip on uith no EFW , no to repressurization loss of SC margin wuld occur from 600 to HPI 2. Core uncovery time 3000 sec 8000 sec 15000 sec 1200 seconds for all three power levels. HPI
- 3. Time req'd for HPI Core uncovery 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> initiation implies one HPI pump only 40%
initiation to prevent dependent upon case repressurires to PORV sgtpoint and core core uncovery EFW initiation uncovers faster than 0.01 ft break which (estimated) does not repressurire as much. Thus, core
- 4. Time req'd for EFW 30 minutes EFW not req'd EFW not reg'd uncover initiation to prevent 0.01ftgfasterforthiscasethanfor case. B-C only needed from full core uncovery power with one HPI pump. Not need(d from (estimated) 40% or 25% power.
i GPUN DeumiSce- t/. 19M Analysis '
Method Or Pane 2 Ivent Reference . Parameter 1001 i'ower 40% Power 75% Power' foernien t s -
loss of Heat 5 ink CSMP & l. SG dryout time 118 sec 130 130 sec -
HPI initiation asstuned to occur at 1200 with One IIPI Pump RETRAN 2. PORV actuation t ine 130 sec l ' ' sec. 160 sec sec seconds. RC5 saturation results in lar rte (Iced & Bleed) . 3. Core uncovery hone None Hone insurge into pressurlier and subsequent lift
- 4. Tine to saturation 850 sec No saturation fio saturation on both pressurtier safety values. RC pumps
- 5. Time to pressurlier 850 sec No 5/V lift No S/V lift off at time zero to simulate a loss of off-5/V lif t with PORY open site power. For 100% power, subcnoling is
- 6. Time to regain sub- 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> Never lost Never lost not required.
Cooling loss of feedwater BAW -1610 & 1. Peak RCS pressure 3464 psig 2750 psig 2750 psig Based upon BAW -1610 and sensitivity study ATWS BAW -10099 2. Moderator temp co- -1.05 x 10-4 -0.841 X 10-4 -0.731 x 10-4 which shows that:
& GPU Cal (s efficient 10% MIC = 120 psi (P2-21 of BAW'-1610) 15% power = 500 psi (Figure 5-1 of BAW -10099)
Turbine Trip RETRAN 1. Steamline relief Bar.ks 1-3 Bank I only None full power turbine trip analysed with a (Assume reactor valves lif t ing reconnended set of ICS settings. With scram on turbine 2. PORV actuat ion time No lift No lift No lift proper secondary side response, RCS pressure trip) 3. Peak RCS pressure 2230 psia 2230 psia
- 2230 psta* reaches pressortier spray setpoint and then turns around. Only 100% power case was analyzed with REIRAN. Other power level results estimated from other analyses. 40%
secondary safet) valve behavior based upon a loss of feedwater analysis. Two small safeties will also lift for 100% and 40% but not for 25% power.
Trip of a Single RETRAN & l. Peak RCS pressure 2235 psia" No change 2235 psia" Two feedwater pumps operaticqal at 40%
Main Feedwater Pump ffiS for 2. Steamline safety None Nnne None power. One feedwater punp capable of encren (Partial tufW) 100% power valve lift removal up to 60% power. Thus, trip of ora case 3. Integrated control 60% No runback Plant scrams pump at 40% pnwer does not result in system runhatt power required transient behavior. Power is below f.0% sun level back point also. Since single feed pump
- 4. Scram No scram No scram Scram on loss running at 25% power then trip of one pump of feed water is a loss of all MfW and is a nore severe transient which results in scram, init lat irai of f.FW and HLS pressurization.
- Not Analyzed
- Pressuriier Spray Setpoint O