ML20097J763
| ML20097J763 | |
| Person / Time | |
|---|---|
| Site: | Dresden, Quad Cities  |
| Issue date: | 06/27/1995 |
| From: | Dezfuli H SCIENTECH, INC. |
| To: | NRC |
| Shared Package | |
| ML17180B458 | List: |
| References | |
| SCIE-NRC-221-93, NUDOCS 9509010251 | |
| Download: ML20097J763 (28) | |
Text
.
~
SCIE-NRC-221-93
~
DRESDEN NUCLEAR POWER STATION TECHNICAL EVALUATION REPORT ON THE INDIVIDUAL PLANT EXAMINATION BACK-END ANALYSIS
~
s Homayoon Dezfuli James F. Meyer Prepared for the U.S. Nuclear Regulatory Commission Under Contract NRC-05-91-068-20 March 1995 SCIENTECH, Inc.
11140 Rockville Pike, Suite 500 Rockville, Maryland 20852 5 0 90 f 016 KA 1s!'il ri
. _ ~
- -.- ~..
I TABLE OF CONTENTS i
Pace E.
EXECUTIVE
SUMMARY
iv E1.
PLANT CHARACTERIZATION iv E2.
LICENSEE IPE PROCESS iv I
E3.
BACK END ANALYSIS.
iv E4.
CONTAINMENT PERFORMANCE IMPROVEMENTS (CPI)
.v E5.
VULNERABILITIES AND PLANT IMPROVEMENTS
.v B6.
OBSERVATIONS vi I.
INTRODUCTICE
.1 T.1 REVIEW PROCESS 1
I.2 PLANT CHARACTERIZATION 1
II.
TECENICAL REVIEW.
3 II.1 LICENSEE IPE PROCESS.
3 s
II.1.1 Completeness and Methodology 3
II.1.2 Multi-Unit Effects and As-Built As-Operated Status
.4 l
II.1.3 Licensee Participation and Peer Review
.4 II.2 CONTAINMENT ANALYSIS /CHARACTERIEATION
.4 II.2.1 Front-end Back-end Dependencies
.4 II.2.2 Sequences with Significant Probabilities
.5 II.2.3 Failure Modes and Timing
.6 II.2.4 Containment Isolat ion Failure
.9 II.2.5 System / Human Response
.9 II.2.6 Radionuclide Release Characterization 10' Dresden IPE Back-End Analysis il March 1995
i 4
t l
TABLE OF COblTENTS (cont. )
i I
II.3 ACCIDENT PROGRESSION AND CONTAINMENT PERFORMANCE ANALYSIS 10 II.3.1 Severe Accident Progression.
10 II.3.2 Dominant Contributors: Consistency with IPE Insights 11 II.3.3 Characterization of Containment Performance 12 II.3.4 Impact on Equipment Behavior 13 II.4 REDUCING PROBABILITY OF CORE DAMAGE OR FISSION PRODUCT RELEASE 14 II.4.1 Definition of Vulnerability 14 II.4.2 Plant Improvements 15 II.5 RESPONSES TO CPI PROGRAM RECOMMENDATIONS 15 II.6 IPE INSIGHTS, IMPROVEMENTS AND COMMIDENTS 16 III. CottTRACTOR OBSERVATIONS AND CONCLUSIONS 17 s
IV.
REFMCES.
18 APPENDIX 19 i
i e
1 Dresde IPE Back-End Analysis lii Ms.ch 1995
E.
EXECUTIVE
SUMMARY
E1.
PLANT CHARACTERIEATION Both. of the Units at the Dresden Nuclear Power Station (Units 2 and 3) are BWR-3s with Mark-I containments.
Each unit is rated at 773 MWE and is equipped with an isolation condenser.
E2.
I.ICENSEE IPE PROCESS The IPE conducted at Dresden was a PRA study based on the support state model.
In performing a PRA, the total plant response to each severe accident scenario was modeled in plant response trees (PRTs) in order to integrate front-end and back-end analyses.
The MAAP computer code was used to characterize success criteria, timing, and containment response.
4 The Dresden IPE back-end submittal reflects the plant as it was configured in January 1991.
Although the " hardened" vent system was not yet operative at Dresden in 1991, the IPE team took credit for it.
The hardened vent became operational for Unit 3 in 1993 and for Unit 2 in 1993.
The IPE team integrated accident management (AM) considerations into the IPE model.
The team modeled the Emergency Operating Procedures -(EOPs) using generic Boiling Water Reactor Owners Group (BWROG) symptom-based guidance.
The IPE team also
~
identified a class of core damage sequences referred to as Containment Success with Accident Management (CAM) sequences in which no containment failure occurs within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
After the 1
first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />, however, accident management is necessary to achieve long-term containment integrity.
E3.
BACK END ANALYSIS Important results reported in the Dresden IPE back-end submittal are as follows:
The conditional probability of containment failure and/or containment venting with significant releases was 89 percent.
The conditional probability of early containment failure with large releases was 3 percent.
The core damage frequency (CDF) was 1.85E-5 per year.
i
Failure of,the suppression pool cooling (SPC) function accounted for 82 percent of the CDF.
The Dresden IPE team concluded that a single accident initiated by a loss of DC power, and in.which the SPC failed, would constitute the dominant accident sequence, accounting for 44 percent of the CDP.
