摘要： 针对表面粗糙度不能及时检测造成的工件浪费问题，首次提出根据主轴电机电流信号进行表面粗糙度检测分类。通过实验采集不同表面粗糙度加工时的主轴电机电流信号，采用小波包分解将电流信号分解成不同频段，借助能量特征和裕度因子对不同频段电流信号进行评估，过滤低相关性频段，再通过随机森林筛选特征，降低特征的冗余性。总谐波失真特征实现了积屑瘤检测，仅依赖构建的电流信号特征工程表面粗糙度检测准确率高达95%以上，并且检测时间在2 s以内，基本实现了工件表面粗糙度的快速准确检测。

中图分类号：TP181 文献标志码：A DOI: 10.16157/j.issn.0258-7998.234208
中文引用格式： 刘雪杰，李国富，任潞. 基于主轴电机电流信号的表面粗糙度检测[J]. 电子技术应用，2024，50(2)：54-59.
英文引用格式： Liu Xuejie，Li Guofu，Ren Lu. Surface roughness detection based on spindle motor current signal[J]. Application of Electronic Technique，2024，50(2)：54-59.

Surface roughness detection based on spindle motor current signal

Abstract： Workpiece waste is usually caused by delayed detection of surface roughness. A rapid surface roughness detection classification based on the current signal of the spindle motor is proposed for the first time. The current signals of the spindle motor under different surface roughness processing conditions are collected through experiments, and the current signals are decomposed into different frequency bands through wavelet packet decomposition. The current signals of different frequency bands are evaluated by the energy characteristics and the margin factors, and the low correlation frequency bands are filtered. Then the features are screened through random forest to reduce the redundancy of features. The total harmonic distortion feature achieves built-up edge detection during the machining process. The workpiece surface roughness detection accuracy is as high as 95%. And the detection time is within 2 seconds. Spindle current signal analysis basically achieves fast and accurate detection of workpiece surface roughness.

Key words : spindle motor current signal；wavelet packet decomposition；random forest；the total harmonic distortion；surface roughness