基于轴心轨迹自动辨识的高速主轴系统故障诊断研究

提出一种轴心轨迹自动辨识方法实现高速主轴系统故障诊断。结合使用中值滤波和小波滤波去除轴心轨迹中的脉冲噪声和高斯噪声,将轴心轨迹转换成轴心轨迹图像,然后采用线矩计算不变矩识别轴心轨迹特征,通过建立的BP神经网络进行高速主轴系统的故障诊断。在高速主轴轴承系统试验器进行实验,通过对不平衡故障和不对中故障的诊断,验证了本方法的正确性,为有效保证数控机床主轴系统的可靠性提供了保障。

Study on Fault Diagnosis of High-Speed Spindle System Based on Automatic Identification of Axis Orbit

A new method is proposed for calculating characteristic vector of axis orbit, which can be used for fault diagnosis of highspeed spindle system. By using median filtering and wavelet filtering method, impulse noise and Gauss noise in the axis orbit are removed. The new identification method of axis orbit feature is proposed, in which the axis orbit is transformed into orbit image and then the invariant moments are calculated. Fault diagnosis of high-speed spindle is realized based on BP neural network. By the experiment of high-speed spindle bearing test bench, reliability of the method is verified through diagnosis of unbalanced and misaligned faults, which provides technical means for effectively ensuring the reliability of computer numerical control machine tools.