A single accident initiated by a loss of DC power, and in which SPC failed due to operator error, would account for 9 percent of the CDP.
Modes of contininment failure and their respective conditional probabilities at Dresden were reported to be the following:
High-temperature structural failure after the containment was vented: 84 percent Rapid, high-pressure failure of the containment:
3 percent 4
Late high-pressure and/or high-temperature failure:
2 percent Based on these data, the conditional probability of the containment remaining intact is 11 percent.
34.
CONTAIMMENT PERFORMANCE IMPROVEMENTS (CPI)
Based on Generic Letter No. 88-20, Supplement No. 1, the
~
following CPI Program recommendations pertain to Mark-I containments:
Create an alternate water supply for drywell spray / vessel injection (Procedure under consideration)
Enhance reactor pressure vessel depressurization system reliability (No Action)
Implamant emergency procedures and training (Procedures under consideration in Accident Management Program)
Evaluate and initiate operation of the hardened vent (hardened vent installed).
1 E5.
VULNERABILITIES AND PLANT IMPROVEMENTS The IPE team identified several enhancements that would reduce SPC failures.
These are now under consideration for implementation.
The major enhancement, which the NUMARC Severe Dresden IPE Back-End Analysis v
March 1995
Accident Issue Closure Guidelines reconunends, would be the introduction of a procedure to align low-pressure coolant injections (LPCI) or core spray (CS) pump suctions with the condensate storage tank (CST) when SPC cannot be established.
This enhancement would ensure the retention of coolant injected into the reactor vessel when it would otherwise be lost.
The-coolant is lost when the net-positive suction head (NPSH) for the low-pressure emergency core cooling system (ECCS) pumps is not sufficient as the suppression pool water is heated.
If this enhancement were implemented, the IPE team stated that the CDF would.be reduced substantially.
The team that was responsible for the Dresden Nuclear Station individual plant examination (IPE) back-end submittal concluded that no severe accident issues existed at Dresden that warranted immediate remedial action.
i E6.
OBSERVATICIts l
In conducting the review, SCIENTECH noted the following from the Dresden Nuclear Power Station back-end submittal:
Although the two units at Dresden could share mitigating capabilities, one unit was susceptible to a loss of DC power in the other unit.
Given a core damage accident, the probability of venting the containment and not having a subsequent containment failure was very low.
s Based on SCIENTECH's review, the following strengths and weaknesses in the Dresden Nuclear Power Station IPE back-end submittal are noted:
The IPE back-end submittal is well structured and well written.
As part of the IPE, the IPE team conducted several experiments to investigate lower vessel head cooling.
The team appears to have used a systematic process to gain insights from the IPE.
The IPE team did not fully describe the PRT structures in the back-end submittal.
i Dresden IPE Back-End Analysis vi March 1995
The IPE results indicate that the containment is sensitive to the loss of DC power.
The team did not suggest in the submittal a preventive measure to reduce or eliminate this initiating event.
No independent ' review of the IPE was perfonned.
~
s i
% IPE Back-End Analysis vii March 1995
i 1
l I.
INTRODUCTION I.1 REVIEW PROCESS This technical evaluation report (TER) documents the results of the SCIENTECH review of the back-end portion of the Dresden Nuclear Power Station Individual Plant Examination (IPE) submittal [1,2].
This technical evaluation report complios with the requirements for IPE back-end reviews of the U.S. Nuclear Regulatory Commission (NRC) in its contractor task orders, and l
adopts the NRC review objectives, which include the following:
To determine if the IPE submittal provides the level of detail requested in the " Submittal Guidance Document,"
NUREG-1335 To assess if the IPE submittal meets the intent of Generic Letter 88-20 To complete the IPE Evaluation Data Summary Sheet A draft TER for the Back-End portion of the Dresden IPE submittal was submitted by SCIENTECH to NRC on November 17, 1993.
Based in part on this draft submittal, the NRC staff submitted a Request for Additional Information (RAI) to Commonwealth Edison on August 24, 1994.
Commonwealth Edison responded to the RAI in a document dated October 28, 1994.
This final TER is based on the original
~
submittal and the response to the RAI.
Section II of the TER summarizes SCIENTECH's review and briefly describes the Dresden IPE submittal, as it pertains to the work
~
requirements outlined in the contractor task order.
Each portion of Section II corresponds to a specific work requirement.
Section II also outlines the insights gained, plant improvements identified, and utility commitments made as a result of the IPE.
'l Section III presents SCIENTECH's overall observations and conclusions.
References are given in Section IV.
The Appendix contains an IPE evaluation and data summary sheet.
1 I.2 PLANT CHARACTERIZATION The Dresden containment is a BWR-3 Mark I design.
The primary containment surrounds the reactor vessel and circulation cooling system.
Any leakage from the primary containment will go directly to the reactor building.
The primary containment consists of the following:
Drywell.
The free volume is 158,236 f t" with a gas space Dresden IPE Back-End Analysis March 1995 b
,r
~.
\\
height of 102 feet.
The drywell is a steel pressure vessel with a 66-foot-diameter spherical lower portion and a 37-foot-diameter cylindrical upper portion.
The vessel is enclosed in reinforced concrete with a 2-inch gap between the steel shell and the concrete.
The design pressure range of the drywell is 62 psig and -2 psig at 281
'F.
