A State Monitoring Method of Gas Regulator Station Based on Evidence Theory Driven by Time-Domain Information

Regulator stations are widely used in gas transmission and distribution systems. Their state monitoring is of great significance to the safe operation of gas pipe networks. Due to the complexity of the working environment and the limitation of sensors, the acquired information is uncertain, which makes the state monitoring result prone to errors. The evidence theory has the ability to solve the uncertainty problem effectively. Most of the improvement methods of the evidence theory are in the spatial domain. These methods are not applicable to the fusion of the time-domain information. In this article, an improved method of the evidence theory is proposed for the state monitoring of a gas regulator station. It can meet the requirement of the dynamic fusion of the time-domain information. First, the back-propagation neural network is used to judge whether the evidence conflicts with each other. The simulation results demonstrate that it can judge the conflicts well. On this basis, the relative conflict factor is proposed to modify the evidence, and the calculation method of the adaptive time attenuation factor is proposed to reduce the accumulated error. The dynamic fusion of the time-domain information is realized by combining the time attenuation factor and the relative conflict factor. Finally, the proposed method is applied to the state monitoring of the gas regulator station. The feasibility and effectiveness of the method are verified by experiments. It verifies that the variation of the support degree of the proposed method for the correct proposition is 0.1478 higher than that of the temporal evidence combination based on relative reliability factor when the evidence is strongly conflicting.

Automated summary

This article proposes an improved method of evidence theory for the state monitoring of a gas regulator station. By using a back-propagation neural network to judge evidence conflicts, introducing a relative conflict factor to modify evidence, and calculating an adaptive time attenuation factor to reduce accumulated error, the dynamic fusion of time-domain information is achieved. The feasibility and effectiveness of the method are verified through experiments.