Fault Diagnosis of Reciprocating Compressor Valve Using Acoustic Emission

Author(s):  
Y. F. Wang ◽  
X. Y. Peng

The faults of a reciprocating compressor valve can be diagnosed using the acoustic emission. Four typical valve faults including the crack, rupture and deformation in the valve discs and leakage through the flow passage were investigated. The fault features were extracted by comparing the acoustic emission signals from the failed valves with those from the normal valves. The results show that the feature locations where the discharge valve opened and closed could easily be identified by the envelope waveform of acoustic emission signal, and they changed when the valve failed including the rupture and deformation in valve discs and leakage through the flow passage and changed with the variation of the discharge pressure. The extent to which the valve failed could be estimated by the deviation degree between the opening/closing locations and the standard ones. The leakage caused by these valve faults could also lead to the increase in the amplitude of the acoustic emission wave. However, the fault of crack in valve disc couldn’t be identified by acoustic emission signal effectively.

Author(s):  
Yuefei Wang ◽  
Ang Gao ◽  
Sulu Zheng ◽  
Xueyuan Peng

The failure of suction/discharge valves is the most common cause of unscheduled compressor shutdowns; therefore, the in-time fault diagnosis of valves is crucial to the reliable operation of reciprocating compressors. Major valve faults include leakage, valve flutter, delayed closing, and improper lift. To determine the features for diagnosing these typical valve faults, this paper presents an experimental study of the fault diagnosis of reciprocating compressor valves with acoustic emission technology and simulated valve motion. The measured AE signals and simulated valve motions of normal and failed valves are studied. The results of the fault diagnosis indicate that an earlier occurrence of the suction process can diagnose suction valve leakage and that an earlier occurrence of the discharge process can be used for detecting discharge valve leakage. The leakage also causes an increase in the amplitude of the continuous acoustic emission signal. Valve faults resulting from improper valve lift can be diagnosed by the amplitude of the burst acoustic emission signal. The number of burst acoustic emission signals and the shape of the simulated valve motion can be used to monitor the valve flutter conditions. The location where the valve closes can diagnose a valve-delayed closing fault.


2020 ◽  
pp. 61-64
Author(s):  
Yu.G. Kabaldin ◽  
A.A. Khlybov ◽  
M.S. Anosov ◽  
D.A. Shatagin

The study of metals in impact bending and indentation is considered. A bench is developed for assessing the character of failure on the example of 45 steel at low temperatures using the classification of acoustic emission signal pulses and a trained artificial neural network. The results of fractographic studies of samples on impact bending correlate well with the results of pulse recognition in the acoustic emission signal. Keywords acoustic emission, classification, artificial neural network, low temperature, character of failure, hardness. [email protected]


Sign in / Sign up

Export Citation Format

Share Document