The ambient temperature range is 135 to 150
'F.
Pressure suppression pool torus (Wetwell).
Used to control drywell pressurization under accident conditions.
Interconnecting vent pipes.
Eight circular vent pipes connect the drywell to the wetwell.
The pipes enclosed in sleeves are provided with bellows to accommodate differential motion between the drywell and the wetwell.
The pipes are connected to at vent header located in the air space of the wetwell.
The wetwell and drywell may be vented through use of a Standby Gas Treatment (SBGT) system or preferably through a 10-inch
' hardened" vent to the 310-foot chimney.
In addition to the suppression pool, the IPE team used the following systems to model primary containment pressure control:
Low-Pressure Coolant Injection (LPCI).
This can be aligned with either the drfwell or wetwell spray headers.
Operator actions that are taken to restart the drywell coolers (Upon activation of the core spray system, the i
drywell cooler fans trip.)
Drywell or wetwell venting.
Dresden IPE Back-End Analysis 2
March 1995
-._ _.. ~. _ _ _ _. _ - _. _ _ _ _ _ _ _ _. _ _. _ _ _. _ _ _ _ _ _. _
j l
i l.
4 i
II.
TECENICAL REVIEW i
II.1 LICENSEE IPE PROCESS i
i i
II.1.1 Completeness and Methodology i
i The Dresden Station IPE back-end submittal is essentially complete l
with respect to the level of detail requested in NUREG-1335.
i The Dresden IPE team used an integral approach to model the plant j
response from initiating event through the entire accident
}
progression, including the containment response.
This integrated l
approach combined the traditional Level I and Level II analyses
{
into a single model.
Each initiating event was tracked through a PRT to evaluate the success or failure of each plant system, j
operator action, and containment system.
The PRTs developed for
{
use in the Dresden IPE were complex and contained many top l
events.
For example, the PRT for the general transient 1
initiating event included 26 top events.
To deal with complexity of the PRTs, the team broke them down into several subtrees.
A i
complete list of the PRT top events appears in Table 4.1.3-4, j
pages 4-31 and 4-32 in Volume 1 of the submittal.
The PRT structures are provided in Volume 2.
Key support systems were
}.
modeled using support state methodology.
In Section 4.5.3 of the submittal, the split fractions for support system event trees are j
provided.
Also provided is the split fraction for PRT nodes for l
the dominant-100 core damage sequences.
i i
The team used MAAP analysis to develop physical plant models to l
(1) define PRT nodal success criteria, (2) establish the timing of accident progression, and (3) determine accident outcomes.
i PRT outcomes.(end states or plant damage states) were categorized j
as any of the following:
1 SCS --
Success SAM --
Success with accident management, i.e.,
accident sequences that require accident management activities 24 I
hours after the core damage event in order to achieve an ultimately safe, stable containment state Core damage CD Sequences ending in core damage (i.e., CD endstates) each received a 5-character' code to characterize their unique core and containment response characteristics.
The team identified many CD plant damage states.
Dresden IPE Back-End Analysis 3
March 1995
II.1.2 Multi-Unit Effects and As-Built As-Operated Status The Dresden IPE back-end submittal reflects the plant as it was configured in January 1991.
Although the " hardened" vent system was not yet operative at Dresden in 1991, the IPE team took credit for it, The hardened vent became operational for Unit 3 in 1993 and for Unit 2 in 1993.
Although the two units at Dresden could share mitigating capabilities, one unit was susceptible to a loss of DC power in the other unit.
II.1.3 Licensee Participation and Peer Review Commonwealth uMarm Company (CECO) engaged the Individual Plant Evaluation Partnership (IPEP) to support the conduct of the Dresden IPE.
'Ibe IPEP participant companies were Westinghouse, Fauske and Associates, Inc., and TENERA.
IPEP personnel performed the basic modeling and analysis.
The CECO staff performed the success criteria analysis using MAAP and they conducted detailed reviews of the models, assumptions, and results.
The sensitivity analyses that the team performed and i
reported on in the submittal were based on NUREG-1335 and EPRI-TR-100167 recommendations.
A " Tiger Team" composed of IPEP and CECO personnel evaluated the insights gained from the IPE.
The IPE team stated that,.an~IPEP Senior Management Support Team (SMST), whose members were not involved in the day-to-day conduct of the IPE, also reviewed the key insights and key results.
On page 1-2 of the submittal, it is stated that:
As noted in the initial CECO responses to the Nuclear Regulatory Commission (NRC) on Generic Letter 88-20, no separate independent review of the Dresden IPE was performed.
II.2 CONTmmerr ANALYSIS / CHARACTERIZATION II. 2.1 Front-end Back-end Dependencies The PRTs developed for the Dresden IPE reflect that reactor systems and containment systems were treated collectively.
Using the PRT concept, front-end and back-end analyses were integrated into one model.
The front-end back-end dependencies were accounted for in the structure of the PRTs.
Dresden IPE Back-End Analysis 4
March 1995
I
)
l The core damage frequency by plant damage state (PDS) is shown in Table 1.5.1-2, page 1-25 of the submittal.
II.2.2 Sequences with Significant Probabilities The core damage frequency for Dresden Station was calculated to The IPE submittal states that, of this be 1.85E-5 per year.
total, the frequency of core damage where the containment remained intact and was not vented was 2.1E-6 per year.
