Development Of Internal Valve Leakage Detection Instrument And Experimental Research

The internal leakage of the valve has always been a major hidden danger for the safety of natural gas pipeline transportation.Therefore,the quantitative identification analysis of the leakage of the valve is particularly important for the safe operation of natural gas pipelines.This article has done an investigation on the detection method,technology and instruments of the internal valve leakage based on acoustic emission thoroughly.Based on the many years’ research results of the laboratory,the third generation internal valve leakage detection system based on acoustic emission has been developed.The internal leakage sample data of the ball valve and plug valve were collected.Due to the complexity of the experimental environment,the collected signals contain a large amount of environmental noise.In order to obtain a purified internal leakage signal,in this paper,Based EMD threshold noise reduction method and EMD spatio-temporal filtering noise reduction method,the research for reducing the noise interference in the internal valve signals has been carried out.Finally,the RBF neural network quantitative models are built based on ball valve and plug valve sample data of no reduction,EMD threshold noise reduction,EMD spatio-temporal filtering.Four evaluation indexes are used to evaluate these models,which are Mean Square Error(RMSE),Mean Absolute Proportional Error(MAPE),Mean Absolute Error(MAE)and Pearson correlation coefficient(Pearson).The experimental results show that the EMD spatio-temporal filtering is a good noise reduction method for the valve leakage signal with the environmental noise.The prediction models of ball valve and plug valve are superior to the other two models.In addition,in the field test,the instrument can well perform qualitative and quantitative analysis of ball valve and plug inner leakage and also can satisfy the requirements of natural gas station.