ML20053D194
| ML20053D194 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 06/01/1982 |
| From: | Hukill H GENERAL PUBLIC UTILITIES CORP. |
| To: | Eisenhut D Office of Nuclear Reactor Regulation |
| References | |
| REF-GTECI-A-49, REF-GTECI-RV, TASK-A-49, TASK-OR 5211-82-132, NUDOCS 8206040172 | |
| Download: ML20053D194 (19) | |
Text
.
GPU Nuclear IJ ?
8 'g g gp P.O. Box 480 Midcietown, Pennsylvania 17057 717-944-7621 Writer's Direct Dici Number:
June 1, 1982 5211-82-132 Office of Nuclear Reactor Regulation Attn: Darrell G. Eisenhut Director of Division of Licensing U. S. Nuclear Regulatory Commission Washington, D.C.
20555
Dear Sir:
Three Mile Island Nuclear Station, Unit 1 (TMI-1)
Operating License No. DPR-50 Docket No. 50-289 Presaurized Thermal Shock (PTS) to Reactor Pressure Vessel This latter is in response to your August 21 and December 18, 1981 letters requesting information concerning the PTS issue.
As we indicated to you in our March 17, 1982 correspondence, we have proceeded with our analysis in-house and with B&W by using a realistic mixing model as well as TMI-1 plant specific data. We have placed a high priority on complet-ing the analysis so that we could report to you in June.
Since you need this report to brief the Commission during the early part of June, we are, with this letter, summarizing the results of our analysis. We feel that this information, much of which is excerpted from what will be the final report, will provide you with sufficient information to formulate recommendations.
We will issue the final report, currently under detailed review, the last week of June.
The analysis is very encouraging in that it concludes the following:
Reactor Vessel integrity will not be compromised due to low to moderate frequency events and anticipated transients during the designed lifetime of the vessel.
The analysis includes significant conservatisms that add to the margin of safety in maintaining RV integrity.
The current rate of embrittlement of the TMI-l vessel may be reduced further if the plant switches to a low leakage fuel scheme in near term reloads.
8206040172 820601 c)
PDR ADOCK 05000289
/,
P PDR l
GPU Nuclear is a part of the General Public Utikties System 1
Mr. Darrell G. Eisenhut 5211-82-132 Because~of the concerns raised by the PTS Issues, operator responae will be significantly improved through increased awareness and additional training.
We hope that this letter and attachment will assure the Staff that there are no short term or long term concerns in regard to PTS. The final report the Staff will receive at the end of June will elaborate on these conclusions in greater detail. Likewise, GPU Nuclear's ongoing participation in Owner's Group Materiah Surveillance Program will provide the periodic confirmation of the integrity of the reactor vessel.
Sincerely,
. D.
kill Director, TMI-1 HDH:PGD:vj f Attachment f
e i
GPU Nuclear - TMI Unit 1 NRC letter of August 21, 1981 - REQUEST FOR ADDITIONAL INFORMATION Question 1.
Geometry Geometrical description including design and as-built (when available) dimensions of the core, assemblies, shroud / baffle, thermal shield,
downcomer, vessel, cavity, and surrounding shield and/ or support structure.
Response
The following references contain the requested geometrical descriptions for THI-1:
o TMI-1 FSAR - Chapters 1, 3 & 4 o BAW 1628 "RV Brittle Fracture Analysis During SBLOCA Events With Extended Loss of Feedwater" o BAW 1646
" Thermal Mechanical Report - Ef fect of HPI on Vessel Integrity f or SBLOCA With Extended Loss of Feedwater" Question 2.
Material Description Region-wise material composition and material isotopic number densities (atoms / barn-cm) for the core, near-core regions and RPV, suitable for neutron transport calculations.
Response
The following references contain the available material descriptions:
o BAW 1511P
" Irradiation-Induced Reduction in Charpy USE of RV Welds" o BAW 1439
" Analysis of Capsule TMI-lE From Met-Ed Co.
TMI-l - RV Materials Surveillance Program"
=_._ _.
Question 3.
Neutron Source e
f Present and expected EOL:
a) Assembly-wise and core power history (EFPY) b) Rod-wise and core power history (EFPY) for peripheral assemblies,
c) Core average axial power history distribution.
