基于深度学习的电机故障在线检测技术优化

Optimization of online motor fault detection technology based on deep learning

为有效提高电机故障检测效率,研究建立了深度学习模型,并利用深度学习模型构建在线检测函数。试验结果表明,基于深度学习模型可以降低电机轴承振动干扰信号,以保证清晰识别特征频率,所建立的深度学习检测模型对电机轴承外部的外圈磨损、轴承缺口且有裂纹及保护架破损的精度及准确率均大于98.5%,而电机轴承内部的检测精度与准确率较低,但检测精度仍大于90%,可满足电机故障在线检测要求。研究成果可为电机故障在线检测提供理论参考。

To effectively improve the efficiency of motor fault detection,establishes a deep learning model was established,and an online detection function was constructed using the deep learning model.The experimental results showed that the deep learning model reduced the vibration interference signal of motor bearings,ensuring clear recognition of feature frequencies.The accuracy and accuracy rate of the established deep learning detection model were all greater than 98.5%for the outer ring wear of the motor bearing,the bearing notch and crack,and the damage of the protective frame.The internal detection accuracy and accuracy rate of the motor bearing were relatively low,but the detection accuracy was still greater than 90%,which could meet the requirements of online detection of electrical faults.The research results can provide theoretical reference for online detection of motor faults.