基于贝叶斯理论的结构件健康状态评估方法研究

Approach for structural health assessment based on the Bayesian theory

以结构件裂纹扩展过程中的健康状态评估为研究对象,针对结构件健康状态评估过程中广泛存在的Weibull型数据的特征,通过将Weibull型分布数据类型与贝叶斯理论相结合来提出了一种基于贝叶斯理论的结构件健康状态评估模型,并通过采用MCMC算法对复杂后验量的计算问题进行了求解。运用声发射手段对实例构件的健康状态数据进行了采集,基于贝叶斯理论的结构件健康状态评估模型对实例构件的健康状态进行了评估,评估结果表明5%与95%两个分位数构成了结构件健康状态评估参数的90%置信区间,后验参数的分布集中,且置信度达到90%,可以实现结构件健康状态的精确评估。

The structural health assessment in the crack propagation process was studied.Considering the wide existence of Weilbull distribution data characteristics in the assessment process,an approach for structural health assessment based on the Bayesian theory was proposed through combining the data of the type of Weibull distribution with the Bayesian theory.The numerical solutions for the complex posterior problems were provided by using the MCMC algorithm.The acoustic emission approach was adopted to collect the health assessment data for an industrial component.The health result obtained by the proposed model shows that two fractile parameters of 5%and 95%constitute the 90%of confidence interval of the assessment parameter for structural health.The posterior parameters are centrally distributed and the confidence level is up to 90%,which verifies the accuracy of the health assessment method.