基于铣削力仿真样本和降维分类算法的刀具状态监测方法

Tool Condition Monitoring Method Based on Simulation Samples of Milling Force and Algorithms of Dimensionality Reduction and Classification

针对铣削加工过程中刀具磨损造成的主切削力增大现象,提出一种基于计算仿真、降维和分类模型的铣削力判断方法。该方法基于瞬时切厚推导铣削力计算公式,模拟铣削加工过程的时域切削力。依据刀具磨损和切削力增大的关系,生成模拟样本。对生成的样本进行主成分分析(Principal Component Analysis,PCA)和核支持向量机(kernel-based Support Vector Machine,k SVM)分类训练。试验表明,该方法对铣削力样本具有很强的辨认能力,成本较低,可以应用到刀具磨损判断领域。

According to the increase of the cutting force caused by tool wear in milling process,a judging method on milling force in time domain is brought up,which is based on calculation,dimensionality reduction and categorization model. This method simulates the cutting force in the milling process by deducing the formula of milling force based on undeformed chip thickness. Principal component analysis（ PCA） and kernel-based support vector machine（ k SVM） classifier are used on those generated samples. The experiment results show that this method works well on distinguishing milling force samples and it could be applied to the field of tool condition monitoring in consideration of its good economy and generalization ability.