数据驱动下孪生贝叶斯理论非齐次泊松过程的结构损伤评估方法

A damage assessment approach for structure by twin Bayesian theory for non-homogeneous Poisson process is proposed, in which the damage for structure in the crack propagation is selected as the object. The prior parameter distribution for different damage situation is obtained by combine the reliability sequencing strategy based on crack tip field energy with the non-homogeneous Poisson model, and having an effective integration on the test information with the gradual relation by using Bayesian method. At the same time, the posterior distribution calculation approach based on Bayesian is achieved by using the prior information, based on the progressive factor of crack tip field energy and the likelihood function. Furthermore, the typical structure is selected as a case study. The evaluation results are compared with the test results. It can be clearly found from the results that the average accuracy of the proposed approach is 92.1%, and this approach can be used to complete the purpose of damage assessment by using a small amount of test information.