Rotating machinery fault diagnosis method based on multi-attention convolution neural network

Abstract

The invention discloses a rotating machinery fault diagnosis method based on a multi-attention convolution neural network. Firstly, acceleration vibration signals of rotating machinery without faultsand with different faults under different operation states are collected, fault state labels are set according to the fault states corresponding to each acceleration vibration signal, standardizationprocessing is carried out on each acceleration vibration signal, a multi-attention convolution neural network model is trained as a training sample, the multi-attention convolution neural network model comprises six convolution layers, five combined attention modules, a global average pooling layer and a Softmax layer, and then current acceleration vibration signals of the rotating machinery are collected and sent to the multi-attention convolution neural network model for fault diagnosis. According to the method, the attention modules are introduced into the convolution neural network so as to enhance the learning of the network on distinguishing characteristics and the fault impact signal sections, irrelevant noise is restrained, and the fault diagnosis performance is improved.

Description

technical field [0001] The invention belongs to the technical field of fault diagnosis of rotating machinery, and more specifically relates to a fault diagnosis method of rotating machinery based on a multi-attention convolutional neural network. Background technique [0002] Rotating machinery is a widely used component in industrial equipment. Once a failure occurs, it will inevitably lead to deterioration of equipment performance, resulting in economic losses and even safety accidents. Therefore, it is of great significance to carry out fault diagnosis on rotating machinery. [0003] In recent years, deep learning technology, as an efficient feature extraction and pattern recognition algorithm, has solved the important problem of manually extracting features in the past. Therefore, fault diagnosis research based on methods such as denoising autoencoders, deep belief networks, and convolutional neural networks has achieved a large number of research results. In particula...

AI Technical Summary

Problems solved by technology

1) CNN technology tends to learn all the features of the input signal, and cannot learn fault-related features in a targeted manner, resulting in a huge waste of resources; 2) For the issue of how CNN learns discriminative features, published Outcomes lack of adequate research

Images