TNAG-1506-FCO40-2064-Guangdong-nuclear-power-station-project-at-Daya-Bay-safety-c-1986 — Page 146

: ་་ པས་

Commercial in Confidence

1

The following information requirements refer to a proposed analysis of the Daya Bay Nuclear Plant to the extent required to give approx- imate estimates of the probabilities of the more severe nuclear reactor accidents, involving releases of activity from the reactor

core to the environment. Radioactive releases from lesser active

inventories will not be analysed and information on fuel storage, waste storage and disposal, refuelling procedures is not required.

Basis for the Information Request

Although a comprehensive analysis of a proposed nuclear reactor system, along the lines of recent Probabilistic Risk Assessments, is the most enlightening and informative procedure currently available, it is quite expensive and time consuming to perform, costing typically a few million US dollars in professional effort.

As an alternative to the high cost state of the art PRA, SRD has developed rapid but crude) assessment methods, designed to provide preliminary assessments on short timescales (months) and at a small fraction (circa one percent) of the price of a full PRA. These methods make use of generic system reliability information, coupled with expert judgement of specific plant features, to assess probabilities of typical accident sequences for a nuclear plant of a type which has been previously studied by comprehensive analysis.

In view of the existence of several past PRAS for the generic PWR type on which the Daya Bay plant will be based, the feasibility of the pilot study, at a small fraction of the cost of a normal PRA, is anticipated. The inf- ormation requirements are significantly less than for a full PRA, and should be fulfilled by provision of the necessary documents followed by replies to relatively few supplementary questions, once critical sequences have been identified. However, it must be made clear that the pilot study is not a complete substitute for a proper PRA. It does not provide truly plant specific analysis, but rather a mixture of generic analysis and plant specific judgements, it does not recognise very recent developments which have not yet been manifested in documented PRAS, and it does not provide an effective mechanism for detecting any quirks in a system which might have an unbalancing effect on a plant's nuclear safety strategy.

These are matters which can only be elicited by comprehensive analysis.

Despite these reservations, the pilot study method should provide a cost- effective assessment, appropriate to analysis of a plant at the design stage rather than at the operating stage. Together with general reviews of the state of the art, and a proper appreciation of the uncertainties involved in nuclear risk analysis, it should enable the prospective safety of a nuclear plant to be appreciated.

The information requests which follow are based on the requirements for implementing a pilot risk study of the type which SRD has developed and implemented on other projects.

i