Availability of the Emergency Core Cooling System of a CANDU Pressurized Heavy-Water Reactor Following a Small Loss-of-Coolant Accident

The availability of the emergency core cooling system (ECCS) as a long-term safety backup system following a small loss-of-coolant accident (LOCA) has been analyzed for the Pickering NGS Unit A, a Canada deuterium uranium (CANDU) type of pressurized heavy-water reactor (PHWR). Fault tree analysis methodology was used to assess the unavailability of the ECCS. The PREP and KITT computer codes were used to estimate the failure probabilities. From these computations, the unavailability of the ECCS to supply sufficient coolant to the core is estimated as 3.63 × 10⁻³. This figure is higher than the failure probability target 3 × 10⁻³ that is specified by the Canadian Atomic Energy Control Board for the safety systems of CANDU PHWRs. It has been found that human error might make a very important contribution to ECCS unavailability, especially if the human error rates have been assigned the upper bound values in the fault tree calculations. That should be the case, therefore, for any fault analysis and reliability assessment of nuclear generating stations. Unlike the case for light water reactors, the ECCS in a CANDU PHWR is not the last defense against the LOCA, because of the availability of quite a large amount of D₂O moderator in the calandria around the pressure tubes.