ML20211K468
| ML20211K468 | |
| Person / Time | |
|---|---|
| Issue date: | 11/03/1986 |
| From: | Bernero R Office of Nuclear Reactor Regulation |
| To: | Pickens T BWR OWNERS GROUP |
| Shared Package | |
| ML20209C630 | List: |
| References | |
| FOIA-87-10 NUDOCS 8611240005 | |
| Download: ML20211K468 (7) | |
Text
.
O
[gha agg o
~, UNITED STATES
[' s #
3 j
E NUCLEAR REGULATORY COMMISSION WASHING TON, D. C. 20555 s i g, .....f NOV 0 31986 Mr. Terry Pickens BWR Owners Group Chairman Northern States Power Company 414 Nicollet Minneapolis, Mn. 65401
Dear Mr. Pickens:
The purpose of this letter is to provide you with our coments on the August, 1986 IDCOR " Evaluation of BWR Accident Mitigation Capability Relative to Proposed NRC Changes". In general, we have found the evaluation useful.
Specifically, we are enclosing an assessment of the evaluation for your consideration. The following summarizes our present views of the IDCOR evaluation for each area of potential containment performance improvement:
- a. Combustible Gas Control - Insufficient information has been provided on a plant specific or generic basis to indicate that reducing the number of allowable hours of power cperation while deinerted would result in decreased safety.
Moreover, no bases have been provided that either a significant decrease in operational flexibility, or power operation, would result.
- b. Filtered Venting - The evaluation of potential vent size guidelines of 6 to 36 inches in diameter, with the larger size necessary to cope with an ATWS, is useful. What is not clear is whether significant modifications are required at any plant or plants to ensure reliable and remote manual opening and closing of one or more valves that vent up a plant stack. Furthermore, it is not clear what the studies that refer to other means to mitigate an ATWS encompass, nor when and if the staff will be informed of the results. Lastly, as the evaluation points out, we recognize that venting is a risk tradeoff. As a last resort to prevent a large and uncontrolled release, however, we have tentatively concluded the risks from venting virtually only noble gases through a plant stack significantly outweigh the risks of not venting.
- c. Reliable Drywell Spray - The evaluation points out that sprays could be counted on to achieve goals such as debris cooling, containment cooling, fission product attenuation.
We agree. We conclude that any potential drywell structural problems may not exist after realistic calculations are completed, or can be avoided by proper operator action, spray system modification, or by a combination. Lastly, we also agree that further study is desirable to identify the minimum l flew rate, system modifications, and related alternative power sources and water supplies necessary to achieve the goals.
QIh 2( 7 -/O w Obb&0$. _s
~
Mr. Terry Pickens ,
- d. Debris Control - Bases for the IDCOR conclusions have not been provided to allow us to agree or disagree with the conclusion that torus room barriers do "not appear to have significant benefits to justify further consideration."
Although more discussion of containment barriers was provided, the same general conclusion was reached.
Comparisons of IDCOR evaluations with Brookhaven assessments of core melt progression and core debris were provided which indicated a potential for relatively large uncertainties in debris characteristics. For example, no assessment of the uncertainties in the volume of core debris and steel was provided. Also, evaluations of the benefits and costs of various design options were not discussed. We conclude that this area needs further investigation, since it is just the uncertainties in debris characteristics that 4 indicate debris control may be necessary.
- e. Training and Procedures - We agree that emergency procedure guidelines of the type presently being proposed by the BWR owners (Revision 4) appear to be of the type necessary to implement satisfactory emergency procedures and related operator training. The staff intends to provide specific coments on the proposed guidelines (i.e., to establish appropriate water level reductions and related power levels to be better able to cope with ATWS type events). What is not clear is the positten of every BWR owner with respect to both complete adoption of the guidelines, nor the anticipated
, implementation dates for emergency procedures.
Based upon our assessment you may wish to comment further.
