Introduction  The rapid development of clean energy sources, such as wind and solar power, has led to significant changes in the energy structure of the power system, which consequently has increased uncertainty in safe grid operation and imposed new challenges in accurately forecasting electricity consumption. Among the numerous influencing factors on grid electricity consumption, meteorological factors exert a significant impact. Therefore, it is imperative to analyze the influence of meteorological factors on the refined forecast of grid electricity consumption.

  Method  The influence of meteorological factors on electricity consumption was investigated, based on the daily electricity consumption data and meteorological elements in 2017, including the maximum temperature, average temperature, minimum temperature, atmospheric pressure, relative humidity and wind speed, and using the combined method of ensemble empirical mode decomposition (EEMD) and back-propagation (BP) neural networks.

  Result  This study reveals a significant correlation between the average temperature, maximum temperature, minimum temperature, atmospheric pressure, and relative humidity with the low-frequency component of the electricity consumption series processed by EEMD, and an insignificant correlation with the high-frequency component and periodic component.

  Conclusion  The electricity consumption forecast using the BP regression model exhibits considerable deviations when compared to the actual status. The electricity consumption forecast by the EEMD-BP regression model shows a significant improvement in accuracy, attributed to the incorporation of meteorological factors, indicating that the combined forecast method of EEMD-BP based on meteorological factors effectively enhances the accuracy of electricity consumption forecast. Consequently, it can serve as an effective technical support for improving short-term electricity consumption forecast methods.