Density Topology Optimization Method Based on Neural Network

A neural network used for structural topology optimization is designed and built in this paper. In the proposed method, density-based SIMP method first applied to obtain data of three different mechanisms with specified design conditions. Image processing screening methods are raw material learning training process. Then, a according funnel-shaped structure, whose tasks optimize purpose outputting model, lightweight algorithm improve generalization effectiveness network. The verified through classic minimum compliance problems.