基于直方图均衡化和卷积神经网络的轴承故障诊断方法

Bearing Fault Diagnosis Method Based on Histogram Equalization and Convolutional Neural Network

针对传统的滚动轴承故障诊断方法依赖人工特征提取和专家经验,难以自适应提取强噪声信号微弱故障特征的问题,提出一种直方图均衡化和卷积神经网络(CNN)相结合的智能诊断方法。首先,将传感器采集到的一维振动信号通过横向插值法转换为便于模型识别的二维振动图像,利用直方图均衡化技术拉伸像素之间灰度值差别的动态范围,突出纹理细节和对比度,以增强周期性故障特征;然后构建深层CNN模型,采用优化技术降低模型参数量,逐层学习监测数据与故障状态之间的复杂映射关系。实验结果表明该方法具有高达99%以上的准确率,对不同负载下的故障信号仍具有较高的识别精度和泛化能力。

To solve the problem that traditional fault diagnosis methods of rolling bearing rely on artificial feature extraction and expert experience,and it is difficult to self-adapt to extract weak fault features of strong noise signals,an intelligent diagnosis method combining histogram equalization and convolutional neural network(CNN)is proposed.First,the one-dimensional vibration signal collected by the sensor is transformed into a two-dimensional vibration image that is easy to be recognized by the model through Transversal interpolation.Histogram equalization technology is used to stretch the dynamic range of gray value difference between pixels,highlight texture details and contrast,and enhance periodic fault characteristics.Then,a deep CNN model is constructed,and the optimization technology is used to reduce the model parameters,and the complex mapping relationship between monitoring data and fault state is learned layer by layer.The experimental results show that this method has a high accuracy of more than 99%,and still has a high identification accuracy and generalization ability for fault signals under different loads.