基于BP神经网络的破碎机主轴系统参数优化

为了优化破碎机主轴系统的主要工作和结构参数,采用离散变量型正交试验法得到部分参数水平组合,并利用有限元分析软件ANSYS进行分析,得到在这些组合下主轴系统各个性能指数,应用层次分析法和综合评价法得到了主轴系统的综合性能指数,用以上数据训练反向传播(BP)神经网络,并利用完成训练的BP神经网络,预测在其余参数水平组合下的主轴综合性能数指数。根据主轴系统性能指标,优选出利于提高破碎机主轴系统的综合性能的结构和工作参数水平组合,实现废金属破碎机主轴系统参数优化。

Study of Optimization Method for Parameters of the Crusher Spindle System Based on BP Neural Network

To optimize spindle system's main working and structural parameters, this paper uses the orthogonal test with discrete variables to obtain part of the combinations of the spindle system's parameters. According to the analysis results of ANSYS the finite element analysis tool, each performance index of spindle system based on the previous combination is obtained. After each performance index is gained, the synthetic performance of crusher spindle system is also obtained with the analytic hierarchy process and comprehensive evaluation method. Then a BP(Back Propagation)neural network is trained based on the above data, which is able to predict the synthetic performance under the rest part of the combinations of working and structural parameters. Upon the synthetic performance, the combination of working and structural parameters is obtained, which is most beneficial to improving the performance, realizing the optimization of the crusher spindle system's parameters.