基于参数识别的结构损伤概率诊断方法

A STATISTICAL DAMAGE ASSESSMENT METHOD BASED ON PARAMETER ESTIMATION

为解决观测噪声等不确定性因素影响下确定性损伤识别方法的损伤误判问题,提出了结构损伤识别的概率可靠度法。通过正则化修正改善参数识别问题的不适定性,利用数据摄动和蒙特卡罗法研究观测噪声对参数识别结果的影响并获得各待识别参数的统计特征,在此基础上基于概率统计理论推导了损伤概率的计算公式。根据损伤识别试验验证了方法的有效性和准确性,分析并比较了概率可靠度法和确定性方法的损伤识别结果。试验研究表明,损伤识别的概率可靠度法能够显著改善传统确定性损伤识别方法的损伤误判问题。

To avoid the damage misidentification of traditional damage assessment methods induced by indeterminate factors such as measurement noise, a statistical damage assessment method is proposed. Regularization technique is introduced to alleviate the ill-posedness of the parameter estimation problem. A data perturbation method is employed to study the impact of measurement noise on identification results. The probability nature of estimated parameters is obtained by Monte Carlo method. The formula for statistical damage assessment is then established based on probabilistic method. The validity and accuracy of the proposed method were demonstrated through damage detection experiment and the damage detection results of both statistical and determinate assessment methods were compared. Experimental results show that the statistical damage assessment algorithm can improve damage misidentification notably in contrast with determinate method.