Division of Nuclear Power

The overall objective of this collaborative project was to contribute to an international consensus on the definition of the reliability of passive systems that involves natural circulation, and on a methodology to assess this reliability.

RMPS (reliability methods for passive systems) and APSRA (assessment of passive systems reliability) methodologies were used to assess the performance and reliability of passive decay heat removal (DHR) system of the French GFR (gas cooled fast reactor) design for station blackout (SBO) and loss of coolant accident (LOCA) combined with loss of off-site power, respectively. The benchmark was comprised of two different phases: deterministic calculations and reliability calculations.

1. Two different computer codes (CATHARE and RELAP5) were used for the deterministic analyses to evaluate the performance of the DHR for the above transients. The analyses showed that both codes could predict the system performance and failure states satisfactorily. The results were consistent with each other as far as the effects of sensitive parameters on system performance are concerned. However, some differences were observed with regard to values of the output parameters associated with the failure criteria.
2. RMPS was used to calculate the reliability of DHR for SBO and LOCA, combined with loss of off-site power. APSRA was used to calculate the reliability of DHR for SBO condition only.
3. Both methodologies showed that DHR is capable enough to remove the decay heat in case of SBO without any failure. However, to understand the applicability of the methodologies, the failure criteria were made more stringent artificially (maximum allowed gas outlet temperature was reduced to 1050 ^oC instead of 1200 ^oC). With these criteria it was found that DHR has a high probability of failure as predicted by both methodologies.
4. RMPS predicted the probability of failure of the DHR as 0.316 and APSRA predicted a failure frequency of 7.3 x 10-6/ hour for the SBO condition. Since APSRA attributes the process parameters variations to the failure of the components or the systems which control them, the failure probability of the system is predicted to be a failure frequency unlike RMPS method. For the 2nd transient, the failure probability estimated by RMPS was relatively high and equal to 0.49.
5. A common definition of reliability of T-H passive systems was agreed which is the probability that a system will perform its intended function in a satisfactory manner for a given period of time [0, t], when used under specified operating conditions. This definition has been used by both methodologies.
6. Similarities and differences of the two methodologies were identified, and the possibility of unifying the features of the two methodologies in order to develop a generic methodology was also discussed.