A Sparse Constrained Generative Adversarial Network Implementation Method for Vibration Signals of Rotating Machinery

Abstract

The invention discloses a method for implementing a sparsity-constrained generation confrontation network of vibration signals of rotating machinery, including: constructing an input layer dimension and an output layer dimension as w , the hidden layer dimension is m The sparse autoencoder of ; the dimension obtained after preprocessing the collected vibration signal is w The training sample of the vibration signal trains the constructed sparse autoencoder to obtain a trained sparse autoencoder; use the trained sparse autoencoder to construct a sparse constrained generative adversarial network including a generator and a discriminator; use the dimension for w The sparse constrained generative adversarial network is trained with the vibration signal training samples and noise samples, and the sparse constrained generative adversarial network that can use noise to generate vibration signals of rotating machinery is obtained.

Description

technical field [0001] The invention relates to the technical field of vibration signal generation of rotating machinery, in particular to a method for implementing a sparse constraint generation confrontation network of vibration signals of rotating machinery. Background technique [0002] As an important part of industrial equipment, rotating machinery plays a key role in the operation of the equipment. Therefore, the operating status of the rotating machinery can greatly affect the overall operating status of the equipment. Once the rotating machinery fails, it is easy to cause the overall failure of the equipment. Lead to adverse consequences such as equipment downtime, economic losses and damage to personal safety. However, rotating machinery usually operates under harsh environmental conditions such as high loads and variable working conditions, and is prone to degradation and failure. Therefore, health management such as fault detection, fault diagnosis, and health as...

AI Technical Summary

Problems solved by technology

However, neither spectrum generation nor feature generation can preserve all the information of the original vibration signal

Existing methods lack the ability to stably generate the original time-domain vibration signals of rotating machinery, and all require complex network structures, which further lead to training instability

Images