多裂纹梁不确定性损伤识别和实验研究  

Uncertain Damage Detection and Experimental Study of Multiple Cracks in Beams

作　　者：梁岗[1] 史单艳 沈奎双 Liang Gang;Shi Danyan;Shen Kuishuang(Logistics Engineering College,Shanghai Maritime University,Shanghai 201306,China;SMDERI,Shanghai 201108,China)

机构地区：[1]上海海事大学物流工程学院,上海201306 [2]中国船舶重工集团公司第七一一研究所,上海201108

出　　处：《机械科学与技术》2020年第9期1335-1345,共11页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家高技术研究发展计划(863计划)项目(2005AA11Z123);上海市重点学科建设项目(S30602)资助。

摘　　要：对于多裂纹梁识别过程中的测试数据的不确定性问题,基于传递矩阵法推导了悬臂裂纹梁的频率特征方程,构建损伤参数(损伤位置和损伤深度)和动态特征参量(频率和模态)的映射关系。基于贝叶斯推断理论建立了损伤参数识别的贝叶斯模型,选择均匀分布作为损伤参数的先验分布,以悬臂裂纹梁固有频率和模态振型的测试值作为样本信息,建立损伤参数的后验概率分布。采用马尔科夫链蒙特卡洛模拟方法来解决贝叶斯方程中存在的高维积分的问题,通过仿真算例分析和实验研究,实现了对多裂纹梁损伤参数的有效识别。For the uncertainty of test data in the identification process of multiple cracks in beams,the frequency characteristic equation of a cantilever cracked beam is deduced based on transfer matrix method,and the mapping relationship between damage parameters(damage location and depth)and dynamic characteristic parameters(frequency and mode)is constructed.Based on Bayesian inference theory,a Bayesian model for damage parameter identification is established.The uniform distribution is chosen as the prior distribution of damage parameters.The posterior probability distribution of damage parameters is established by taking the measured values of natural frequencies and modal modes of the cantilever cracked beam as sample information.The markov chain monte carlo simulation method is used to solve the problem of high dimensional integral in Bayesian equation.The effective detection of damage parameters of multiple cracks in beams is achieved through the analysis of simulation examples and experimental research.

关 键 词：多裂纹 损伤识别 贝叶斯推断 马尔科夫链蒙特卡洛模拟方法 

分 类 号：TH215[机械工程—机械制造及自动化]