Research Of The ESO Detection Algorithm Based On Virtual Material And Its Application In Thin-Walled Structures

Evolutionary structural optimization(ESO) algorithm which based on Finite Element Method(FEM) is one of the most active algorithms on structural topology optimization. On the basis of good operation and simple idea that the optimal structure can be produced by gradually removing elements with low efficiency from initial design domain, ESO algorithm make great strides in the development of both theoretical research and engineering application.ESO is a heuristic algorithm, Rozvany questioned the validity of ESO algorithm by the failure of structural optimization in Tie-beam problem. So far, many researchers has proposed various algorithms for solving Tie-beam problem and revising ESO algorithm.The heuristic detection algorithm using virtual material remedies the drawbacks of ESO algorithm fundamentally through checking the integrity of load path. Further investigation in this paper shows that there is no need to check most of iterations in the optimization procedure using initial ESO. Thus, this paper divides the detection progress using detection algorithm into two steps: Firstly, prejudging the iterations where structural strength or stiffness exists sudden change, then, making heuristic detection just for the abnormal iterations where structural connectivity failure exists. By doing this, the efficiency of heuristic detection algorithm is greatly improved.This paper applies the detection algorithm which is designed for plane problems to shell cases, then obtains the detection criterion and designs the procedure of algorithm.The examples demonstrate that the algorithm proposed successfully tackles the deficiency of failure when optimizing thin-walled structures with initial ESO algorithm.Besides, the author will apply the algorithm proposed to the layout problem of thin-wall structure of garbage container body and obtain the optimal topology of thin-wall structure.In this paper, the improved algorithm proposed can avoid the failure in the topology optimization of thin-wall structure when using initial ESO algorithm and does not affect the ability to exploring optimization. Consequently, the contribution proposed above means certain theoretical and practical value.