Bayes方法在水工泄洪闸门事故率分析中的运用

Application of Bayesian approach to accident rate analysis of outlet gates

泄洪闸门事故率λ受设计制造、运行管理等各种复杂的不确定因素影响。随着工程的老化,影响闸门安全运行的各种随机因素的时变特性更为明显。采用Bayes方法,可以利用适时的闸门运行抽样检查结果,修正和改进已积累的原观统计信息,从而将闸门老化的时变特性引入事故率λ的定量分析中,为全面、准确地评估漫坝风险率创造条件。文中根据Bernoulli模型的假设,按“等值样本”概念,认为λ的先验信息等值于来自实际运行过程的样本信息,其分布符合Beta共轭分布簇条件,从而简化了λ后验分布的计算。

The accident rate of outlet gates is affected by the uncertain factors in design, manufacture and Operation, Various stochastic factors influencing the operation security of gates have time-varying characteristics due to ageing of the hydro project, By using the Bayesian approach the results of suitable sampling inspections for the operation of gates can be used to modify and improve the accumulated statistical data obtained from prototype observation in order to introduce the time-varying characteristics of gate ageing into the quantitative analysis of accident rate, which is essential to the determination of the overtopping risk rate of dams. Based on the hypothesis of Bernoulli model and concept of equivalent samples, the prior information of accident rate may regarded as the sample information obtained from engineering practice, Its distribution is accord with that of Beta conjugate function, By this method the computation of the posterior distribution of accident rate can be simplified.