基于遗传神经网络的压铸机座板结构优化设计

Structure Optimization of Fixed Die Plate of Locking Mechanism in Die Casting Machine Based on Genetic Algorithm and Neural Network

在对压铸机合模机构进行结构设计时,利用神经网络的非线性映射能力,通过少量样本的有限元分析结果,训练出表述结构参数间函数关系的神经网络模型,然后利用遗传算法的全局寻优性找到神经网络模型表述的目标函数的最优结构参数,从而解决结构优化设计的瓶颈和智能问题,利用这种优化设计策略,设计了压铸机合模机构座板,结果表明了该方法的高效性。

During structure design of fixed die plate of locking mechanism in die casting machine, using its nonlinear mapping ability, a neural network model is trained by small number of samples of finite element analysis results. This neural network model is able to predict the structure results with respect to different structures. Then the genetic algorithm method is applied to find the best structure parameters of goal function described by neural network model. This strategy solves the bottle neck and intelligence problem of structure optimum design. This strategy is applied to design the fixed die plate of locking mechanism in a die cast machine and shows a good efficiency.