Analyzing adjustment and verification errors in electric metering devices for smart power systems considering multiple environmental factors

Environmental conditions can greatly influence the precision of electric metering devices, resulting in heightened measurement errors. This paper presents a correction model for electric metering equipment, considering a range of environmental influences. Initially, the potential impacts of various environmental elements on electric metering devices are analyzed. Subsequently, mutual information methods are employed to screen out the environmental factors with significant influence on the electric metering devices. Then, an error adjustment model based on gated recurrent unit-attention is constructed, and the particle swarm optimization algorithm is adopted for the purpose of optimizing hyperparameters. Ultimately, various evaluation metrics are selected, followed by experimental validation to confirm the suggested method’s efficacy. Case studies demonstrate the proposed method performs well across different seasons, with the lowest RMSE reaching 1.24.