基于大数据深度迁移模型的机械故障诊断

Mechanical Fault Diagnosis Based on Big Data Deep Migration Model

针对现有机械故障诊断方法在小样本条件下检测率低的不足,提出一种基于深度迁移学习模型的机械大数据故障诊断方法研究。构建深度学习模型,计算模型的稀疏特性和分类错误率指标,并基于此提取机械大数据的故障特征类型;针对实际检测中有效样本较少的不足,利用迁移学习方法将实验数据用于辅助机械故障特征大数据的训练与测试,不断地调整输出结果并提高对故障点的定位与诊断精度。实验结果表明,提出诊断方法的G-Mean指标优于现有方法,在故障比为1:1000的条件下,故障查准率仍可达到96.34%。

In view of the low detection rate of existing mechanical fault diagnosis methods under the condition of small samples, this paper proposes a research on mechanical big data fault diagnosis method based on deep transfer learning model. Build deep learning model, calculate the sparse characteristics and classification error rate index of the model, and extract the fault feature types of mechanical big data based on this;in view of the shortage of less effective samples in the actual detection, use the migration learning method to use the experimental data to assist the training and testing of Mechanical fault feature big data, constantly adjust the output results and improve the location of fault points Diagnostic accuracy. The experimental results show that the g-mean index of the proposed diagnosis method is better than that of the existing method. Under the condition that the fault ratio is 1:1000, the accuracy of the fault can still reach 96.34%.