This implies that the frequency of core damage and containment failure (including venting) was 1.85E-5 -'2.1E-6 = 1.64E-5 Stated differently, the conditional' containment failure, given a core damage accident, was approximately 89 percent.
Table 1.5.1-2 of the submittal lists the dominant core damage frequency accidents with their associated plant damage states.
The top three dominant accident sequences are shown in Table 1 of this report.
Table 1 ConIribution PDS Description Frequency (per year) to CDF DLCO' Loss of DC power with 1.06E-05 57.2 late core damage (6-24 hours) and SPC failure LLCO Loss of offsite power 3.27E-06 17.7 (single or dual unit) with late core damage (6-24 hours) and SPC failure MLCO Medium LOCA with late 7.56E-07 4.1 core damage (6-24 hours) and SPC failure The three sequences described in Table 1 accounted for approximately 80 percent of the total CDF.
Each sequence 5
March 1995 DresdenIPEBad-End Aantysis i
described a plant damage state in which there was a low-pressure vessel melt-through with no water applied to the debris bed.
The containment failed after having been vented previcusly.
As can be seen from the table, the most dominant PDS involved the loss of DC power and subsequent loss of SPC.
Using the data contained in the submittal, which were based on the top-100 dominant core damage accidents, SCIENTECH attempted
)
t to construct the Dresden conditional containment failure probabilities. (These values, with minor changes, were later j
verified in the Commonwealth Edison 10/20/94 response to RAI.)
i l
Table 2 i
Failure Mode Conditional probability of j
Containment failure 1
Intact 11.2%
Vented and failed late 84.2%
Late, high-temperature /high-1.5%
~
s pressure Early 3.0%
Bypass / isolation 0
II.2.3 Failure Modes and Timing In the Dresden IPE, the fission product releases for core damage sequences represented on the PRTs were calculated using a time
'This value is for when the containment is intact and not vented for at least 21 hours2.430556e-4 days <br />0.00583 hours <br />3.472222e-5 weeks <br />7.9905e-6 months <br />.
The value of 0.3% given on page 7-1, third paragraph, and in Table 7.1-3 on page 7-5 of the subnittal is for when the containment is intact and not vented for 11 hours1.273148e-4 days <br />0.00306 hours <br />1.818783e-5 weeks <br />4.1855e-6 months <br />.
(Note that the value of 0.3% was corrected to 0.4%
in the 10/28/94 response to RAI.)
~
Dresden IPE Back-End Analysis 6
March 1995
J r
i a
\\
frame of 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> from the initiation of the accident.
The IPE I
team extended the IPE window, which is 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
j ii The Dresden containment fragility curve presented in the 1
i submittal-shows that the mean failure pressure was approximately 1
105 psig for temperacures below 281
'F.
Low-pressure containment-i failures were dominated by drywell head closure.
If the containment were to fail at a relatively high pressure, the likely location would be the vent line bellows.- On pagS 4-113, i
the IPE submittal provides mean failure pressures for critical l
locations as shown below 4
Drywell shell 125 psig j
Equipment hatch
>165 psig j
Personnel airlock 150 psig i
Mechanical penetrations 140 psig Electrical penetrations
>150 psig
{
Drywell head closure 125 psig (leakage) j Vent line bellows93-253 psig (leakage) i Wetwell shell 125 psig Using the terms "unlikely" and "likely' in a conditional l
probability sense, the IPE team defined unlikely and likely i
failure modes in the IPE submittal, which are sununarized below.
i Unlikely failure modes included steam explosion, vessel thrust l
forces, molten core-concrete attack, direct containment heating, j
thermal attack of containment penetrations, hydrogen ccanbustion, and containment isolation.
Likely failure modes were stated to be containment high pressure, containment high temperat,ure, j
unscrubbed venting of the drywell, and liner melt-through.
Table 4.3.3-1, page 4-112 of the submittal, summarizes the IPE i
team's phenomenological evaluations of all the containment j
failure modes postulated for Dresden.
l~
Steam Explosion (unlikely). The IPE team concluded that the slumping of molten debris into the RPV lower plenum could not result in an in-vessel steam explosion.
Moreover, the submittal states that an ex-vessel steam explosion would i
l not threaten containment integrity.
Vessel Thrust Forces (unlikely). The submittal states that, l
j based on a bounding analysis, this. failure mode is highly unlikely.
Even if vessel thrust forces did occur, they l
would not threaten containment integrity.
Molten Core-Concrete Attack (unlikely). The IPE team assessed the probability of a molten core-concrete attack within the Dresden containment, based on the most conservative scenario in which the debris is not coolable-1 j
by direct contact with an overlying pool of water.
The IPE 1
j Dneden IPE Back-End Analysis 7
March 1995 i
l l
.w
=..
team concluded that melt-through of the pedestal walls would not occur until well beyond the mission time of the IPE.
Direct containment Beating (unlikely). The submittal states that DCH is not a potential early containment failure mode.
The ADS operation and " tightly packed" geometry of the l
drywell would inhibit the pressure rise associated with l
.DCH.
Thermal Attack of Containment Penetration (unlikely). In accident scenarios where debris coolability is absent, drywell integrity may be a concern.
However, seal degradation generally will occur within a few days, if drywell airspace is not coo 3 =d.
Eydrogen Combustion (unlikely).