I i
Response
The following references contain the requested neutron source inf ormation:
BAW 1511P* " Irradiation-Induced Reduction in Charpy USE of RV Welds" o
BAW 1485P* "Pressur'c Veisel Fluende Analyris o
for 177 FA' Reactors" 1
Question 4
~4.
Vessel Fluence I
a) \\ Description of available calculations of the vessel fluence including fluence values, locations, and corresponding powee histories (EFPY), including 1/4T, I 1/2T and 3/4T thrcugh the RPV.
l j
)
s l
!r ) - Description of available capsule-inferred vessel fluences including fluence values, locations,iand s
sdcorresponding power histories (EFPY).
~
\\
- Response:
4 The following references contain the requested vessel fluence descriptions:
o 3;# 1511P*,
"Irrediation-Induced Reduction in Charpy USE of RV Welds" i
'BAV 1485P*
" Pre.ssure Vessel Fluence Analysis for 177 FA Reactors" o
i i
Ouestion 3
5.
Surveillance Capsules a) Capsule materials, tradial and aidal dimensions.and i
+
?*\\'
locations.
N.
3,
,s s
s
(
s
' AQ' c
\\
g
- B AW 14 85 P and B AW.1311P do,not ref1"ect new f uel shuf fle.; scheme (i.e.,
18 month LBP Low Leakage Cycle.)!
j 3
.7 1
A f
I g
't t-g
/
}.
g
- ^
{ [
s
.4 I
. _.. ~ -.
,, x
b) Capsule fluence measurements, together with the accum-ulated power history (EFPY) and a description of the lead factors used to extrapolate the measurements" to the peak wall fluence location.
I R7sponse:
The following references contain the reque s t ed information en surveillance capsules:
BAW 1439
" Analysis, of Capsule TM1-lE from Met-Ed Co. TMI RV o
Materials Surveillance Program" BAW 10006, Rev.3 "RV Material Surveillance Program" o
BAW 1543
" Integrated Reactor Vessel Material o
Surveillance Program" Question 6.
Vessel Welds Axial and aximuthal locations of vessel weld-seams with respect to the Overlay of current fluencemap with weld locations.
Identify the core.
critical welds, vertical and circumf erential, and give the weld wire heat numbers. Give weld chemistry for the critical welds.
For each weld wire haat number, report the estimated mean copper content, the range and the standard deviation, based on all the reported measurements for that weld wire heat.
The welds may be surveillance veldments for your vessel or others, nozzle dropouts that contain a weld, weld metal qualification data, or archive material.
In the ab sence of any inf ormation, assume that cooper content is at its upper limit (0.35 percent when using R.G.
1.99, Rev. 1) and that the nickel content is high.
R2sponse The following references contain the requested vessel weld data:
BAW 1511P
" Irradiation-Induced Reduction in Charpy USE of RV Welds" o
BAW 1439
" Analysis of Capsule TMI-lE from Met-Ed Co."
o MW 10006, Lv.3 "RV Material Surveillance Program" o
BAW 1500P
" Chemistry of 177 FA B&W Owner's Group RV Beltline Welds" o
BAW 1485P
" Pressure Vessel Fluence Analysis for 177 FA Reactors" o
Question 7.
Systems Analysis c)
Provide a list of transients or accidents by class (for example:
cxcessive feedwater, operating transients which result from multiple failures including control system f ailures and/or operator error, steam line break and small break IDCA) which could lead to inside vessel fluid temperatures of 300'F or lower.
Provide any Failure Modes an! Ef fects Analyses (fMEAs) of control systems currently available or reference any such analyses already submitted.
Provide the analysis of the most limiting transient or accident with regard to vessel thermal shock
and considerations. Estimate the frequency of occurrence of this event provide the basis f or this estimate.
Discuss the assumptions made regarding reactor operator actions.
b)
Identify the computer programs used to calculate the limiting or accident.
Indicate the degree to which the computer transient programs used have been verified and any other additional verification requested to demons.trate th a t the computer program models adequately treat the identified important physical models (i.e., ECC mixing, heat transfer, and repressurization).
Response
There are a vast number of possible overcooling event sequences that can be postulated to occur.
Design basis events, such as loss of coolant accident (LOCA) and main steamline break (MSLB) have already been addressed previously during the original licensing process.
Thu s, we logical approach to identify those overcooling transients have chosen a to analyze in detail - that is, we chose event sequences that provide an
.vi severity of appropriate balance between likelihood of occurrence cooldown.
