不同信息融合方法在结构损伤识别上的应用和分析

APPLICATION AND ANALYSIS OF DIFFERENT INFORMATION FUSION METHODS FOR STRUCTURAL DAMAGE DETECTION

在工程结构的损伤探测领域,不同的信息融合方法和方式对结构的损伤敏感程度以及计算的复杂程度往往不同,而且适用条件也不同。为了解决以上问题,描述了基于结构损伤识别的功能信息融合模型,并在此基础之上采用了多种融合方法进行了数值仿真和分析。数值仿真结果表明,采用了信息融合技术的结构多损伤位置识别,可以产生比单一信息源更精确、更完全的估计和判决,而且不同的信息融合算法的应用往往取决于研究对象和实际条件的要求。

Different information fusion approaches possess different sensitivity and computational complexity in structural damage detection field. In order to solve the problem, a functional fusion model for damage identification is presented. Three main fusion methods, i.e. Bayesian fusion method, Dempster-Shafer evidence theory and fuzzy set theory, which are based on the functional fusion model, are proposed to identify the structural multiple damage sites. Simulation results demonstrate the excellent performance of information fusion technique to identify damage locations as compared with the methods based on single information source. The applicability scope of different fusion approaches is pointed out.