The Research On The Identification Of Damage Degree For Structural Members And The Location Of Joint Damage Of Long-Span Spatial Lattice Structures

Structural damage often leads to heavy accident and results in the loss of life and property; therefore, there is important theoretic and practical significance about the research on the damage identification.The detection, location of structural damage and the estimated residual life of the damaged structures have been the subject of much research in recent years. The technique of the damage identification has been widely applied in the field of civil engineering and astronautics and so on.Once the structure is damaged, it is certain that the damage can be expressed completely by the change of modal parameters. So the structural damage can be recognized and identified based on the change of the structural modal parameters. In this paper, the damage degree of the member in the long-span spatial lattice structures is identified using the method of neural network on the assumption that the structural damage has been located accurately. The input of neural network is the combined parameter composed of the location of damage and the rate of the modal strain energy. The severity of damage is estimated using the trained BP neural network through a numerical simulation study on Tianjin Olympic Stadium. It is found the damage degree can be reasonably estimated when the modal data is not severely corrupted with noise.Also, the research in this paper deals with the location of joint damage in long span space structure, and a two-step method is proposed to locate the joint damage. At the first step, the damaged substructure is detected employing the rate of the mode curvature as damage index, and then through analyzing the substructure the location of joint damage is accurately identified using the rate of element end strain mode. In this paper, the research is focused on the location of joint damage using the method mentioned before through a study model of Tianjin Olympic Stadium. It is proved that the method proposed is reliable for the detection of joint damage of long-span spatial lattice structures.