Optimization And Test Of Fault Self-diagnosis Technology For Medium-low Gas Pressure Regulator

Under the call of energy conservation and emission reduction,gas energy has great potential for development and played an important role in the energy revolution in recent years.As the key facility of the gas pipeline network,the gas pressure regulator plays an important role in the gas transmission and distribution system.In view of the current problem that the failure of medium-low gas pressure regulators rely on artificial discrimination,it is necessary to realize the intelligentization of fault diagnosis,which has important social and economic significance for the safe operation and development of gas systems.Firstly,this paper introduces the research status of fault diagnosis of medium-low gas pressure regulators,and makes a brief analysis of the classification of regulators and the characteristics of different operating states.Secondly,the fault self-diagnosis system for medium-low gas pressure regulator based on EMD decomposition method and SVM algorithm is established.Finally,the actual case verification proves that this method achieves the ideal machine learning effect.The above provides an effective and feasible way for the intelligent diagnosis of medium-low gas pressure regulators.The specific research method is: collecting the outlet pressure data for EMD decomposition to obtain the energy moment value,and calculating the dynamic fluctuation amplitude index(DB)of each sample as supplementary criterion.The output pressure radar chart is drawn,and the output state of regulator is refined in depth,which is divided into 7different output states.The precedence chart method is used to determine the artificial diagnosis results of expert technicians reasonably.In the process of machine learning,the problem of less sample size is solved by cross-validation method,and the penalty parameter and kernel function parameters are optimized by grid search method.Based on this,the model of optimal support vector machine is established,and an effective fault self-diagnosis system for medium-low gas regulators is formed.The optimized self-diagnosis technology for medium-low gas pressure regulators is tested.And the results show that in the actual project,the optimized SVM model can classify the fault status of the medium-low gas pressure regulators correctly,The accuracy rate has increased to 91.40%.The on-line application of this technology optimizes the operation and management of medium-low gas pressure regulators,which is of great significance for ensuring the safe and efficient operation of gas pipeline networks.