Results of various independent programs have contributed to the selection processs and have resulted in a significant reduction in the number of overcooling and SBLOCA events that require detailed analysis, as well as a better understanding of system response once an event is postulated to occur.
Therefore, in response to Question M - Systems Analysis - we have chosen the following two classes of events that yield conservative but realistic scenarios that require evaluation of the vessel in terms of thenmal 1
shock:
Overcooling Transient Overcooling of the plant occurs during events which result in increased steam flow with subsequent secondary side depressurization, reactor trip and continued feedwater injection.
Cvents of this type which result from pipe breaks are less likely to occur than those which occur from equipment malfunctions.
In this regard, a single failure in the Integrated Pressure Control portion of Integrated Control System (ICS) results.in a full open signal to the the Turbine Bypass Valves. Severity of a cooldown transient is judged on its initial cooldown rate, the minimum temperature, and the eventual scenario. Events repressurization of the RCS if this also occurs in the with a fast cooldown rate, low minimum temperature, and subsequent repressurization are considered most challenging to the reactor vessel's integrity.
A comparison of the turbine bypass f ailure event with other potential cooldown events is provided in Table 1.
This table is based upon previous analyses, engineering judgment and experience and shows that failures which have a more severe overcooling potential are less likely to occur.
Similarly, more likely failures have a smaller was chosen potential for overcooling.
The Turbine Bypass Valve Failure as the overcooling transient to be evaluated for reactor vessel thermal shock on the basis th a t it has a relatively large likelihood of
TABLE 1 J
COMPARISON OF OTHER OVERC00f;ING EVENTS WITH THE BASE CASE TURBINE BYPASS VALVE FAILURE EVENT DESCRIPTION RELATIVE RELATIVE LIKELIHOOD FAILURE SEVERITY OF OCCURRENCE COMMENT Emergency Feedwater Less Comparable This type of event was analyzed in section 8.0 Overfill of the TMI-l Restart J
Report Stuck Open Secondary Less
- Comparable TMI-l Restart Report Section 8.0 Safety Valve or Atmospheric i
Dump Valves Stuck Open Turbine Stop More Less*
More severe overcooling initially, but more Valve likely to be terminated i
quickly by operator action i.e., the first two post-trip actions are to verify reactor and turbine trip.There-fore, overall severity judged to be less.
Small Steamline Breaks Comparable Less Large Steamline Breaks More Much Less Analyzed in TMI-2 FSAE Appendix 15D (Analysis of Large Steam Line Breaks),which bounds TMI-1 Total Loss of Feedwater More Less Analyzed in this reper:
See Section for Small With HPI Cooling Break LOCA.
SMUD Event of 3-20-78 More Much Less NNI/ICS Power Supply Modifications Reduce
,l Likelihood of this event.
- The turbine control and intercept valves would also have to f ail open, which is considered less likely than a failure of the turbine bypass valves.
i
occurrence as well as degree of severity.
SBLOCA Transient The criteria used f or selecting the SLOCA transient are as follows:
1.
The transient must be realistic.
It may be a result of a normal operating transient and will evolve in a mechanistic manner.
2.
The transient must have a reasonable probability of occuring.
Systems must be challenged to fail and may contain single or multiple f ailures, i.e.,
an initiating event plus one or more failures.
l 3.
The transient provides a realistic challenge to the reactor I
vessel.
The transient and resulting f ailures provide a thermal shock to the reactor vessel.
~
The initiating event (i.e., 1oss of all feedwater) for the SBLOCA provide a mechanistic manner in which to transient was chosen to challenge the system and provide the necessary criteria for obtaining :
the rm al shock of the vessel, i.e., by providing a once-through mode of cooling.
The mode of cooling is the cold HPI flow which is injected into the cold leg of the plant, passing through the core, and out of the system through a break.
Therefore, the initiating event that will challenge the code safety in a mechanistic manner is a loss of all feedwat er.
In order to allow the code safety valve to be challenged during the transient, the PORV is assumed to be isolated.
The mechanistic break size chosen using these criteria was a single TMI-1 pressurizer code safety valve (0.0182ft ).
The PORV break size was considered in BAW-1648 since it led to repressurization of the RCS.
- However, the transient provides a less severe the rmal shock than the large safety valve break because the RCS temperature remains considerably warmer th an with the 0.018-f t break, and the HPI flow is less than with 1
1 the 0.018-ft break.
