ML062360115
| ML062360115 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 08/22/2006 |
| From: | - No Known Affiliation |
| To: | Office of Nuclear Reactor Regulation |
| References | |
| FOIA/PA-2006-0007 | |
| Download: ML062360115 (4) | |
Text
LERF Significance Determination for Dresden HPCI Water Hammer Finding The dominant CDF sequences from the SDP worksheets (for internal events) serve as a starting point for the ALERF determination. For LERF considerations, only those sequences at the highest order of magnitude (in this case 1 E-6) and one order of magnitude lower need to be considered. For Mark I containments, CDF sequences of importance for LERF are ISLOCA, ATWS, and transients (including SBO) that can lead to reactor vessel breach (VB) at high pressure (irrespective of drywell floor conditions), or at low pressure with a dry drywell floor.
The dominant CDF sequences from the SDP worksheets that meet these criteria are listed in Table 2 below.
A simplified containment event tree (CET) can be constructed with the following top events:
- operator depressurizes RCS after core damage but prior to reactor vessel breach (probability of successful depressurization = POPdep)
- RCS depressurizes passively due to a stuck open safety relief valve or other mechanisms (probability of passive depressurization = PsoRv
- DW is flooded at time of VB due to either LOCA break flow, DW spray operation, or containment flooding (probability of a wet drywell floor = Pflod)
- Containment remains intact following VB, i.e., no liner melt-through. The split fractions assigned for branch correspond to the LERF Factors in Appendix H for the 4 possible RCS pressure/drywell floor combinations. (These values are shown in the shaded portion of Table 2 below based on the current draft of Appendix H.)
The effective/weighted LERF Factor is the sum of the frequency of all CET endstates that do not result in an intact containment. This can be expressed as:
Factor = Poeprdps * (0-PlPd)+ (0 POPdep) * [(PsoRv) * (1- Pfljod) + (1-PSORv) * (1- 0.4* Pfod)]
These probabilities would be evaluated on a sequence by sequence basis.
Licensee Assessment For the TPCS sequence, the licensee assumed a probability of 0.16 that the RCS would not be depressurized through either:
- operator action (0.52 probability of failure), or passive means, including
- SORV (0.55 probability of no SORV),
- temperature-induced failure of the RCS during core damage (0.7 probability of no RCS failure), or
- pressure-induced failure of the RCS during core damage (0.8 probability of no RCS failure).
Supporting documentation from the PRA notebooks was provided for each of these values.
In follow-up discussions, licensee staff (R. Johnson) indicated that the operator failure probability is very conservative (given failure of depressurization in the TPCS sequence is dominated by operator rather than hardware failure, the significant time available to depressurize prior to VB, and the involvement of the technical support center staff in post-accident decision-making) and that an operator failure probability of 0.05 would be more appropriate for this accident class.
The licensee assumed that the drywell (DW) floor would always be flooded in this sequence.
Supporting justification includes:
consideration of the revised severe accident guidelines, which have relaxed the drywell spray initiation limit to allow spray operation over a broader range of containment conditions, and now direct the operator to initiate DW sprays upon high drywell radiation confirmation that the containment conditions prior to VB would not preclude DW spray initiation
- consideration of the various systems that can be used for injection.
The licensee computes a weighted-average LERF factor of 0.098 for the Class 1A cutsets, which include the TPCS sequence. The licensee's model also includes a small contribution to LERF from accidents in which containment integrity is maintained but the DW is vented. This is not included in the staff's assessment.
Staff Assessment As a base case, the staff adopted a 0.5 probability of successful operator depressurization, a 0.5 probability of depressurization via passive mechanisms, and a 1.0 probability of a wet DW floor. The staff explored the impact of variations in these parameters via sensitivity calculations as summarized below:
The staff acknowledges that the probability of successful operator depressurization is higher than 0.5 and considers a value of 0.95 more realistic for this accident class. The staff notes that additional strategies/systems, beyond those credited for core damage prevention, can be used for late depressurization (such as opening MSIVs or use of the HPCI steam line) lending credence to the use of higher probabilities. The use of the higher success probability was considered in sensitivity calculations.
