Truss structural damage identification based on Bayesian probabilistic inference

A truss is a common structure in structural engineering, but it will inevitably be damaged when used. It is necessary to establish a reasonable and accurate structural damage identification method to effectively detect the damage and ensure the safe state of the structure. This paper uses a Monte Carlo sampling algorithm for modeling programming based on Bayesian theory for damage identification of statically indeterminate truss structures. By comparing different loading schemes and monitoring node measurement schemes, this paper derives the law of truss optimization. It obtains the optimal posterior distribution of damaged members to find the optimal identification scheme. In addition, using the displacement calculation formula of the truss, this paper gives the optimization direction of the damage identification monitoring scheme. The results show that when the axial force of the damaged member is larger than that of other members, the updating effect of the damaged member is better . And the feasibility of the method is verified by using the actual displacement value and damage value. The method proposed in this paper can effectively update the damage values of bars using known displacement information. However, the method in this paper also has the deficiency of not considering the iterative use of data, and it needs to be improved with the idea of combining the information.