Sincerely, v es
- =
Robert M. Bernero, Director Division of BWR I.f censing
Enclosure:
Staff Assessment of IDCOR Evaluation
. NRC STAFF ASSESSMENT OF THE IDCOR AUGUST, 1986
" EVALUATION OF BWP ACCIDENT MITIGATION CAPABILITY RELATIVE TO PROPOSED NRC CHANGES" The subject report addresses BWR severe accident probabilities and five areas of potential improvement in containment performance. The PRA methodology used by IDCOR in developing the dominant accident sequences and in identifying the plant " damage bins" is very similar to that used in the Shoreham Probabilistic Study, and has been reflected in the IDCOR initial Individual Plant Evaluation for Feach Bottom. In general, we concur with the PRA methodology and the classification of the plant " damage bins" used in the IDCOR report. What has not been addressed are the relatively large uncertainties
- in such assessments, the costs' of improvements compared to potential risks averted, and analysis of costs that could be avoided by closure on any further need to consider mitigation (subject to no plant specific identification of significantly different core melt probabilities).
Each of the five potential areas of containment performance are discussed below:
Combustible Gas Control - Typically, Mark I and II plants are allowed by Technical Specifications to be deinerted for 24 hcurs at startup and 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> prior to shutdown. The staff has proposed the possibility of minimizing the deinerting. Based on the evaluation by IDCOR regarding this initiative, the industry recognizes the increased safety margin of reducing the deinerting window. Any plant operation or safety impacts associated with shorter deinerting times were not identified.
It would appear that plant specific or further generic assessments need to be i performed to determine the feasibility of shorter deinerting intervals. This issue shculd be explored further in order to address any increased assurance that the presence of a combustible mixture inside primary containment is reduced to as low as practicable levels. To this end, future study should include an evaluation of past practice to determine the impacts on plant
. operations and safety due to a shorter allowable deinerting period.
3 1
Filtered Venting - The IDCOR study considered a spectrum of accidents and identified the range of the venting sizing requirements to be between 6 and 36 inch equivalent pipe diameters. An ATWS was identified as the most severe event, requiring a vent size up to 36 inches in equivalent pipe diameter. The study that compares vent size guidelines to individual plant venting capabilities was not provided. However, it appears from an industry a
survey conducted by IOCOR that many plants do not have the large diameter capability needed for an ATWS event, but do have hardware generally available
- *The staff evaluation of the Vermont Yankee Containment Safety Study identified ,
- several uncertainties inherent in methodologies used in the probability calculations (Letter dated October 24, 1986, from NRC to Vermont Yankee Nuclear Pcwer Corporation).
l
__ ,_, _ . , _ , . . , _ . _ _ _ . - , , ,~ __
. ~ - . ~ -. . . - --. . . -- . - - . __ . _ _ _ -
4 a
- 4 to support venting under many postulated severe accidents. IDCOR suggested that venting for an A1WS be considered as an option. The evaluation mentions that industry is currently studying an alternative mitigating scheme for this -
event which may eliminate the need for larp diameter vents, but provided
- insufficient information to indicate feasibfif ty.
The study identifies the need for vents to be remotely and reliably controlled, and have the abf11ty to be re-closed. A conceptual design that meets all the
- potential venting guidelines is described as requiring two DC-powered, fail closed, motor-operated butterfly valves. However, there is no discussion of the need for a capability of vent signal overriding con' tainment isolation signals, nor was the D. C power source fully identified. An operability reliability guideline of .9 to .95 was identified, but no information was provided that would indicate that any of the schemes studied by IDCOR would achieve the guidelines. IDCOR also indicated that valve opening should be achieved at pressures of 76 psi or less, but provided insufficient basis for the guideline, and did not address reciosing.
IDCOR identified the need to ensure that SRV operability should be a principal goal for accident mitigation. Such considerations suggested were identified as setting intial vent initiation at I to 1.5 tines the design pressure, and <
vent termination at 10 to 20 psi belcw the initiation pressure. IDCOR also suggest venting through the SGTS (presumably to insure an elevated release) or, in any case, to outside the reactor building. While we agree with these conclusions conceptually, IDCOR did not supply sufficient infonnation to indicate bases, nor how a utility would convert the guidance into numerical design and performance criteria, i
One aspect of the IDCOR study is that it depicts containment venting consecuences in the light of competing risks. For example, venting strategies intended to avert catastrophic containment failure under severe accident conditions are perceived to lead to a higher frequency of noble gas releases.