Credit for passive depressurization mechanisms is difficult to defend as these failure modes/mechanisms are complex and highly uncertain, especially temperature- and pressure-induced failure modes. A base case value of 0.5 was used to reflect the possibility of an SORV, but even this beneficial failure is speculative and difficult to quantify. The staff considers it more appropriate to not credit these passive mechanisms in a scoping assessment and addressed the impact of eliminating these beneficial failures via sensitivity calculations.
The staff agrees that the probability of achieving a wet drywell floor is very high, given the revised severe accident guidelines (SAG) and availability of support systems in this 2
sequence. However, since drywell floor conditions have the greatest impact on the LERF Factor, the use a lower probability of drywell flooding was explored via sensitivity calculations. A value of 0.9 was used to represent the possibility that drywell flooding, although called for by the SAG, is not accomplished due to either hardware failure or unavailability or a technical support center decision to avoid the use of sprays. A value of 0.5 was used to represent a worst case situation.
The base case and sensitivity calculations are summarized in Table 1 below. The weighted LERF Factor ranges from 0.015 to 0.575 considering the full range of parameters. Using more realistic and defensible probability values, the LERF factor ranges from about 0.1 to 0.2. (e.g.,
Cases 2 and 7). For purposes of the Phase 3 assessment, a weighted LERF Factor of 0.2 is suggested.
Results Use of a LERF Factor of 0.2, in conjunction with the ACDF estimate of approximately 4E-6 from the Level 1 analysis, results in a ALERF of 8E-7, which is considered a WHITE finding.
Alternatively, each of the dominant CDF sequences can be assessed using the LERF Factors and counting rules contained in the current draft of Appendix H. The dominant CDF sequences from the SDP worksheets are listed in Table 2 below. The estimated CDF, based on the proposed counting rules in the current draft of Appendix H, is also provided. The fraction of the CDF associated with each of the four RCS pressure/DW floor combinations (based on the base case probability values) and the corresponding LERF Factors are also provided. For the lower magnitude sequences, the weighted average LERF Factor of 0.2 is used as a simplification.
Marginal changes from the 0.2 value would not be significant since the frequency of these sequences is a decade lower than the TPCS sequence. The total ALERF for these sequences is 7E-7, which is close to the previous estimate.
Conclusions Using the guidance contained in the current draft of Appendix H, and realistic and defensible estimates of the probability of RCS depressurization and drywell flooding, a weighted LERF Factor of 0.2 is obtained. Use of this factor, in conjunction with the ACDF estimate of approximately 4E-6 from the Level 1 analysis, results in a ALERF of 8E-7, which is considered a WHITE finding.
3
Table 1 - Weiahted LERF Factors for TPCS Case POPdep PSORV Pb Weighted LERF Factor Base Case 0.5 0.5 1.0 0.15
- 1. More realistic operator depressurization 0.95 0.5 1.0 0.015 probability
- 2. More realistic probability of drywell flooding 0.5 0.5 0.9 0.235
- 3. Pessimistic probability of drywell flooding 0.5 0.5 0.5 0.575
- 4. No passive depressurization 0.5 0. 1.0 0.3
- 5. No passive depressurization and more 0.5 0. 0.9 0.37 realistic probability of drywell flooding
- 6. More realistic operator depressurization 0.95 0. 0.5 0.515 probability, no passive depressurization, and pessimistic probability of drywell flooding
- 7. More realistic operator depressurization 0.95 0. 0.9 0.127 probability, no passive depressurization, and more realistic probability of drywell flooding Table 2 - ALERF Contribution from Dominant CDF Sequences Dominant CDF Fraction RCS Press Drywell LERF ALERF Sequence Contrib at VB Floor at VB Factor TPCS (6) 3.3E-6 0. 0.lli 0.25 1 igb1 5E-7
- 0. 0.
0.75 Negligible MLOCA (6) 3.3E-6 1.0 Low Wet <0.1 Negligible LOIA (7) 3.3E-7 Use Factor of 0.2 based on assessment for 7E-8 TPCS LOOP (7) with 3.3E-7 Use Factor of 0.2 based on assessment for 7E-8 depress failure TPCS LOOP (7) with 3.3E-7 Use Factor of 0.2 based on assessment for 7E-8 EAC failure TPCS Total ALERF 7E-7 4