Bearing vibration signal characteristic interpretability dimension reduction and fault diagnosis method

Abstract

The invention discloses a bearing vibration signal characteristic interpretability dimension reduction and fault diagnosis method, which comprises the steps of collecting vibration signals of a bearing in different states, and dividing the vibration signals into a training set and a test set; constructing an information maximization generative adversarial network model; inputting the training set into an information maximization generation adversarial network model, and carrying out iterative training to obtain a trained network model; and inputting the test set into the trained network model for feature interpretability dimension reduction and fault diagnosis to obtain a bearing fault result. According to the method, the adversarial network is generated through information maximization, the obtained bearing fault diagnosis model is high in anti-noise capability, the fault type, the fault degree and the load of the bearing can be diagnosed at the same time, and the diagnosis precision is high.

Description

technical field [0001] The invention belongs to the technical field of bearing fault detection, and in particular relates to a bearing vibration signal feature interpretable dimension reduction and fault diagnosis method. Background technique [0002] Rolling bearings are one of the most common and easily damaged parts in wind turbines. In engineering practice, the failure of rolling bearings may lead to huge economic losses and accidents. In response to this problem, it is urgent to study the application of reliable intelligent algorithms, and realize early warning and intelligent fault diagnosis of wind turbine faults by carrying out rolling bearing fault diagnosis, so as to achieve reasonable operation scheduling, planned pre-inspection and pre-repair, so as to reduce the frequency of equipment failures and prolong the service life. The purpose of equipment operation cycle and reducing operation and maintenance costs. However, from the source of the fault to the vibratio...

AI Technical Summary

Problems solved by technology

However, through the attenuation and mixing of different transmission paths from the fault source to the vibration detection point, the monitored bearing signal is often mixed with a large amount of background noise, which has the characteristics of large interference, non-stationary, and nonlinear, which affects the accuracy of fault detection. Analysis and diagnosis

Images