Containment vessels, which contain any radioactive materials that would be released from the primary system in an accident, are the last barrier between the environment and the nuclear steam supply system in nuclear power plants. Assessing the probability of failure for the containment building is essential to level 2 PSA studies of nuclear power plants. Degradation of containment vessels of some nuclear power plants has been observed in many countries, so it is important to study how the corrosion has adverse effects on the capacity of containment vessels. Conventionally, the reliability analysis of containment vessels can be conducted by using Monte Carlo Simulation (MCS) or Latin Hypercube Sampling (LHS) with the deterministic finite element analysis. In this paper, a 3D finite element model of an AP1000 steel containment vessel is constructed using the general-purpose nonlinear finite element analysis program ABAQUS. Then the finite element reliability method (FERM) based on the first order reliability method (FORM) is applied to analyze the reliability of the steel containment vessel, which is implemented by combining ABAQUS and MATLAB software platforms. The reliability and sensitivity indices of steel containment vessels under internal pressure with and without corrosion damage are obtained and compared. It is found that the FERM-based procedure is very efficient to analyze reliability and sensitivity of nuclear power plant structures.
HUMAN RELIABILITY ANALYSIS AT NUCLEAR POWER PLANTS
