基于IAGA-LSSVM的切削加工表面粗糙度的智能预测

Intelligent prediction for surface roughness of cutting based on IAGA-LSSVM

提出了一种基于改进自适应遗传算法与最小二乘支持向量机(IAGA-LSSVM)的切削加工表面粗糙度的智能预测方法。通过设定LS-SVM模型主要参数的取值范围,采用IAGA进行寻优,提高了LS-SVM预测模型的精度。最后采用平均相对预测误差作为检验指标,比较了多元线性回归模型、BP神经网络模型、AGA-LSSVM模型及IAGA-LSSVM模型对表面粗糙度的预测能力。结果表明:IAGA-LSSVM预测模型的建模时间更短,平均相对预测误差更小,对切削加工表面粗糙度的预测具有一定的参考意义。

A new prediction method for surface roughness of milling based on an Improved Adaptive Genetic Algorithm （IAGA） and Least Squares Support Vector Machine （LS-SVM） is put forward. By setting the range of the main parameters of the LS-SVM model, the accuracy of the LS-SVM forecasting model is improved by using IAGA optimizing the parameters. Finally, the prediction accuracy for surface roughness of the multiple linear regression model, BP neural network model, AGA-LSSVM model and IAGA -LSSVM model is compared. The practical experimental results show that the modeling time of IAGALSSVM prediction model is shorter, while the average relative prediction error is smaller, that has a certain guiding significance for the prediction of surface roughness in milling.