Break sizes larger than 0.018-ft 2 result in more rapid depressurization to pressures at which the low-pressure inj ec t ion (LPI) system provides makeup (with little or no repressurication).
The transient response of the s e larger breaks is similar to th a t of the large break LOCA being considered under NRC Task Action Plan A-11.
An examination of the system response for various break locations shows that the limiting condition will occur for a break in the p essurizer or hot leg independent of break size.
This results from the following considerations:
For cold leg breaks, the hot water leaving the core flows (1) through the hot leg, steam generator, and broken cold leg to the break; and (2)
The through the vent valve, downcomer, and broken cold leg to the break.
latter path has the least flow resistance and thus allows a large portion of the hot water to enter the downcomer for mixing.
Furthermore, the
diversion of HPI water to the break reduces the total amount of HPI water entering the downcomer.
For hot leg or, pressurizer breaks, more HPI water is available to enter the downcomer. In addition, less vent valve flow occurs in a hot leg or pressurizer break, thus decreasing the amount of hot water available for downcomer mixing.
The analyses that follow used a pressurizer break to evaluate system conditions.
It is further concluded that breaks within the hot leg will respond in the long term in'a f ashion similar to those in the pressurizer.
- Thus, the combination of system response, break location, and break s ise d e t e rmin ed t h a t-the ev al u at ed accident would be the failure of a preseurizer code safety valve.
The SBLOCA transient thus anarlyzed, is a pressurizer code safety valve (0.018-f t2) failing open af ter being challenged by high pressure as a
l result of a loss of all f eedwater.
The valve cannot be isolated and thus continues the blowdown until the RCS can be placed on the decay heat remov al ( DHR) sys tem.
Since feedwater is not restored during the event, both forced and natural circul,ation RCS flow are assumed to be lost for the duration of the event.
Response to Question 7b ANALYTICAL MODEL DESCRIPTIONS Overcooling Transient The Turbine Bypass Valve f ailure overcooling transient was analyzed by using the system analysis code RETRAN and the dynamic simulation code CSMP (Continuous System Modeling Program).
For the first twelve minutes, during which a best es ticat e of primary and secondary parameters is the one-dimensional multi-node RETRAN program provides the necessary, level of detail required to model the event.
The overcooling is essentially terminated by isolation of the steam leak and termination of the HPI; therefore, the long term system response is an energy balance calculation which can be ef ficiently perf ormed by using the dynamic simulation 1anguage CSMP.
Both models are discussed below.
3 TheTMI-1ftETRANModel I
RETRAN-01;is the first released version of the computer code package developed:by EPRI for analysis of light water reactor operational transients.
It is based on the REIAP series of codes using homogeneous equilibrium flow equations and has been through extensive verification and benchmark against separate ef fects tests and plant transient data j
conducted by GPUNC on events which have occurred at both Dil-1 and TMI-2 f or which actual recorded data was available.
t l
The CSMP Model for Long Term Cooling In order to " terminate" the overcooling event within the first few hours af ter initiation, a stabilized RC temperature mus t be maintained.
The plant is controlled by using either the steam generators as a heat sink or by the primary side feed and bleed mode of operation.
The transient t+
is slow-varying with time and there are no sudden variations expected between the various locations in both the primary and the secondary systems. Therefore, the nodal diagram can te simplifiedto two volumes -
o ne f or the RCS and one for the secondry side.
Instead of using RETRAN, the mass and energy balance equations can be solved very efficiently using the dynamic simulation language CSMP (Contnuous System Modeling Program) for an extended transient time.
SBLOCA Transient The CRAFT Model An eight-node CRAFT model was developed to determine the thermal-hydraulic conditions existing in the RV for the SBlhCA analysis.
The system noding f or the eight-region model is described below:
Node 1:
Reactor vessel downcomer and lower plenum Node 2:
Reactor core and upper plenum Node 3:
Cold legs between RC pumps and reactor vessel Node 4:
Hot legs Node 5:
Primary side of steam generators and cold legs between steam generators and RC pumps Hode 6:
Secondary side of steam generators Node 7:
Pressurizer Node 8:
Containment Node CFT:
Core flood tank The CRAFT code assumes homogeneous mixing of the liquids in a node and determines its the rmod ynamic conditions based on thermal equilibrium between the steam and liquid phases.
