Depth-learning-theory-based fault diagnosis method for rotating machinery

Abstract

The invention relates to a depth-learning-theory-based fault diagnosis method for rotating machinery. An equal-angle interval sampling time sequence signal is constructed by using an equal-interval sampling sequence signal, thereby eliminating a non-stationary property caused by non-constant-speed rotation; training and fault diagnosis of a deep neural network are carried out by using an autocorrelation sequence and Fourier transform of the equal-angle interval sampling time sequence signal and a temperature during working of the diagnosed rotating machinery as inputs of the deep neural network; newly generated data and corresponding fault states are added into a training sample set; and the deep neural network is trained again by using the new sample set. Therefore, self learning and self perfection of the deep neural network (DNN) are realized.

Description

technical field [0001] The invention relates to fault diagnosis of rotating mechanical equipment, in particular to a fault diagnosis method for rotating mechanical equipment based on deep learning theory. Background technique [0002] With the development of modern industry and science and technology and the further improvement of the degree of automation, mechanical equipment is developing in the direction of large-scale, high-speed, continuous, centralized, and automated. Rotating mechanical equipment has been in high-speed operation for a long time (generally more than 3000 revolutions per minute or even as high as tens of thousands of revolutions). Due to the influence of various factors, some failures will inevitably occur, and these failures often cause huge economic losses or even catastrophic consequences. , so it is very important for fault diagnosis of rotating machinery equipment. Vibration detection and diagnosis method is the most commonly used diagnosis method...

AI Technical Summary

Problems solved by technology

However, this method only uses the frequency domain characteristics of the vibration signal, and does not overcome the frequency ambiguity caused by vibration non-stationarity, nor does it use the time domain information of the vibration signal and the temperature information of the monitored object that is highly correlated with the fault. The effect of the non-uniform speed scene is not ideal, and the severity of the fault cannot be judged

In addition, this method does not provide a method for self-learning and self-improvement of the constructed deep neural network.

Images