Reciprocating compressor fault diagnosis system and method based on neural network algorithm

Abstract

The invention discloses a reciprocating compressor fault diagnosis system and method based on a neural network algorithm. The reciprocating compressor fault diagnosis system and method based on the neural network algorithm comprises a field communication module, a data collecting module, a simulation model based on neural networks, an expert system module based on standard fault characteristics, afault diagnosis main machine and an upper computer. According to the reciprocating compressor fault diagnosis system and method based on the neural network algorithm, the neural network algorithm isutilized, measured compressor system status parameters are combined, and thus simulation models in all working conditions are acquired; compressor faults are simulated by changing system parameters tocalculate and obtain the standard fault characteristics, after the abnormal operation of the compressor occurs, through the comparison with the standard fault characteristics, fault parts can be accurately located, and elimination methods can be brought; and the diagnosis success rate is high, and real-time monitoring analysis of high precision equipment is not needed, and the diagnosis cost is lowered.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis of equipment, in particular to a fault diagnosis system and method for a reciprocating compressor based on a neural network algorithm. Background technique [0002] As the core equipment of oil and gas field transportation, the reciprocating compressor has very important significance for its stability, reliability and safety. Because the structure of the compressor is relatively complex, there are many parts, and the assembly is complicated, sometimes the reciprocating compressor produced with the same set of drawings will have different operating characteristics in the field, so it is difficult to construct an accurate overall mathematics for the compressor Therefore, when fault symptoms appear, how to reduce direct or indirect losses caused by compressor faults through fault diagnosis has always been an important topic in the compressor industry. [0003] The current mainstream detectio...

AI Technical Summary

Problems solved by technology

Because the structure of the compressor is relatively complex, there are many parts, and the assembly is complicated, sometimes the reciprocating compressor produced with the same set of drawings will have different operating characteristics in the field, so it is difficult to construct an accurate overall mathematics for the compressor Therefore, when fault symptoms appear, how to reduce direct or indirect losses caused by compressor faults through fault diagnosis has always been an important topic in the compressor industry

[0003] The current mainstream detection methods include vibration detection, thermal parameter detection, stress and strain detection, etc. However, the above detections only detect and judge specific faults of the compressor, and cannot compare and judge the overall and system status of the compressor.

At present, there are some fault diagnosis systems for engine equipment that use online real-time monitoring of engine dynamics and vibration characteristic parameters for fault diagnosis. The method is more prone to false alarms, and ultimately most of the technicians can only use experience to deal with the cause of the failure, which is time-consuming and labor-intensive and is more prone to errors

Images