The team assessed the potential for H, combustion in a worst-case scenario involving an SBO event without initial containment inertion.
The submittal states that H,-combustion-induced containment failure can occur when AC power is recovered and the drywell sprays are initiated without first venting i
the wetwell.
The team concluded that this situation can be l
avoided if wetwell venting is implemented before any attempt is made to use the drywell sprays.
Containment Bypass (unlikely).
According to the submittal, the frequency of interfacing systems LOCA was estimated to
{
be on the order of IE-10.
For that reason, this.4ai" lure j
mode was not analyzed.
Containment Isolation (unlikely).
Addressed in Section II.2.4 of this report.
Containment Eigh Pressure and Eigh Temperature (likely).
The submittal states that the potential for such failure exists in severe accidents where sufficient containment heat removal is not available.
These accidents are primarily ATWS and station blackout events where venting is not possible, or is ineffective.
The containment fragility curve in the submittal shows a probability of failure at 105 psig where temperatures are below 281
- F.
The probability decreases to 62 psig at 500 *F.
Because the containment gas temperature!near the drywell head can reach the 300-500 *F range during severe accidents, leakage through the drywell closure can be expected at pressures less than 105 psig.
Liner Melt-through (likely).
The submittal states that liner melt-through will occur during a core-melt scenario Dresden IPE Back-End Analysis 8
March 1995 i
in which the debris on the drywell floor is not cooled by.
an overlying pool of water.
Once the debris is in contact
- with the shell, the shell will fail rather quickly.
Because the sump in the pedestal region cannot hold all the possible molten debris, some of it may exit the pedestal region, make its way across the drywell. floor, and eventually come into direct contact with the drywell steel shell.
The IPE team treated this model of failure as part of its sersitivity analysis.
1 II.2.4 containment Isolation Failure 4
The IPE team did not consider any containment isolation failures, although three conditions are described on page 4-119 of the sulhamittal that may lead to containment isolation failures.
The substittal states that none of these conditions would be likely because the Dresden containment is inerted during normal operation and would be isolated from the outset of a severe accident.
II.2.5 SystesVHuman Response The submittal states that the most significant operator-related contributions to the loss of heat removal from the containment are the failure to implement suppression pool cooling, the failure to make up to the shell side of the isolation cond,enser, and the failure to depressurize the reactor vessel when" required.
' The second most likely sequence belonging to the most dominant
- PDS, i.e., DLCO (see Table-1 of this report)'is related to a loss of DC power in one unit and the subsequent loss of suppression pool cooling due to operator error.
This sequence accounts for about 9 percent of the total core damage frequency.
The IPE team identified a class of accident sequences referred to as success with Accident Management (SAM) sequences, which do not result in core damage within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of the initiating event, but do require additional operator action af ter the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> in order to achieve a long-term safe, stable containment state.
These sequences are primarily ATWS events.
Table 4.6.5-4, page 4-252 of the submittal, lists the dominant operator failure modes for SAM accident sequences.
The data show that these failures are the results off operator errors of omission, which, according to the submittal, can be remedied long af ter the errors are made.
d DresdenIPEBack-End Analysis 9
March 1995 y
m--
._ma
l i
i j
l
?
II.2.6 Radionuclide Release Characterization f
i The IPE team perfonned source-term calculations based on the 100 highest-frequency sequences.
To further reduce the fission 3
product release calculations, these sequences were grouped into seven source-tem bins.
1 For each bin, the dominant sequence belonging to the bin was j
selected for source-term evaluation.
For example, the sequences listed in Table 1 of this report belong to a source-term bin labeled "CO."
Table 4.5.5-3, page 4-191 of the submittal, j
tabulates the source-term analysis results for each source-term bin.
The sequences were analyzed for 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br />.
The fission i
product releases were reported after 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> of elapsed sequence time.
For the dominant source-term bin, i.e.,
CO, the following data were provided:
Fraction of clad reacted in vessel 0.1067 t Noble release 99.9
% Volatile FP release (CsI and RbI) 5.9
% Non-volatile FP release (SrO) 0.16
% Tellurium-based FP release (Te2 TeO2) 32.3
)
l II.3 ACCIDENT PROGRESSION AND CONTAINMENT PERFORMANCE ANALYSIS
~
s II.3.1 Severe Accident Progression
'Ibe IPE team used a special version of the MAAP code (MAAP BWR 3.0B Revision 7.03B) co develop its accident progression analysis.
The si*mittal states that this version of the code was modified as to its in-vessel fission product retention capabilities.
By varying the use of MAAP options, the IPE team investigated the influence of the core-melt progression model, and ex-vessel core debris coolability on containment failure timing.
Key results of the MAAP analysis for dominant sequences are sununarized in Table 4.5.5-3 on page 4-191 of the submittal.
For example, for the Sequence DLCO, (see Table 1 of this report) which is represented by the source-term bin, CO, the following data were provided:
Time of core uncovery (hr) 11.4 Time of core relocation (hr) 13.1 Time of vessel failure (hr) 16.2 Time of containment failure (hr) 27.5 Dresden IPE Back-End Analysis 10 Man:h 1995 l
i
~.
Time of venting (hr) 16.5 Maximum drywell pressure;(psig) 57.1 Maximum drywell temperature (*F) 668.
