ML20080B975

Forwards Technical Comments Based on Review of Suppl 1 to Des.Nrc Should Consider Risks Associated W/Range of Events Expected to Occur Offsite.Documents Rated as ER-2

ML20080B975
Person / Time
Site:	Limerick
Issue date:	02/03/1984
From:	Brubaker H ENVIRONMENTAL PROTECTION AGENCY
To:	Martin R Office of Nuclear Reactor Regulation
References
NUDOCS 8402070340
Download: ML20080B975 (5)

Text

_ _ _ _ _ _ _ _ - _ _ _ _ _ .

r .,.'

, e f *%

~ A ,,g I UNITED STATES ENVIRONMENTAL PROTECTION AGENCY REGION lil 6TH AND WALNUT STREETS PHILADELPHI A. PENNSYLVANI A 19106 February 3, 1984 Mr. Robert E. Martin Project Manager Office of Nuclear Reactor Regulation U.S. NRC Washington, D.C. 20555

Dear Mr. Martin:

EPA has completed its review of Supplement No. 1 to the Draft EIS for the Limerick Generating Station. The document represents what appears to be final results of a great deal of mathematical analyses. However, in many places it is difficult to know what is based upon analysis of historical events as they may relate to Limerick, and what is based upon probability theory related and derived from design parameters specific to Limerick.

The concensus here is that NRC looked at the risks associated with the plant itself, but not with the range of events that can be expected to occur off-site. An example is the Even transportation of fuel both to and from the facility.

though this subject has been dealt with generically in another publication, events for this site and its surroundings should be thoroughly analyzed. Perhaps this could be properly dealt with through a comparison with non-nuclear facilities for shear numbers of accidents, and then working out comparative analysis for the gravity of events using parameters related to radioactive materials. In addition, the AMA carried out a comparative study in an attempt to qual 4fy the risks from air pollution from alternate means of power generation. NRC may find this useful.

4

~

8402070340 840203 PDR ADOCK 05000352 D pg

. - - ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ }

e .*

The technical comments are categorized below for your consideration in preparing the final EIS.. It is hoped you will consider them seriously inasmuch as the analysis reflected in the document seems to be of sufficient depth to be acceptable, but at the same time summarizes possible events all too briefly. Because of this, the document is rated ER-2, as was the Draft EIS of which thie risk assessment is a part.

If you have any questions or if we can be of further assistance, please contact Mr. Robert Davis of my staff.

Sinc rely H nry uba Analys s & Services Section l

h

f . .

i Technical Comments

1. Statistics and probability:

In spite of the uses made of information from historical events, it appears that much of the assessment for PRA is design based. As we understand it, this means that plant incorporates certain parameters respresenting upset design'so events that'their severity is minimized. Mathematical models can then be used to analyze in advance for SCRAM or worse, and develop a probability that has resulted in the reassuring scenario presented in this Supplement. It is assumed that designers have availed themselves of the broadest statistical bases possible, but the reader is often confused by the hybridization of probability and operational experience.

It is implied throughout the document that GE has adjusted its design in accord with operational experience, but it is apparent that the statistical evidence for some of these changes is not in hand. The Zircaloy issue mentioned below is an example of where information has been culled out when it probably should not have been.

In other words, terminology may be the culprit, or at least choice of words, because TMI was co accident: it was a design failure regardless of operator error, and looking at it even further it was a designed accident because if the plant had been designed differently the event presumably would never have occurred.

Appendix H relates to accident sequences and release categories that all seem to describe the ultimate cumulative disasters and their probabilities. It appears that uncertainty in the available data allows only a rough estimate of conseuences of events during reactor operation. It this an attempt to say at the same time that worst case analysis and minor events are all lumped together and that the minor events really fall into the same unlikely category, but since their consequences are so insignificant that they need not be analyzed? That is, no great consequence exists if a part of the operation represents a smaller risk than a large event, which is analyzed. However, no mention is made of the possibility that a small event may make an associated small event more likely by defeating a safety system. An example is the limits of temperature

. measurements. Design should include parameter monitors that operate independent of design operational limits, i.e.,

monitors that measure the entire spectrum of each integral part to be monitored. For example, the monitors and gauges related to core temperatures should be able to read possible range of core temperature and not be designed to the limits of plant operation. If this isn't done minor upsets can lead quickly to major ones regardless of the proLability of any one occurring individually. Table H-1 attempts to tabulate all of the

~

initiating events, but does not exhaust the possibilities. Is

.this table representative of the design or experience?

t .- -.

