Defect Assessment Methods For Reactor Pressure Vessel Based On Probabilistic Analysis

As one of the important parts containing pressure in nuclear power plant, the safety of reactor pressure vessel (RPV) should be ensured during its lifecycle, to prevent radioactive product leaking. However, some abnormal conditions will cause RPV suffering pressurized thermal shock (PTS) transients, during reactor operation. If the RPV material has enough toughness, these PTS transients will not cause crack initiation or RPV failure. But the toughness of RPV material will decrease with the reactor operation, caused by neutron irradiation. In this case, some severe PTS transients will bring about crack initiation and propagation, and become a penetrated crack. Therefore, it is necessary to assess these defects for the safety operation of nuclear power plant.The safety goals defined with core damage frequency and large early release frequency of fission products at operating nuclear plants are admitted by the world. However, the current defect assessment methods are usually recognized as a deterministic fracture mechanics method with safety factors. These methods can not give the failure probability directly, to correspond with the safety goals. But the probabilistic mehod needs a special program and many input data, which is difficult to be used in engineering assessment at the present stage. Therefore, based on the analysis of both deterministic and probabilistic method, the relationship is investigated and some results are obtained. The main work and achievements are shown as follow.1) According to the elastic fracture mechanics, the relationship between the signle safety factor of deterministic method and the failure probability of probabilistic method is established on the probability sufficiency factor. At the same time, a concept with critical failure probability is proposed to calculate the failure probability under the deterministic acceptance criteria with safety factors. The acceptance criterion of ASME section XI is analyzed. The result shows that under the assumption, the critical failure probability can not meet the safety goals.2) Based on the lognormal distribution, the computational expressions for partial safety factors are established, with the coefficient of variation for crack depth, fracture toughness and load. At the same time, based on the simplified probability method, a reliability calibration method for partial safety factors used in the integrity assessment for the structures with defects is proposed, which can be used to draw up the partial safety factors used in the standards or rules. For the assessment of RPV with defects, many uncertainties are discussed. A guide to partial safety factors is proposed and verified for different conditions with different safty goals.3) The estimation of characteristic values of fracture toughness is evaluated with probability statistics technique. The interpretation of fracture toughness data with the minimum of three equivalent concept is given in the statistical sense. A case result shows that the minimum of three equivalent will have 18.5% likelihood that over valuated the characteristic values, than ASME lower boundary for the sample number not exceed 15. But, the tendency will drop with the increment of sample number. The estimation procedure of characteristic values of fracture toughness is proposed for given confidence level and probability, based on the distribution type and master curve, respectively. The requirement of sample number is proposed. It is suitable for Weibull distribution, when the sample number is not larger than 6. But lognormal distribution is suggested when the sample number is larger than 6.4) The deterministic and probabilistic analyses are performed for the integrity of RPV with defects under PTS. According to the deterministic analysis, the allowable reference temperature for a preset crack. The critical crack size and rapid fracture behavior are investigated for a supposed PTS trainsient. The prototype software for RPV PTS probabilistic analysis is developed. The probabilistic analysis is performed for a typical RPV under PTS. The results show that the failure probability of the RPV is lower than the nuclear safety requirement for 60 years.However, it is only the assumed conditions and a discussion of analysis methods, because of the lack of actual data. Further verification need to be investigated. It is necessary to develop our application database for the development of nuclear structure reliability.