Status Monitoring and Predictive Maintenance of Reciprocating Compressor Based on GA-BP
At a booster station,the existing compressor status monitoring system cannot effectively identify fault signals,but fail to provide warning before faults occur.To solve this problem,a reciprocating compressor sta-tus monitoring and prediction system for predictive maintenance was developed,and a complete software and hard-ware system was built.Air valve is an important moving part of reciprocating compressor,and the measuring point for its status monitoring is located at the center of the outer surface of the valve cap.Taking the monitoring of the air valve status for example,a vibration mechanics model of the air valve cap was built.Then,the time domain,frequency domain and wavelet feature extraction methods were comprehensively used to conduct vibration signal a-nalysis,obtaining the signal features of normal and fault states,and extracting the state indicators.Finally,the genetic algorithm was used to optimize the neural network model and build a GA-BP based reciprocating compressor vibration trend prediction model.The test results show that the prediction error of the model is less than 0.04.Field application demonstrates that the system achieves all-weather real-time online monitoring and prediction as well as predictive maintenance,reduces maintenance pressure of working personnel,improves the safety and efficiency of unit operation,and saves operation and maintenance costs.
万方数据
维普