The IPE team performed sensitivity analyses, based on the following information sources:
Table A.5 in NUREG-1335 EPRI report EPRI-TR-100167, " Recommended Sensitivity Analyses for an Individual Plant Rwamination Using MAAP 3.0B' IPE analyst insights.
Tables 4.5.6-1 through 4.5.6-4 on pages 4-204 through 4-214 of the submittal' tabulate the scope and results of the sansitivity analyses performed during the IPE.
II.3.2 Dominant Contributors: Consistency with IPE Insights In Table 3 below, the conditional probabilities for the occurrence of the various containment failure modes are compared as they were set out in the Dresden IPE submittal, in several other BWR plants IPE submittals, and in the Peach Bottom NUREG-1150 scudy.
~
/
Table 3 Cantalement Fitapotrick Oyster Creek tre e s Ferry tune arnetd Peach gottem/
Cooper Dresden Feiture IPE -
IPE 1FE WE umEG-1150 IFE IPE IDF (per yeer) 1.9E 6 3.2E-6 4.8E-5 T.8E-6 4.5E 6 T.1E 5 1.8E 5 Early felture 60 16 4
47 56 36 3
typass no 7
no 9
ns 0
0 Late felture 26 26 26 32 16 31 86 Intact 3
0 5
21 18 33 11 to vesset 11 51 25 mm 10 ne no breech Dresdes IPE Back-End Analysis 11 March 1995
l l
i Based on the submittal data, most of the core damage postulated j
for Dresden would occur-late (6-24 hours).
Rapid,.high-pressure failure of the containment was calculated to make only a small l
contribution to the total CDFs.
ATWS sequences would account for most of the rapid, high-pressure failure of the containment.
Of the late containment failures, 84.2 percent would be associated with sequences in which the containment failed due to high temperature after having been vented previously.
The contribution of sequences resulting in late, high-temperature /
high-pressure containment failures would be less than 2 percent.
The IPE team concluded that the contribution of interfacing systems LOCA to CDF would be negligible.
The most significant l
contributor to core damage frequency would be the loss of suppression pool cooling.
The plant's capability to remove heat from the containment would be reduced severely if there were a loss of 125VDC in one unit that resulted in the loss of isolation condenser and suppression pool cooling.
l II.3.3 C15aracterisation of Containment Performance l
l The Dresden containment is a BWR-3 Mark I design.
The primary containment surrounds the reactor vessel and circulation cooling system.
Any leakage from the primary containment will go directly to the reactor building.
The primary containment consists of the following:
Drywell.
The free volume is 158,236 ft* with a gas space height of 102 feet.
The drywell is a steel pressgre vessel with a 66-foot-diameter spherical lower portion and a 37-l foot-diameter cylindrical upper portion.
The vessel is enclosed in reinforced concrete with a 2-inch gap between the steel shell and the concrete.
The design pressure range l
of the drywell is 62 psig and -2 psig at 281 *F.
The ambient temperature range is 135 to 150
- F.
Pressure suppression pool torus (Wetwell).
Used to control drywell pressurization under accident conditions.
i Interconnecting vent pipes.
Eight circular vent pipes connect the drywell to the wetwell.
The pipes enclosed in sleeves are provided with bellows to accommodate differential motion between the drywell and the wetwell.
The pipes are connected to a: vent header located in the air space of the wetwell.
The wetwell and drywell may be vented through use of a Standby Gas Treatment (SBGT) system or preferably through a 10-inch
- hardened" vent to the 310-foot chimney.
Dresden IPE Back-End Analysis 12 March 1995 3...-----
- - ~,
l 9
In addition to the suppression pool, the IPE team used the following systems to model primary containment pressure control:
i Low-Pressure Coolant Injection (LPCI).
This can be aligned with either the drywell or wetwell spray headers.
Operator actions that are taken to restart the drywell coolers (Upon activation of the core spray system, the drywell cooler fans trip.)
i Drywell or wetwell venting.
The nodes of the PRTs used to perform the IPE reflected both l
reactor systems and containment systems.
In the review that is the subject of this report, SCIENTECH examined the 'PRT for Lass of DC Power in Unic 2" to identify the type of nodes that the IPE team considered.
Only the containment-related PRT nodes are listed below:
e i
LP - -
LPCI pumps LV --
LPCI injection valves OSPC --
Operator action to align for suppression pool cooling SPC --
Hardware for suppression pool cooling OCNTS --
Operator action to initiate containment spray CNTS --
Hardware for containment spray OVNT --
Operator action to vent containment SVW --
Hardware for 2 IN wetwell vent
~
SVD --
Hardware for 2 IN drywell vent LVW --
Hardware for 10 IN wetwell vent LVD --
Hardware for 10 IN drywell vent WW/DW --
Location of containment failure The nodes shown above are typical of other PRTs.
II.3.4 Impact on Equipment Behavior Equipment survivability is addressed in Section 4.4.5, page 4-157 of the submittal.
Table 4.4.5-1 lists components located in areas that could become environmentally harsh following an accident.
The suhanittal concludes that, based on survivability evaluations, these components would remain operative within the IPE window (i.e., 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />).
The:IPE submittal also contains a preliminary list of equipment that would be needed to carry out post-24 hour containment accident management following a core damage event.
This equipment is listed in Tables 4.4.5-2 and 4.4.5-3, pages 4-160 and 4-161 of the submittal.
