Rapid calculation method for Bayesian update of logarithmic normal distribution of nuclear power plant

Abstract

The invention belongs to the field of nuclear power plant equipment reliability, and particularly discloses a rapid calculation method for Bayesian update of logarithmic normal distribution of a nuclear power plant, which comprises the following steps: S1, according to a prior distribution function of logarithmic normal distribution equipment failure and actual power plant failure sample data, and according to a Bayesian formula and a lambda probability density function of posteriori distribution; s2, calculating an integral interval; s3, caculating integral terms in the posterior distribution function, so that the posterior distribution function without the integral terms is obtained; and S4, fitting posterior distribution parameters. Compared with a conversion method commonly used in engineering, the method has the advantages that conversion errors caused by conversion from logarithmic normal distribution to conjugate distribution are avoided, and the calculation accuracy is improved; meanwhile, the numerical integration calculation amount is greatly simplified by estimating a posterior distribution function integration interval, and the calculation speed and stability are improved by adopting an anti-calculation overflow processing method.

Description

technical field [0001] The invention relates to the field of nuclear power plant equipment reliability, in particular to a fast calculation method for nuclear power plant logarithmic normal distribution Bayesian update. Background technique [0002] As the basis of the probabilistic safety assessment (PSA), the reliability data of nuclear power plant equipment directly affects the quantitative analysis results of PSA in nuclear power plants. For equipment reliability data, the PSA in the power plant design stage is generally directly obtained from the industry's general database, while the PSA in the power plant operation stage needs to be updated in combination with the actual operation of the power plant. [0003] The update of reliability data generally adopts the Bayesian method, that is, the data before the update is used as the prior distribution, and the Bayesian update is performed according to the actual operation failure data to obtain the posterior distribution. ...

AI Technical Summary

Problems solved by technology

Although the calculation of the conversion method is simple, the conversion process will introduce a large error

Images