Investigation on recognition method of acoustic emission signal of the compressor valve based on the deep learning method

The valve affects the reliability and efficiency of the reciprocating compressor. The Acoustic Emission (AE) technology is applied to nondestructive measurement of compressor valve movement in this paper. Furthermore, the AE signal is analyzed to predict the valves dynamic characteristics of based on the deep learning method. The results show that the prediction accuracy of dynamic characteristics of valve by Convolutional Neural Network (CNN) artificial neural network and Long Short-Term Memory (LSTM) artificial neural network is 94.49% and 96.14%, respectively, which shows that the deep learning method can effectively predict the valve dynamic characteristics. The prediction accuracy of LSTM network is slightly higher than CNN. And the prediction speed of CNN is higher. Based on the models, the delay closing of the valve is analyzed. This paper provides the experimental and theoretical basis for the application of AE technology to the fault diagnosis of the reciprocating compressor. (C) 2021 The Author(s). Published by Elsevier Ltd.

Automated summary

The paper utilizes Acoustic Emission (AE) technology combined with deep learning methods to successfully predict the dynamic characteristics of reciprocating compressor valves, providing an experimental and theoretical basis for fault diagnosis.