Small sample rolling bearing fault diagnosis method under multiple working conditions

Abstract

The invention discloses a small sample rolling bearing fault diagnosis method under multiple working conditions. The method comprises the following steps: a, constructing a task set; b, inputting data in the task set into ISDAE for reconstruction to obtain a reconstructed signal matrix which retains effective characteristics in original signals and reduces noise; c, classifying the reconstructed signals by using the MAML, and training model parameters of the MAML to obtain an optimal network model; and d, inputting an original vibration signal of the monitored rolling bearing into the trained MAML model, and judging whether the rolling bearing has a fault or not and the type of the fault. According to the method, the bearing fault is diagnosed by adopting a method of combining model independence and improved sparse noise reduction self-coding, not only can the separable characteristics in the original vibration signal be extracted and the anti-noise capability of the signal be improved, but also the generalization capability of the model can be improved, so that the bearing fault of small sample data under multiple working conditions can be accurately diagnosed, and the fault diagnosis accuracy is improved. And safe operation of mechanical equipment is ensured.

Description

technical field [0001] The present invention relates to a small-sample rolling bearing fault diagnosis method under multiple working conditions combining model-agnostic meta-learning (MAML) and improved sparse denoising autoencoder (ISDAE), It belongs to the technical field of diagnosis. Background technique [0002] Rolling bearings are one of the key components of rotating machinery and are widely used in modern large-scale mechanical equipment. According to statistics, in rotating machinery using rolling bearings, about 30% of mechanical failures are caused by bearings. In recent years, intelligent fault diagnosis methods for rolling bearings based on deep learning have emerged one after another, providing a powerful tool for monitoring the safe operation of mechanical equipment. However, with the continuous improvement of bearing machining accuracy and material properties, the service life of bearings is continuously extended, and when the bearing is detected to be dam...

AI Technical Summary

Problems solved by technology

However, with the continuous improvement of bearing machining accuracy and material properties, the service life of bearings is continuously extended, and when the bearing is detected to be damaged, it will be replaced in time, resulting in the inability to obtain sufficient bearing fault data

Since the training of the deep learning model needs a large amount of data as the research object, the failure to obtain sufficient bearing fault data will make the training of the model in trouble, resulting in a more limited generalization ability of the deep learning model, and it is difficult to improve the recognition accuracy

In addition, the operating environment of rolling bearings is usually time-varying, and the fault conditions are complex. Different sampling frequencies, different fault degrees, different fault locations, different fault types, different fault sizes, etc., lead to extraction of dominant Separable depth features are very difficult, which increases the difficulty of intelligent diagnosis of bearing faults

Images