This assumption will result in complete mixing of the cold and hot fluids entering the downcomer region from the cold legs and vent valves, so the downcomer node temperatures calculated by CRAFT are mixed mean temperatures.
However, this temperature is not used in the mixing or fracture mechanics analysis.
The primary purpose of the LOCA analyses was to determine the HPI flow rate, vent valve flow rate and temperature, and RCS pre s sure for use in the fluid mixing analysis.
The SBlDCA analyses used the CRAFT code only during the blowdown stage cf the transient until the core outlet temperature became 100'F subcooled.
After the RCS refilled with water, a steady-state analysis was performed to d e t e rm ine the RV conditions.
This steady-state analytical method was benchmarked in BAW-lo48 against the CRAFT analysis for the 0.007-ft2 pressurizer break with no operator action.
The results indicate that the steady-state code predicts the downcomer temperature approximately 9%
above the CRAFT prediction.
The 9% deviation in the downcomer temperature was used as an adjustment factor for the breaks analyzed in BAW-1648 as well as the present analysis.
The COMMIX-1A Model The COMMIX-1 A (advanced version of COMMIX-1) is a three-dimensional, transient, single phase computer code for thermal hydraulic analysis.
It i
i l
i l
i uses the ICE technique of Harlow and Amsden to discretize the conservation equation of mass, momentum and energy.
The set of discretization equations are then solved us'ing the cell by cell (point by point) iterative procedure.
To permit the analysis of a flow domain with solid objects, the porous medium formulation with surf ace pe rmeab ility,
volume porosity, distributed resistance and distributed heat source are incorporated in the conservation equations. The " Force Structure" model and the " Thermal Structure" model in COMMIX-1A permit calculations of distributed resis'tances and distributed heat sources, respectively.
The.se models are designed such that we can use different correlations at different grid loations.
In addition, the code has various options permitting a large amount of flexibiility (e.g., use of Cartesian or cylindrical coordinate systems, various rebalancing schemes for speedy convergence, au t om a t ic time step selection, implicit energy option for low flow cases, etc.).
The code has been developed and re fined over a number of years, and j
already a large number of computations for complex situations have been performed.
The following is a list of some of the problems analyzed using COMMIX-1A.
I A.
Pretest Prediction of the W-1 SLSF Experiment i
B.
Hexagonal Fuel Assembly with a Planar Blockage C.
Nineteen-Pin LMFBR Fuel Assembly in A Hexagonal Duct with Power Skew D.
Flow Stratification in a Horizontal Pipe E.
Simulation of LMFBR Outlet Plenum Mixing F.
Analysis of LOPI transient G.
Simulation of P2 Transient Free Convention Test H.
CRBR Upper Plenum Under Thermal Stratification I.
German 7-Pin Flow Rundown Test J.
Solar Pond Heat Loss
- Although the COMMIX-1A has been develped for thermal hydraulic analysu of LMFBR fuel assemblies, the code is designed to permit applications to other components and other reactor types,
e.g.,
PWR.
This can be seen from some of the applications listed above.
In addition the COMMIX-1A code has been benchmarked and produced reasonable agreement with (a) experimental mixing tests conducted by B&W Alliance Research Center in Alliance Ohio, and (b) test data obtained from the one fif th scale model CREARE test facility that approximated the B&W 177 FA geometry.
NRC Letter of December 18, 1981 - Enclosure (2)
AMPLIFICATION OF THE "150-DAY" REQUEST TO THE AUGUST 21, 1981 IETTER Question (1)
Identification of the PTS events that were considered in reaching your conicusions, and a justification for PTS events th a t you did not consider.
You should include a quantitative assessment of the probability of occurrence of the various PTS events considered and not J
m--
p r-
considered and an accompanying as sessment of the likelihood of vessel failure vs. EFPY for the events.
The man,ner in which you considered multiple failures of systems, components, and those resulting from operator actions should be described in detail.
Response
of our response is contained in the response to Question #7 of the Most 8/21/81 NRC letter. '
Additionally, although GPU Nuclear has not embarked on a Probabilistic Risk Assessment (FRA) program at this time, work conducted by Duke Power Co. for Oconee III and reported to the NRC on January 15, 1982 in failure to response to the PTS issue, has shown probabilities of vessel be extremely low based on the relative threat of severe overcooling events.
GPU also believes, th at as further refinements are made to the models used in the PTS analyses, the. likelihood of events that may lead to vessel f ailure will further diminish.