The IPE team deferred survivability evaluations of this equipment to the Dresden IPE Back-End Analysis 13 March 1995
l future, as part cf the implementation of an Accident Management Program.
j l
The IPE submittal assumes that the status of equipment at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> af ter a core damage event would apply throughout the 48-bour mission time.
II.4 REDUCING PROBABILITY OF CORE DAMAGE OR FISSION PRODUCT 4
pur.nans The team's quantitative assessment of the benefits of the proposed BCCS pump suction realignment procedural change is described in Section 6.2 of the submittal.
Implementation of this change would reduce the core damage frequency by a factor of 5 to 3.7E-6 per year.
The frequencies of the PDSs in which the containnumt were vented would be reduced by a factor of 8.
The frequency.of the PDSs in which the containment failed after o
having been vented would be reduced by a factor of 20.
The IPE team concluded that this procedural chnnge would have the most impact on the loss of DC power.
Table 7.1-3 of the submittal compares the containment performance probabilities for the IPE-based model and for the model based on the ECCS pump suction alignment procedural enhancement.
II.4.1 Definition of vulnerability J
In Section 7, page 7-9 of the submittal, it is stated that:
i Although actions will be taken to comply with the NUMARC Closure Guidelines, it 10 concluded that there are no vulnerabilities for Dresden Station which require immediate attention to improve the plant risk profile.
The IPE team grouped all of the accident sequences with respect to the NUMARC Severe Accident Closure Guidelines.
The grouping of the top-100 sequences is provided in Table 4.7.2-1, page 4-260 of the submittal.
Only Class II accidents require action under the Closure Guidelines.
The frequency of Class IB accidents calculated for Dresden fell just below the cutoff frequency for requiring implementation of Severe Accident Management Guidelines.
The plant improvements that the IPE team proposed in response to the Closure Guidelines are discussed in the next section of this report.
Dresdeo IPE Back-End Analysis 14 March 1995
R II.4.2 Plant Improvements Based on the IPE results and the NUMARC Severe Accident Issue Closure Guidelines, CECO is considering the implementation of two procedural enhancements of the Dresden Emergency Operating Procedures.
4 unhancement in Dresden Emergency Operating Procedures (DEOP 100).
The purpose of th9 procedures is to maintain the reactor vessel level when the ability to remove heat from the suppression pool has been lost.
The procedural change would enable the suctions for either LPCIs or the CS pumps to be aligned to the condensate storage tank when the NPSH limits for the ECCS pumps were reached.
The reactor vessel level would then be maintained by intermittent operation of the ECCS pumps supplied from the CSTs.
unhancement in procedures involving loss of all AC power.
o The purpose of the current procedures is to maintain the operation of the isolation condenser during extended SBO sequences.
The submittal states that this procedure could be enhanced by instructing operators to manually open the circuit breakers to the isolation condenser's 250VDC MOVs prior to depletion of the 125VDC batteries.
This would allow for continued operation of the ICs, even under SBO conditions.
~
s II.5 RESPONSES TO CPI PROGRAM RECOMMENDATIONS Based on Generic Letter No. 88-20, Supplement No.
1, the following CPI Program recommendations pertain to Mark-I containments:
Create an alternate water supply for drywell spray / vessel injection (Procedure under consideration)
Enhance reactor pressure vessel depressurization system reliability (No Action)
Implement emergency procedures and training (Procedures under consideration in Accident Management Program)
Evaluate and initiate operation of the hardened vent (hardened vent installed).
With the exception of the recommendation regarding pressure vessel depressurization, all of the above are discussed in the Dresden IPE Back-End Analysis 15 March 1995
i i
l back-end submittal as part of the IPE insights and AM program -
development.
4 II.6 IPE INSIGHTS, INPROVEMENTS AND CCtG(I' DENTS 6
In Section 5.3 of the submittal, the process is described that 1
the IPE toana used to gain insights into Dresden Station.
Two key l
AM guidelines that CECO is now considering whether to implement l
are summarized below.
Implementation of these AM guidelines i
would require hardware modifications to the plant.
Prevention of reactor pressure vessel failure.
Through its l
AM Program, the IPE team was able to verify that submersion of the bottom portion of the reactor could prevent the failure of the reactor vessel after relocation of the
)
damaged core to the lower head, assuming that the RPV l.
support skirt were modified to allow the egress of steam.
3 This modification would require placing holes in the RPV support skirt near its junction with the RPV.
The holes would provide for the escape of both air and steam from the skirt during pedestal flooding.
A related insight on the part of the team was that provisions could be made to flood the reactor pedestal independent of the rest of the containment.
This isolated flooding, which would It require structural modifications, offers several advantages.
It also would would result in the rapid submergence of the RPV.
prevent -the complete flooding of the torus, which is what the existing guidelines require.
The suppression pool as a heat sink could be lost if the primary system ruptured following floodup.
Alternate Sources of Containment Spray for Source-Term Reduction.
In recognition of the potential for large fission product releases af ter SBO and A'IWS sequences, the IPE team proposed an AM strategy to use an alternate source of contminment spray to control the fission product releases.
An alternate source of containment spray could be provided for an SBO in one unit via a cross-connection to the other unit's low-pressure core injection system or to the plant's fire protection system.
In Section 2.1.4.2 of this reports a procedural enhancement related to realignment of ECCS pump suction is discussed.