Specifically, the study points to cases where containment pressures may rise above design basis levels, but stop short of causing a containment failure.
Venting in these cases-is perceived to lead to noble gas releases, even thcugh the containment survives the pressure transient. However, the study does not quantify the frequency of such releases and reconnends that an additional evaluation be made. Another example of competing risks brought out in the study is the possibility of induced core melt due to the loss of water in the suppression pool in the context of an ATWS event. Aside from the concerns regarding competing risks, the study notes that certain venting strategies can lead to undesirable onsite consequences, even when containment failure is I averted by venting. Specifically, the study concludes that venting can lead to l in-plant contamination and inaccessibility to critical plant safety areas and 1 equipment which are vital for post-accident functions and control. !
The staff initiatives are focused on the reduction of containment vulnerability through debris control; enhanced temperature, combustible gas and source
' tem attenuation; and improved procedures and training. That this initiative can lead to competing risks is a possibility also recognized by the staff. The j staff has concluded however, that the effect of possible increases in the i
higher frequency of noble gas releases caused by venting is offset significantly s
l
. . _ _ _ _ _ , _ _ _ _ - , ,_m... , , , , . . .,. . , , , , - , _ - , . _ . _ , m , _ . . _ . _ _ . . _ , . . _ , . , _ _ _ _ _ _ . , , , , , _
by the dose and econcmic consequence reductions afforded by containment venting.
Moreover, the costs of further plant-by-plant studies of containment perfomance and risk and potential mitigation improvement evaluations, could be eliminated.
With respect to the possibility of plant contamination, the staff agrees that venting has the potential for making parts of the plant inaccessible under severe accident conditions. However, compensating measures involving hardware and procedure modifications (including "hard pipes" to the plant stack) can be made to mitigate these consequences at what appears to be moderate costs.
Drywell Sprays - IDCOR identified four objectives for drywell sprays; 1) to lower containment pressures, 2) to cool vulnerable equipment, 3) to quench debris and, 4) to scrub aerosols. The staff agrees, and suggests not only would aerosols be scrubbed, but so would certain volatile fission products (particularly if good drywell spray coverage is assumed). IDCOR also indicated that such sprays could orovide other benefits in the fann of somewhat 1cwer primary system pressures, inhibited fission product revolatilization, and as a
" steam generator" mechanism for pushing heat and contaminants into the suppression pool. The staff agrees.
The IDCOR evaluation describes a method for spraying during severe accidents associated with A. C. blackouts by using an onsite diesel fire pump to supply the containment sprays at a flow rate well below existing design levels. What was not fully discussed by IDCOR was whether such a relatively low flow rate can be expected to achieve all the goals for spraying. The staff concludes that additional power and water sources shculd be considered to ensure all the goals are met. Cooling and scrubbing with the suppression pool could require an emergency A. C. power source for the systems necessary to cool the pool for station blackouts. The staff suggests that additicnal emergency power sources, such as commercial grade portable generators, and local power circuit modifications be considered for such situations.
The study coints out problems related to creating a significant negative pressure cifferential on the wall between the drywell and wetwell that might cause a failure. IDCOR indicated this concern arose frcm a conservative evaluation of such conditions. The staff concludes that realistic calculations may not indicate such a failure at the containment pressure conditions at which such sprays may be initiated. If failure was indicated, IDCOR described various A. C. powered pumping configurations for supplying the containment sprays at the rated flow rate, but recomended further study of the potential problems associated with spray initiation and the rapidly depressurization of the drywell. Rapid depressurization of the containment may be avoided in the staff's view by modifying the plant emergency procedures and spray system design to specify acceptable initiating pressures, variable initial spray rates, or a combination.
