Abstract
Due to insufficient test data, insufficient constraint equations and uncertain objective function, the local optimal solution and the global optimal solution of the objective function in finite element model updating may represent the actual parameters of the structure. Based on this, this paper proposes an improved artificial fish school algorithm. By combining the niche technology with the artificial fish school algorithm, the improved algorithm can systematically find multiple global optimal solutions and local optimal solutions of the objective function. Aiming at the difficulty of determining the niche radius, an adaptive niche radius mechanism is proposed. The improved algorithm is used to study the multi-alternatives problem of finite element model updating after verifying its feasibility through numerical simulation analysis. In the case of benchmark framework model updating, it is confirmed that multi-alternative problems exist and the global optimal solution of the objective function does not necessarily represent the true parameters of the structure. In case 2, the improved algorithm combined with the Kriging model is applied to the model updating of a cable-stayed footbridge, and 15 sets of solutions are obtained, in which the error objective function values of the measured and theoretical values of the bridge modes are close but the solutions are completely different. Combining with the actual bridge condition and reanalysis technology, the author takes the suboptimal solution 2 as the most representative solution of the bridge parameters, which reduces the possibility of misjudgment of structural parameters.
Highlights
The finite element model is usually established according to the design drawings
In order to obtain the benchmark model, the finite element model updating (FEMU) aimed at reducing the structural theoretical responses and the actual responses error has been extensively studied by scholars at home and abroad
The results showed that the global optimal solution of the objective function was not the real damage of the structure, on the contrary, the local optimal solution was the most representative of the real parameters of the structure
Summary
The finite element model is usually established according to the design drawings. In the process of modeling, there are geometric parameters, physical parameters, boundary conditions, and other assumptions, which make the theoretical values of the model have errors with the actual responses of the structure. Some research results show that the updated finite element model based on global optimal solution still cannot guarantee the agreement with the actual structural response [15,16,17,18,19,20,21,22]. HSJA was applied to the model updating of a cable-stayed bridge, and good results were obtained; Caicedo et al [17] proposed a steady-state genetic algorithm (SSGA) which can find the global and local optimal solutions of multi-peak functions. The “correct solution” may be missed due to the poor performance of the algorithm in finding multiple solutions; in addition, the existing research of MGA focuses on the model updating of laboratory structures, the decision-makers can choose the “right key” from multiple alternatives by preset damage [21,22,23].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
