精细复合多尺度模糊熵在电机轴承损伤检测中的应用

Application of Refined Composite Multiscale Fuzzy Entropy in Motor Bearing Damage Detection

针对电机轴承的损伤检测问题,本文提出一种基于精细复合多尺度模糊熵算法(RCMFE)和粒子群优化最小二乘支持向量机的诊断模型,通过提取振动信号中的关键特征有效识别轴承的运行状态.该方法以粗粒化过程和模糊熵的优化改进算法,克服了传统多尺度模糊熵(MFE)因粗粒化尺度不断增大而引起的熵值不规则浮动、误差增加等缺陷.之后,结合电机轴承不同工况的振动数据进行实验验证.结果表明,相较于其他方法,RCMFE计算熵值的误差更小,一致性更强,且分类结果显示RCMFE获得了更高的诊断精度.

Aiming at the problems in detecting damage of motor bearings,a diagnosis model based on refined composite multiscale fuzzy entropy and least squares support vector machine based on particle swarm optimization is proposed to effectively identify the operating state of bearings by extracting key features from vibration signals.This method is improved by optimizing the coarse-graining procedure and the fuzzy entropy,to overcome the shortcomings of traditional multiscale fuzzy entropy,such as the irregular floating of entropy values and increase of errors,caused by the increase of coarse-graining scale.Then,the vibration data of motor bearings under different operating conditions are verified by experiments.The results show that compared with other methods,RCMFE has a smaller error in calculating entropy and a stronger consistency,and the classification results show that RCMFE has a higher diagnostic accuracy.