基于神经网络和遗传算法的液压机上梁轻量化和刚度优化设计

Lightweight and Stiffness Optimization of a Hydraulic Press′s Upper Beam Based on Neural Networks and Genetic Algorithms

提高刚度和轻量化是液压机设计中重点研究的内容。针对传统设计方法难以解决上梁刚度和轻量化之间的矛盾问题,提出了基于神经网络和遗传算法的液压机上梁轻量化和刚度优化设计方法。在液压机设计过程中,建立了上梁有限元分析的参数化模型。采用正交试验设计安排试验方案,获取试验数据。以试验数据为训练和检测样本,建立了设计参数与刚度和质量目标之间的非线性映射关系的神经网络模型。运用NSGA-Ⅱ遗传进化算法对神经网络模型进行优化,在指定参数区域内找出设计参数的Pareto最优解集。结果表明:该方法对于液压机上梁的多目标优化具有明显的效果。

Good stiffness and light weight are two important goals in the design of hydraulic machine structures. Facing with the contradictory problem between the stiffness and light weight in traditional design method,lightweight and stiffness optimization method of a hydraulic press′s upper beam based on neural networks and genetic algorithms is proposed. The parameterized analysis model of the upper beam is established. Using the method of orthogonal experimental design,test data is obtained. The nonlinear neural network model between the design parameters and the stiffness and quality objectives is established based on the test data. Through optimizing the neural network model by NSGA-Ⅱ genetic algorithms,Pareto optimal results are obtained. The result shows that the stiffness of upper beam is increased and the weight of the upper beam is decreased. This design method is also valuable for other machine′s optimal design.