Question (2)
A description of the steps, if any, you are taking now or plan to take in the near future to delay the rate of further embrittlement of your vessel, and your assessment of the effectiveness of those steps.
Response
Currently TMI-1 employs a 12 month " Modified Out-In" refueling pattern.
For economic reasons, GPU Nuclear is evaluating the trans.ition to 18 month re fueling cycles in future operations.
Preliminary estimates indicate th at the Lumped Burnable Poison (LBP) "In-Out-In" reloading,
t that can be u s ed in the longer cycles, would reduce the vessel fluence rate by 25%.
I Question (3) Your assessment of the sensitivity of your analyses to uncertainties in input values, such as initial crack size, copper content, fluence, and initial reference temperatu e at welds.
Response
of the long range program on RV Integrity, the B&W Owner's Group As part is currently investigating quantitatively, the sensitivity to the various parameters used in the PTS analysis.
Results of this analysis will be communicated to the Staf f as part of the Owner's Group continued ef fort to keep the Staff appraised of significant phases in the materials program.
Question (4) A list of assumptions relied upon in reaching your conclusions.
l a.
If this list includes " credit" for operator actions,
f describe the basic instructions given the operators (for example, if a "sub-cooling" band is used, des-cribe it).
Sinbait the procedures the operator will follow, and describe the training being given to establish operator readiness to cope with PTS events.
b.
If the list includes credit for the ef fects of warm prestressing f or some event sequences, include your justification and analyses showing that such events will follow a pressure-temperature pathway f or which warm pre-stress is ef fective.
R2sponse:
Overcooling Transient Assumptions The transient is initiated by a secondary side upset (i.e., all six turbine bypass valves suddenly open) The blowdown is further increased by the assumption that the Turbine Control Valve remains full open.
The initiating event and control valve failure result in a maximum blowdown of 127% of normal steam flow.
This excessive heat removal causes the PCS tempe ra ture and pressure to decrease rapidly.
Owing to the negative tem pe r a t u re coef ficient of the core, excessive reactivity is introduced cnd the reactor is tripped on high neutron flux.
The following is a complete list of the analysis assumptions.
Initiating Event -
Simultaneous opening of all Six Turbine Bypass Valves.
Rsactor Trip -
105% neutron flux.
HPI initiation -
1600 psig primary pressure (ESAS)
Internal Vent Valves - 0 Core Flood Initiation - 600 psig primary pressure Core Flood Temperature-100*F BWST/HPI Temperature - 40*F Core Decay Heat -
1.0 ANS Break Flow -
(a) 0.523 ftf Turbine Bypass Area maximum flow = 27% of full power steam flow.
(b) Turbine control valve remains full open.
RC Pump trip -
ESAS plus 30 sec.
EFW S ys t em -
Maximum flow to OTSG 1evel setpoint SBLOCA ANALYSIS ASSUMPTONS Consideration of various requirements f or system response and its effect on thermal shock, as outlined, resulted in the selection of a transient that unfolds as a result of a pressurizer safety valve f ailing open with a loss of all feedwater.
This results in the core being cooled in the onc e-through HPI cooling mode of operation.
The cold HPI water is injected in the cold leg, enters the downcomer, and could result in large thermal gradients in the vessel wall.
The TMI-l plant specific SBIDCA anlysis assumptions are as follows:
CRAFI Model
- 8 Node Model used A
Break Size
.018ft
, one TMI-1 pressurizer code safety valve Location
- Top of pressurizer, code safety valve Initiating Event
- Loss of all Feedwater at time zero Reactor Trip
- Reactor TRIP on high RCS Pressure, 2355 Psig RC Pump
- TRIP RC pumps immediately on loss of subcooling margin or on ESFAS initiation, whichever occurs first HPI Initiation
- Initiate HPI on ESFAS, 1600 psig HPI System
- TMI-I plant specific - 2 HPI pumps with venturis BWST/HPI Temperature
- 50*F Internal Vent Valves
-8 Core Power Level
- 2568 Mwt Decay Heat Assumption
- 400 EFPD with 1.0 ANSI curve Includes all structural metal heat Structural Metal Heat Break FLOW Model
- Subcooled, Bernoulli with C
.7
=
3
- Saturated or steam, Moody Correlation with CD = 1.0 OPERATOR ACTIONS In bo th the overcooling and SB LOCA transients, the analyses assumed the operator throttled HPI to prevent the system from becoming more th an 100*F subcooled.
