高斯过程的车床刀具磨损预测研究

Gaussian Process of Lathe Tool Wear Prediction Research

为了保证产品质量,并能及时、准确、有效的更换刀具,提出利用高斯过程建立模型并对刀具磨损程度进行预测。首先利用Deform软件仿真车床刀具切削过程,建立刀具磨损随时间变化而变化的样本,然后利用该样本建立高斯过程的刀具磨损预测模型。最后进行刀具实际切削实验,利用测量工具测量刀具磨损量,并建立刀具随时间变化的实际磨损样本,利用实际实验数据对预测值进行验证。数据分析结果表明:预测模型可以有效地学习并预测刀具磨损中的非线性关系,而且刀具磨损的预测精度较高。因此在预测刀具磨损程度时,该模型可以作为重要的预测手段。

In order to ensure the quality of products and replace cutters timely, accurately and effectively, thearticleput forward toestablish model based on the process of Gaussian, which could predict the different degrees of cutting-tool wear. First, using the Deibrms of twaresimulated the cutting process of lathecutter and built the specimens of cutting-tool wearthat varied with time. Then, applying these specimens establish edtheassessment model of cutting-tool wear. Finally, the actual cutting experiment was carried out to measure the level of cutting-toal wear and establish practical samples. Experimental data was used to validate the predicted value of the model. The result indicated that prediction model could effectivelyemulate and prediction-linear relationship of cutting-tool wear. In addition, the accuracy of prediction was extremely high. Therefore, this model could be employed as a significant prediction method for cutting-tool wear.