Methodology for the evaluation of the reliability of passive systems

Within the co-operation between ENEA and University of Pisa (Contract No. 9840, serie 3A, signed in 1998), a synergic study including applications of PSA (Probabilistic Safety Assessment) and thermohydraulics to the design of new reactor concept, is foreseen. An application-oriented project was proposed by ENEA. This is identified hereafter as PSA-TH Project. In this framework a working group has been set-up and supported by ENEA, involving scientific personnel at ENEA, at Polytechnic of Milan (Department of Nuclear Engineering – CESNEF) and at University of Pisa (Department of Mechanical, Nuclear and Production Engineering - DIMNP). Specialists in safety assessment and in thermohydraulics are part of the working group. Following the first meeting ‘exploratory activities’ have been conducted by each of the three branches (i.e. ENEA, CESNEF and DIMNP) of the working group, with the aim of identifying a common way for the study and of finalizing the PSA-TH Project proposed by the ENEA. The ‘exploratory activity’ performed at DIMNP is documented in ref. [1] (see also App. 4 of the present report). At the end of the year 1999, a draft procedure was agreed within the group. The procedure and the results from its application are discussed in the report. The attention was focused toward passive systems. The general objective of the activity was to characterize the transient thermalhydraulic performance of the passive system in probabilistic terms. A natural circulation loop including an Isolation Condenser and the Reactor Pressure Vessel where power production occurs, has been selected for the analysis. The following steps of the procedure can be distinguished: (a) Characterization of design/operational status for the system, (b) Characterization of parameters that are critical for the operation of the system, (c) Assigning probability values to the status and to the parameters from steps (a) and (b), (d) Modeling of the system by a qualified thermal-hydraulic system code, (e) Performing a best estimate calculation and assigning failure criteria for the system performance, (f) Deterministic selection of sets of boundary and initial conditions (6 sets), (g) Statistic selection of sets of boundary and initial conditions – discrete probability distribution (69 sets), (h) Statistic selection of sets of boundary and initial conditions – continuous probability distribution (69 sets), (i) Association of each set to a code run (each set is also characterized by a probability value) and execution of 144 code runs (6+69+69), (j) Based upon failure criteria defined under item (e) and upon results of code calculations, item (i), system performance indicators could be derived (these consisted of ensembles of tables and plots). The deterministic selection, item (f), combined with each of the statistic selections, items (g) or (h), causes the achievement of two ensembles of system performance indicators. ‘Curves of merit’ for the system performance are significant results achieved from the application of the procedure. These are assumed to be characteristic of the selected thermalhydraulic system and can be used to judge the system acceptability and to compare the selected system with different systems