Identification and Number Prediction of Natural Gas Heating Users in a Company in Chengdu

Abstract

Nowadays, more and more residents in Sichuan and Chongqing area choose to use gas for heating in winter, and the demand for gas is increasing. Therefore, it is necessary to improve the refined management of residential gas consumption. Firstly, this paper uses K-means clustering according to the gas consumption data from 2017 to 2021 provided by a company in Chengdu, so as to obtain the regulation of residents’ gas consumption. According to the different users’ gas consumption habits, the gas heating users in 2017 are identified. Then, a better SVM model is selected in SVM, KNN and logistic regression to classify the gas consumption data in 2018-2021. Finally, according to the classification results, the grey model is used to predict the number of gas heating users in the next two years, and compared with the real data in 2022, the results shows that the model effect is good. This study provides strong theoretical and technical support for gas companies to rationally plan gas consumption and optimize the operation of urban gas pipeline network.