基于遗传算法及BP网络的主轴热误差建模研究1.PDFVIP

基于遗传算法及BP 网络的主轴热误差建模研究1 + 马 驰，杨 军， 梅雪松， 赵 亮， 王新孟 （西安交通大学机械制造系统工程国家重点实验室，陕西 西安 710049） 摘要：针对基于多输入多输出(MIMO)反向传播(BP)神经网络的热误差建模方法过度依赖于训练 样本、通用性与收敛性较差的问题，利用灰色聚类分组与相关分析法对温度变量分组并提取热敏感 点，并利用遗传算法(GA)将预测输出与期望输出的误差绝对值和的倒数作为判断隐含层节点数的准 则对MIMO-BP网络的拓扑结构进行优化，设定输出层残差误差限实现了网络阈值与权值的有效优化。 建立了基于MIMOM-BP与GA-BP的主轴轴向热伸长与径向热倾角的热误差模型。以精密坐标镗床主轴为 研究对象，采用五点法对热误差进行测量，验证了测量及建模方法的有效性，且GA-BP模型可实现不 同工况下主轴空间位姿状态的高精度预测，更适合作为热误差补偿模型。 关键词：坐标镗床主轴；热误差；灰色聚类分组；遗传算法；反向传播神经网络 中图分类号：TH161；TG532 文献标识码：A High-speed spindle thermal error modeling based on genetic algorithm and BP neural network + MA Chi, YANG Jun, MEI Xue-song , ZHAO Liang, WANG Xin-meng (State Key Laboratory for Manufacturing Systems Engineering，Xian Jiaotong University ，Xian 710049, China) Abstract：In order to avoid the disadvantages of the thermal error modeling method based on MIMO-BP, including the excessive dependence on the training samples, the poor convergence and worse generality, the gray cluster grouping and correlation analysis were used to group temperature variables and optimize thermal senstive points. Subsequently, a genetic algorithm (GA), which regarded the reciprocal of the absolute value sum of the differences between the predictive and desired outputs as the number of nodes in the hidden layer, was used to optimize the topology of the network. Moreover, the thresholds and weights of the network were optimized by setting the residual error limits of the output layer. Then the elongation and thermal tilt angle models were established based on MIMO-BP and GA-BP modeling. The five-point method was utilized to measure the spindle thermal errors of Jig-boring and the effectiveness of the measurement and modeling was val