The operator controls the flow to maintain the system between 50*F and 100*F subcooled.
These ac tions are consistent with the plant HPI throttling criteria that requires the operator to bypass the ESAS signal and throttle HPI only if one or more of the following criteria are met:
1.
HPI must be throttled to prevent violation of the applicable brittle f racture curve limitations.
/
2.
HPI may be throttled if LPI flow is greater than 1000 gpm in each t
line and stable for 20 minutes.
3.
HPI may be throttled if the required subcooling margin (50*F except for OTSG tube rupture, then 20*F above 1600 psig, 50*F below 1600 psig) exists and pressurizer level is established 100".
NOTE:The margin to saturation is determined by the saturation margin meter and/or the average of the five highest oper-able incore. thermocouples.
The above required actions are independent of the event and are posted in the control room.
In addition to throttling HPI, the analysis of the overcooling transient assumed the operator isolated tlie steam leak at twelve minutes into the event. We believe this is very conservative since we would expect the operators to identify an overcooling event within two (2) minutes using the Pressure-Temperature plot. Plant procedure EP-1203-24 requires the l
operator to isolate the steam leak as soon as it is identified.
The first action the operator finally takes is to isolate the Turbine Bypass Valves.
OPERATOR TRAINING The operators continually review the throttling criteria of HPI as well as the basis for the criteria. Included training activities include the studying of the procedures
- in great detail in preparation for NRC examina t ion s.
As an example, one of'the classes on mitigating core damage entitled "Potentially Damaging Conditions" deals with the HP1 throttling criteria and basis supplemented by the use of the P/T (Pressure / Temperature) plot trainer.
Likewise, the operators receivedinstruction at the B&W simulator on the basis of the Heatup/Cooldown curves where they also receive instruction or OTSG overfill and Non-LOCA overcooling events.
Addtionally, the
- operators are trained on the simulator using procedures to respond to various IDCA transients.
During 1982, additional training will be conducted on PTS.
In fact, our management is currently evaluating detailed lesson material under preparation by B&W, for an operator training program that familiarizes the operators specifically with reactor vessel thermal shock and the concerns, should a severe transient occur.
Implementation of the course material in the training program is scheduled for this sumer and will generally follow this outline:
- The following procedures contain HPI throttling criteria:
EP-1202-4, EP-1202-5, EP-1202-6B, EP-1202-6C, EP-1202-39, EP-1202-6B, EP-1202-2A EP-1202-2, EP-1202-14, EP-1202-26A, EP-1202-26B, EP-1202-36A, EP-1202-36B, OP-1102-16, OP-1105-3 i
l l.
o Reactor Vessel Thermal Shock (T/S) Description Purpose of T/S Lessons Discussion of T/S Scenarios Description of Brittle Fracture Failure Scenarios as it applies to Reactor Vessels Factors that Increase the Possibility of Brittle Fracture Failure o
of Reactor Ves,sels Severe Cooling of Vessel Inner Surface due to transients Irradiation Damage to Vessel Material Assumed crack geometry Operator actions Mixing in Cold Leg & Dokncomer o Effects of these Factors and Why They Are of Concern to Thermal Shock Reasons for Operator Actions Assumed in T/S Analysis o
Actions for SBLOCA Actions for Overcooling Transients Actions for LOCA Recognizing Symptoms and the Use of the Symptom Oriented Procedure o
That Will Minimize the Occurrence of Pressurized Thermal Shock o Recognize RC System Response o Use of Symptom Oriented Procedure Operator Actions in Symptom Oriented Procedures That Will Minimize o
the Occurrence of Pressurized Thermal Shock Warm Prestressing Finally, although we have not taken credit for varm prestressing, we feel that warm prestressing is e demonstrable and valid physical phenomena that has been experimentally preven.
We feel that warm prectressing is a phenomenom that adds additional margin to the service life of the vessel and can be considered when certain highly unlikely but challenging transients are po s tulated.
NRC Letter of December 18, 1981 - Enclosure (1) - AMPLIFICATION OF THE "150-DAY" REQUEST 10 THE AUGUST 21, 1981 LETTER Question
- 1. RTnd t
Response
1
The Staff has accepted our values pracsntsd in rasponse to the "60 Day" letter.
