State Monitoring of Gas Regulator Station Based on Feature Selection of Improved Grey Relational Analysis

Abstract

The in-depth application of the IoT technology in the gas industry has improved the intelligence of the gas system. As an important part of the gas system, the optimization of state monitoring of regulator stations is of great significance. The efficiency of monitoring can be improved by feature selection, but the reduction of important features will reduce the accuracy of identification. Therefore, a feature selection method based on gray relational analysis is proposed. First, the distance between the comparison matrix and the reference matrix is transformed using the geometric probability distribution, which solves the problem of data initial transformation. Then, an adaptive value method of resolution coefficient based on the cuckoo algorithm is proposed. Finally, a weight coefficient is constructed by combining the redundancy and the correlation to improve the gray relational degree applied to nonsequential systems. The maximum classification accuracy and the number of selected features are used to compare the performance of feature selection methods for several different types of data sets. Simulation analysis shows that the proposed method has higher maximum classification accuracy and smaller selected feature set. Finally, the proposed method is applied to the state monitoring of gas regulator stations. Seven feature selection methods, including the proposed method, the classical methods, and the advanced methods are combined with a support vector machine, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> -nearest neighbor, and decision tree, respectively. The experimental results indicate that the comprehensive performance of the proposed method combined with the three classifiers is good. It enables the subset containing the most identifiable features to be obtained quickly, which is beneficial to improve the efficiency of state monitoring.