Core Debris Barriers - The use of core debris barriers to prevent molten core debris (including steel from reactor internals and the lower head) from penetrating the steel centainment shell or torus vent pipes, or to provide a water cover for molten material which reaches the torus room floor, were l discussed by IDCOR. The IDCOR report disnisses the torus room barrier with the I simple statement "It does not appear to have significant containment related i benefits to justify further consideration". Drywell barriers to prevent attack l cn the shell and torus downcomers are discussed in some det&il. The impression ;
i l
i
given by IDCOR, however, is that realistic analyses would show that debris barriers are not needed.
The IDCOR analyses performed to date have shewn that at the time the lower portion of the core melts and slumps into the lower plenum of the reactor pressure vessel, approximately 20-30% of the core is molten. This means that at vessel failure approximately 20-30% of the original core inventory could flow out into the containment. This sequence progression could be followed by melting of the retraining core over the next 3-5 hours. This core melt progression differs from the 60-805 instantaneous melt discharge used in a Brookhaven (BNL) calculation, and has a very strong influence on drywell shell failure. Furthermore, after the core material is expelled from the vessel, IDCOR suggests the wcter remaining in the 1cwer plenum will flow out and cool the core debris. The correct debris temperature history would, therefore, be one in which the initially expelled debris mass (including related steel) is cooled by the lower plenum water and then slowly heats up as the water is boiled away. This scenario influences the calculations performed by Brookhaven and would tend to result in lower estimated and, therefore, more coolable debris temperatures. As shown by BNL, lower debris temperatures tend to result in no shell failure. The IDCOR analyses also suggest that the BNL assumptions on debris mass and temperature for the drywell shell failure analyses are overly conservative and should be replaced with more realistic boundary conditions.
The staff concludes that both the IDCOR and BNL analyses reflect the range of some uncertainties associated with contemporary severe accident assessments. It is just these types of uncertainties that indicate that debris control may be necessary.
The IDCOR report also identified some negative impacts of debris barriers as follows:
o Personnel (ALAPA) issues related to installation and maintenance.
o Prevention of the collection of normal leakage and it's detection (Technical Specification Recuirement).
o Seismic related issues regarding the irrpact of the barriers on other safety related equipment for deterministic evaluations required by existing regulatory guidance.
o Potential interference with the assumed downcomer flow area in LOCA assessments.
o Disintegration of the barrier during other severe accidents (seismic, LOCA, etc.) which may cause the barrier to be blown into the torus.
The IDCOR report gave no indication, however, of attempts to identify design options which would eliminate these negative impacts.
t
{
/ ,
Lastly, the IDCOR report alludes to bases which would indicate that debris barriers are not needed. Although many of their arguments appear plausf ait, they have not been given substance in the IDCOR report.
Severe Accident Procedures And Training - Although the BWR Owners' Group Emergency Procedure Guidelines (EPGs) are not designed with the specific intent of providing guidance to the operator for an event which has progressed to the point of vessel failure, the procedural guidance prior to vessel failure is thought to be also appropriate folicwing vessel failure. Therefore, we conclude that implementation of Revision 4 to the EPGs should be expedited. The IDCOR report is sfient on expedited implementation of the EPGs.
A staff review of Revision 4 to the BWR Emergency Procedure Guidelines is underway. The staff expectation is that substantial improvements in Revision 4 will result in several areas. For example, even though it is generally agreed that durirg an ATWS event, operators need to reduce the reactor water level in order to reduce the core power, it is not clear at what level the reactor water should be maintained and what would be the corresponding core power. It appears that there presently exists large uncertainty associated with estimating the core pcwer for a given reactor Wdter level. For example, Option B among the containment venting options proposed by IDCOR, calls for a wetwell vent (6 - 18 inches) capable of removing approximately 6-8% power associated with controlling reactor water below the top of the active fuel (TAF) for ATWS related events.
Option C calls for a wetwell vent (36") capable of removing 18-30% power associated with contrc111ng reactor water level near the TAF for ATWS related events.
-_ _