In the caemonwealth Edison 10/28/94 response to the staff of RAI, it is noted that of the 130 insights derived from the IPE, 82 are related to containment and containment systems.
The enhancement above is number DR-057/IP, in response to RAI 34.
On page 7-7; IPE states that:
DresdenIPEBack-End Analysis 16 March 1995
Based on the NUMARC Severe Accident Closure Guidelines, action any be taken to' implement a procedure to realign the suction of the ECCS pumps to the condensate storage tanks.
III.
CONTRACTOR OBSERVATIONS AND CONCLUSIONS The IPE team presented its methodology in a well-structured manner, appears to have searched for insights in a systematic way, and integrated accident managenent considerations into the IPE performance.
The results of the IPE are well presented in susmary tables.
An independent, detailed review of the IPE might be worthwhile since the team used the support state /large event j
tree approach, which is inherently difficult to review at a high i
level.
j In conducting the review, SCIENTECH noted the following from the Dresden Nuclear Power Station back-end submittal:
Although the two units at Dresden could share mitigating j
capabilities, on'a unit was susceptible to a loss of DC power in the other Unit.
-Given a core damage accident, the probability of venting the 1
containment and not having a subsequent containment failure was very low.
Based on SCIENTECH's review, the following' strengths and weaknesses in the Dresden Nuclear Power Station IPE back-end submittal are noted:
The IPE back-end submittal is well structured and well written.
4 As part of the IPE, the IPE team conducted several experiments to investigate lower vessel head cooling.
The team appears to have used a systematic process to gain insights from the IPE.
The IPE team did not fully describe the PRT structures in the back-end submittal.
i The IPE results indicate that the containment is sensitive to the loss of DC power.
The team did not suggest in the submittal a preventive measure to reduce or eliminate this initiating event.
No independent review of the IPE was performed.
Dns&m IPE Back-End Analysis 17 March 1995 w
r 6.--
-se y
r---
i
..~ _...
IV.
REFERENCES 1.
"Dresden Nuclear Power Station Units 2 and 3, Individual Plant Examination Submittal Report," Volumes 1 and 2, Connonwealth Edison Company, January 1993.
2 Commonwealth Edison, "Dresden Nuclear Power Station Units 2 and 3 Response to NRC Request for Additional Infortration (RAI),* October 28, 1994.
o
~
s j
Dresdem IPE Back-End Analysis 18 March 1995
APPENDIX IPE EVALUATION AND DATA
SUMMARY
SEEET BWR Back-end Facts Plant Name Dresden Station, Units 2 and 3 Containment Type Mark I Unique Containment Features The suppression pool (wetwell) or drywell may be vented through.either the Standby Gas Treatment (SBGT) system or directly to the 310-foot chimney through the 10-inch, hardened vent system.
The latter system, which is the called Augmented Primary Containment Vent (APCV), was not installed at Dresden at the time of the IPE.
However, for purposes of its examination, the IPE team assumed that the APCV was operational.
Low-Pressure Coolant Injections (LPCIs) can be lined up to discharge to either the drywell or wetwell spray headers for suppression pool cooling (under consideration).
Torus water and condensate storage tank inventory are available for transfer from one Dresden unit to the other (under consideration).
Dresden is equipped with an isolation condenser.
The capability exiscs for long-term ECCS injection via pump suction realignment to the CST (to be implemented via a procedural enhancement).
Unique vessel Features None were addressed.
Number of Plant Damage States 20 PDSs were defined for the top-100 core damage accident seguences.
Dresden IPE Back-End Analysis 19 -
March 1995
. ~. -
Ultimate Containment Failure Pressure 105 psig l
Additional Radionuclide Transport And Retention Structures
~"
seccodary containment system (reactor building)
Conditional Probability That The Containment Is Not Isolated
\\
Failure to isolate sequences was precluded from consideration during the IPE analysis because of the inerted condition of l
the Dresden containment.
Important Insights, Including Unique Safety Features Although two units could share mitigative capabilities, one unit was susceptible to the loss of DC power in the other unit.
Given a core. damage accident, the probability of venting the containment and not having a subsequent containment failure was very low.
The conditional probability of containment failure and/or cont =4== ant venting with significant releases was 89 percent.
The conditional probability of early containment failure with large releases was 3 percent.
The core damage frequency (CDF) was 1.85E-5 per year.
Failure of the suppression pool cooling (SPC) function accounted for 82 percent of the CDF.
Implemented Plant Improvements Of the several enhancements reported to be under consideration, none had yet been implemented, according to the j
submittal.
C-Matrix Construction of a C-Matrix was not possible because supporting data were unavailable.
I Dres&m IPE Back-End Analysis 20 March 1995
i 1
l ENCLOSURE 4 DRESDEN 4 WITS 2 & 3 INDIVIDUAL PLANT EXAMINATION TECHNICAL EVALUATION REPORT (HUMAN RELIABILITY ANALYSIS) i 1
l
.a-.,..e a
e-ns a
-+
a.-sm
+...
amana---.=-
s a m s w= s.
.a ma,m-a m...ex, e mu ss m m a m-n-w.n..u.aw mr--m w.a m m
- - ~ -
mm mm.=,mnn I
J.'
i 1
i e
i O
4 d
W e
11 s
a t
f 4
i 1
1 i
I 1
I l
_..,..._. - -