'In sum, we agree with the wording on 5-8, "The evidence of accident' frequency ...'is a useful indicator of ... probabilities and impacts.",.but hasten to add that several different and

evolving design configurations are being used, all of which are closely linked to research and development. This is why we emphasize the importance of the statistics used.

Two last points regarding statistics:. 1) page 5-3 (1st para.)

relates to off-site dispersion of radioactive materials and implies that such events are minimized by wet weather conditions. The failing here is to analyze for tha vnrat case.

- ' for : example, hot dry windy _ conditions. 2) As admitted in the.

report, dosage over time is a very important consideration.

The NRC.has rules covering the control of-extreme dosing, but

.this document dismissed this with a statement that plans to control.this are in the works, and more importantly, that the probability of. heavy dosage are negligible. Setting the mathematicalLaodels aside, EPA-feels that the regulations

~3 should be the issue here, namely adherence to.them. The plans.

to do this'need to be addressed in the Final;EIS.

@ 2. Cladding Technology:

O The_first line of safety in the reactor is the cladding system

^

~ surrounding the uranium and'its fission products. From what we gather this is shelf technology or at least state of the' art,.

^

but the description on page 4-11 of the SER (NUREG 0991) is not ;

. very; reassuring with regard for the documentaion of control technology for waterside corrosion. As we understand it, this

is a' minor _and fairly well recognized problem, but has yet to be completely nailed down with regard for both causes and cures, although operational regulations probably minimize any hazards. The Final EIS. should pay some attention to detailing the status of this problem and current efforts to pin point it ~

Land correct it. Could this, as discussed above, ever exacerbate a minor event such as unexpected shut-down of cooling systems with associated pressure or loss rise?

3. Benefits:

Page-6-3 relates'the benefits to be realized from operation of the two units. Presumably these benefits will accrue to the consumer, although this is not stated in the~ text, and they amount to $34 million (1985 dollars). How is this figure reconciled with.PECO's recent statement that they will petition the Pennsylvania Utility Commission in spring of '84 for a 20%

orate raise to cover costs associated with unit one and for

-another~ rate raise of 20% in '85 to cover the cost of unit 2?

Translated into current dollars, the initial request will amount'to $477 million cost to the consumers. Perhaps NRC should' carry out.an analysis of conventionally fueled power production so that.the benefits claimed will be put into a

'better comparative perspective.

In' addition, the decommissioning costs appear to be underestimated, even though set in a large range. Following the logic associated with the escalating costs of construction, it would seem that decommissioning should follow the same path.

4. Evacuation Plans:

Evacuation plans are mentioned on page 5-12 as iueomplete, but in preparation. Response systems are hardly mentioned, and then in only'a generic way. The Final EIS should present these in full detail, with a full disclosure of how it was related and explained to the subject population prior to Final EIS.

This must be done prior to completing the final EIS because a major objective is to assure the safety in times of accident.

This may have an ancillary positive effect in that people in the area will soon come to realize that the plant operators and NRC really care about the concerns of the local populace.

Miscellaneous '

On page 5-7, it is stated, and presumably attributed to the National Research Council and to Land, that "...(radiation has) been studied more exhaustively than any other environmental contaminent." This kind of majestic statement, regardless of who makes it, is highly controversial and has no place in a document of this kind. Even if it does not over state the case, it says something about a technology that feels compelled to make it.

The fourth paragraph on the same page refers to the questionable fact that "Most authorities agree..." on the statistics regarding health effects of radiation from such plants. It is suggested that NRC rely upon specific citations rather than inexact and unspecified references. On the whole, the document reflects a great deal of highly sophisticated mathematical manipulation, but NRC would do well to recognize

'that "... chaos results when design is assumed to be primarily a problem in mathematics." (E.S.Furguson, Sci. 26, Aug. '77).

Attention should also be given to non-design factors and public awareness.

The review coordinator tried to contact the appropriate people at NRC to discuss some of the points raised in this letter, but was unable to do so in a timely fashion and faced with the deadline. Therefore, some of the points raised here may either have been either alrcady dealt with by the NRC staff or are amenable to easy explanation.

_ _ _ _ _ _ _ _ _ _ _ - _ _ _ - _ - _ _