Reliability Analysis On FPSO Inner Turret Based On Bayesian Theory

Due to set up in the high load and corrosion environment continuously, the material of marine equipment is easily getting degradation, seriously affecting its security and reliability.In the process of the quantitative risk assessment, the probability of the accidents with small samples which is calculated through the traditional statistical methods have a large uncertainty.For this reason, using the bayesian statistical method considering a prior information calculate the probability of oil and gas Leakage of FPSO Inner Turret, reducing the uncertainty caused by less sample size.Firstly, combing the theory involved in the bayesian statistical analysis. The building of bayesian statistical models (including likelihood distribution and prior distribution),the calculation of posterior distribution using MCMC method(including M-H algorithm and Gibbs algorithm),the convergence diagnostic of posterior samples, the theory of model checking(including leave one out cross-validation and posterior predictive densities) and the model comparison (including BIC and AIC criterion) are proposed.Secondly, through establishing FPSO inner turret system FMEA by relevant information, the leakage of oil & gas in inner turret system with a higher risk value is of significance to the security of FPSO system; then collect failure samples?Finally, establish bayesian statistics model for the leakage of oil & gas. According to the corresponding failure mechanism, the likelihood distribution of the two models are Weibull distribution and lognormal distribution; then calculate posterior distribution and compare the models based on BIC and AIC criteria. The reliability indexes for the subsequent risk assessment are calculated on the basis of the optimal model.