Methodological Validation of MERMOS by 160 analyses

Electricite de France has developed MERMOS, a new PHRA method for the probabilistic safety assessment of the N4 1450 MW units, the most recent type of French reactors with a computerised control room. The distinctive feature of Ed's PSAs is that they are carried out on a set of identical units. Two facilities have N4-type units with a computerised control room, Chooz and Civaux, each with two units. A set of validation actions marked the development of the MERMOS method, and its subsequent application in the frame of the N4 PSA in which 160 MERMOS analyses were performed. These validation actions revolve around two axes which are, on the one hand, the theoretic validation of the method, i.e. its scientific validation, and, on the other hand, its industrial validation. No objective and universal criterion is available to this day to assess a PHRA method, and validation must still be performed case by case based on criteria which remain to be defined with respect to the objective of the implementation of the method. This PHRA feature, already highlighted in the conclusions of the benchmark performed at ISPRA in 1989 [12], is still very true today: ' Where the HF-RBE (Human Factor Reliability Benchmark Exercise) failed is in establishing consensus procedures and methods. From the results and discussions it appears that human reliability analysis is still rather an art than a science and it is too early to try to specify preferred ways of performing the analysis...'. (authors)