龙门机床进给系统热误差模型关键点优化

Optimization of thermal error model critical point for gantry machine tool feeding system

为了解决温度测点数量和测点间复共线性对龙门机床热误差模型精度和鲁棒性的影响,提出了一种测点综合优化方法。首先,采用模糊聚类法和相关性分析筛选出对热误差影响较大的关键点,达到优化测点数量的目的;然后采用特征提取法得到预测模型的自变量,特征提取消除了复共线性对预测模型精度和鲁棒性的影响。在一台龙门精密镗铣加工中心上实验验证,结果表明:所提出的测点综合优化法优于自适应模糊聚类热关键点优化,采用该方法进行误差预测模型关键点优化可有效提高模型精度和鲁棒性。

In order to deal with the effect of the number of temperature test points and multi-collinearity among the test points on the thermal error model accuracy and robustness of gantry machine tool, a test point synthesis optimization method is proposed. First of all, the fuzzy clustering and correlation analysis methods are adopted to screen out the critical points that have great effect on the thermal error and the purpose of optimizing the number of the test points is achieved. Then, the feature extraction method is used to obtain the independent variables of the prediction model, which eliminates the effect of muhi-collinearity on the model accuracy and robustness. An experiment was conducted on a gantry type precision boring milling machining center, and experiment result indicates that the proposed test point synthesis optimization method is superior to the thermal critical point optimization method based on adaptive fuzzy clustering. Adopting this model to carry out the critical point optimization of the error prediction model can effectively improve the robustness and accuracy of the model.