中文摘要：

基于静动力特性改变的结构损伤识别方法在过去几十年中得到了长足发展。相对于确定性的损伤识别，基于概率统计理论框架的损伤识别方法能够更好地反应损伤问题的不确定性本质，是一种非常有发展潜力的损伤识别方法，有望解决当前大多数损伤识别方法对于测量误差和噪声敏感的问题，从统计意义上实现复杂工程结构健康监测早期损伤的诊断。在回顾结构确定性损伤识别方法的基础上，主要介绍不确定性损伤识别方法和有限元模型确认的研究进展，分析比较各种方法的优点和不足之处；最后对有待进一步研究的问题和基于模型确认的概率损伤识别方法发展趋势进行展望。

英文摘要：

Based on changes of static and dynamic structural characteristics, the damage identification methods have been developing rapidly in the past few decades. When compared with deterministic damage identification methods, damage identification methods based on probabilistic statistic theory can better reveal the nondeterministic nature of damage problems. Better resolution of measuring errors and noise problems, to which most of damage identification methods are very sensitive, is expected. Early damage detection in the complicated structures can be carried out using statistical methods. In this paper, deterministic damage identification methods are summarized. The development of probabilistic damage identification methods is described, including the finite element model validation. Their advantages and shortcomings of these damage identification methods are analyzed. Topics worthy of further study and the trend for probabilistic damage identification based on the FE model validation are presented