Research on life prediction method of smart meters in multi environment based on PSO-SVR

In this paper, the life prediction problem of smart meters under multiple operating environments is thoroughly investigated for the demand of high reliability operation of smart meters under harsh environmental operating conditions. A PSO-SVR based smart meter life prediction method is proposed for the existing difficulties of too small samples related to smart meter life of different operating environment, and the difficulty of predicting smart meter life in different environments. Finally, the data collected from four representative provinces and regions in China are used to conduct prediction experiments on the smart meter life under four typical climatic conditions in China. The results of the simulation experiments show that the model prediction accuracy of the method is high and outperforms the BP neural network algorithm performance.