Question 2.
Rate of Increase of ETndt You have provided rates of increase in fluence per EFPY for your reactor vessel. We accept these values.
However, please provide the rate of increase of fluence at ti.s ritical longitudinal weld location taking into consideration any contempitted changes in core configuration.
Response
As we indicated in our response to Question (2) of Enclosure (2) of the NRC Letter of December 18, 1981, we are studying the implementation of an 18 month Lumped Burnable Poison "In-Out-In" fuel scheme.
If implemented the 18 month LBP scheme would have the additional benefit of a reduction in fluence of approximately 25%.
This is applicable to all critical longitudinal and circumferential welds.
Question 3.6 4. RTndt Limit and Basis for the Limit Since the "60 day" re sponse stated that you do not consider e limit on RTndt to be an appropriate basis for continued opera tion that, if implemented, would assure maintenance of a acceptable low risk of vessel failure from PTS event for the near-term, pending longer term results of more detailed analysis or research. We will be developing this criterion considering recommendations that you may provide in your "150 d ay" response.
Response
GPU Nuclear feels that an appropriate criterion to assure an acceptable low-risk of vessel failure should derive from the work currently underway under the B&W Owner's Group Reactor Vessel Integrity Program.
This periodic evaluation will allow for advance warning of any potential problems while accomodating c.dvances in technolgy.
These problems could be accomodated within the existing framework of Appendix G by allowing an alternate means of determining material toughness through emerging Elastic Plastic Fracture Mechanics techniques.
Question 5.
Operator Actions We are aware through the TMI-1 restart hearing proceedings of the emphasis placed on the overall concern of PTS at TMI-1.
We are aware that this issue is addressed in procedures, training and management involvement and th a t operators are sensitive to the thermal shock considerations.
However, we cannot determine from your "60 day" response the degree of emphasis which is currently placed on the issue in training I
t I
I I
1
cnd man: gem:nt involven;nt.
We request that you expand your response to provide us a more detailed discussion of what steps have been taken 'to ensure that your operators have a firm grasp of this issue and can be expected to cope with the events which serve to initiate PTS.
Response
Please refer to the response to Question (4) of the NRC Letter of December 18, 1981 Enclosure (2).
NRC Letter of August 21, 1982, p.3 Question You are also requested to submit a plan for Three Mile Island, Unit No. I to the h1C within 150 days of the date of this letter that will define actions and schedules for resolution of this issue and analyses supporting continued operation.
We re que s t that you include consideration and evaluation of the following possible actions:
(1) reduction of further neutron radiation damage at the beltline by replacement of outer fuel assemblies with dummy assemblies or other management changes;
Response
The TMI Plant Specific Report will be submitted to the Staff at the end of June, 1982.
As we discussed in our response to Question (2) of Enclosure (2), NRr.
Letter of 12/18/81, we are evaluating the transition to an 18 t.cnth LBP fuel scheme. An approximate 25% reduction in fluence would result from the LBP refueling approach.
Question (2) reduction of the thermal shock severity by increasing the ECC water tempe ra ture ;
Response
Consistent with the results of our analysis, we do not intend to increase the ECC water temperature.
Our reasoning, based on the parameters in the analyzed transients (i.e.,
ECC temp. of 40* and 50') and based on the benchmarked mixing Commix 1A model, indicate minimal advantage in raising the ECC temperature.
As part of our PTS evaluation, we undertook a brief
study of actual BWST temp during tho history of THI-1 oparctions. The lowest temperature recorded was $8'F for one day while an 80*F temperature is the norm.
Therefore, we conclude that our analysis has a degree of conservatism based on the historical data.
Question (3)
Recovery of RPV toughness by in-place annealing (include the basis for demonstrating that your plant meets the requirements in 10 CFR 50 Appendix G IV C ).
Response
GPU Nuclear within the B&W Owner's Group is continuing to monitor the current EPRI sponsored research program on in place annealing.
Question (4) Design of a control system to mitigate the initial thermal shock and control repressurization.
Response
As our analysis and our discussions on operator actions indicates, we consider no changes are needed to our existing control systems.
Likewise the upgraded EFW, cavitating venturis in the EFW and HPI systems, redundant power supply to instruments and the upgrading of the position indication of the PORV and pressurizer safety valver,, have all contributed to the enhanced system response to mitigate potential PTS challenges to the vessel.
_