Multiple factors-based damage level assessment method of concrete structures based on evidential reasoning and particle swarm optimization

Abstract

Previous damage prediction methods of concrete structures usually fall into subjective and unreasonable assessments, which are usually limited to the special application scene with complex physical modeling depending on professional knowledges and experience. To overcome the above limitations, a multiple factors-based damage level assessment method of concrete structures is proposed based on evidential reasoning and particle swarm optimization. According to the method, multi-source data can be integrated for damage rules generation and quantitative damage level assessment automatically. The number and types of the used multi-source data can be chosen freely according to the special application scene. Finally, multi-source data including the normal strain in X-direction, normal strain in Y-direction, and the corresponding shear strain of two concrete structures under cyclic loads are utilized to verify the developed method. The results show that the damage level assessment results match well with the corresponding experimental results, which has a high reliability index based on the developed method. The ability of the method with good generality and flexibility is verified, which shows that the method can be used to achieve the reasonable damage level assessment of concrete structures.