An Effective Damage Identification Method for Large Truss Structures Using Successive Iteration Method of Bounding Domain and Domain Decomposition Finite Element Technique

Abstract

Damage identification methods using finite element methods have received much attention for health monitoring systems. However, the damage identification methods need a long computational time for large and complex structures. In this paper, we develop an effective damage identification method for large truss structures using a successive iteration method of bounding domain and a domain decomposition finite element technique. The subdomain of truss structures which contains a damaged truss member is first determined by a flexibility method. The subdomain is subdivided into smaller subdomains using a successive iteration method of bounding domain, and finally the damaged truss member is identified. In order to reduce computational time, the domain decomposition finite element technique is used to get frequencies and mode shapes of the whole structure by assembling subdomain frequencies and mode shapes. To examine the effectiveness of the present damage identification method, the damage identification for large two-dimensional and three-dimensional truss structures is conducted. As the results, it is found that the damaged truss member and its residual stiffness can be identified precisely and the present damage identification method can reduce the computational time when the measurement errors are relatively small.