基于混沌时序分析方法与支持向量机的刀具磨损状态识别

Tool wear state recognition based on chaotic time series analysis and support vector machine

为了表征、获取与识别刀具的磨损状态,提出一种基于混沌时序分析方法与支持向量机的刀具磨损状态识别方法。该方法利用混沌时序分析方法重构了刀具声发射信号的相空间,并提取了嵌入维数与Lyapunov系数建立了特征空间。使用支持向量机作为分类器,实现了刀具磨损状态的识别。实验证明,在小样本学习情况下,基于混沌时序分析方法与支持向量机的刀具磨损状态识别方法具有良好的学习能力,获得了较高的识别准确率。

To distinguish and acquire the wear state of tools, a tool wear state recognition method based on chaotic time series analysis and support vector machine was proposed. The phase space was reconstructed by chaotic time series analysis method, and the embedding dimension and Lyapunov exponent were calculated and integrated as the feature space. Support vector machines （SVMs） algorithm was used as the classifier to realized recognition of different wear states. Experiments result showed that the tool wears state recognition method based on chaotic time series analysis and support vector machine were with excellent ability to learn and to get a high recognition accuracy in the case of small sample.