Motor bearing fault diagnosis method and device

Abstract

The invention provides a motor bearing fault diagnosis method and device which can convert a one-dimensional vibration signal of a bearing into a two-dimensional gray-scale map by adopting a sample-image conversion method and reduce complexity of a structure of a two-dimensional convolutional neural network. The method comprises the steps of: acquiring a motor bearing fault data set, converting aone-dimensional bearing fault signal in the fault data set into a two-dimensional fault signal gray-scale map to obtain a training set image; according to the obtained training set image, training a fault diagnosis model by the two-dimensional convolutional neural network; acquiring a one-dimensional bearing vibration signal in the field, and converting the one-dimensional bearing vibration signalinto a two-dimensional field signal gray-scale map; and inputting the two-dimensional field signal gray-scale map into the trained fault diagnosis model, outputting each fault category affiliation weight by the trained fault diagnosis model, and according to each output fault category affiliation weight, determining a fault category of a current bearing. The motor bearing fault diagnosis method and device are applicable to the field of motor bearing fault diagnosis.

Description

technical field [0001] The invention relates to the field of mechanical equipment fault diagnosis, in particular to a motor bearing fault diagnosis method and device. Background technique [0002] Bearing is one of the most critical parts in rotating machinery, and its normal operation is directly related to the performance of the machine. If the bearing breaks down and the machine cannot work normally, it will cause stoppage of production and work, and it will directly endanger people's life safety. Although the traditional one-dimensional convolutional neural network has been used for fault diagnosis of bearings, it is not easy to extract features because of one-dimensional time series signals, and it is prone to feature loss, resulting in low final diagnosis accuracy. However, most of the two-dimensional convolutional neural network structures are not directly applicable to the one-dimensional vibration signal of the bearing. Therefore, in order to suppress overfitting, ...

AI Technical Summary

Problems solved by technology

[0003] The technical problem to be solved by the present invention is to provide a motor bearing fault diagnosis method and device to solve the problem existing in the prior art that the depth of the two-dimensional convolutional neural network needs to be deepened for bearing fault diagnosis, resulting in a complex structure of the neural network

Images