Accumulated Temperature Effect on Rhythmic Urban Electric Load

With the acceleration of urbanization and improvements in residential living standards, the demand of urban electricity in China has been steadily increasing. In particular, prolonged abnormal weather conditions significantly increase the demand for temperature control equipments, leading to sharp peaks of power load in both summer and winter. This presents new challenges for the operation and scheduling of the power system, highlighting the urgent need for effective analysis and forecasting accurately the electric load. To address this issue, we systematically investigate the temperature impact on the electricity demand by analyzing the power load and temperature data of a city in China from 2015 to 2023. By developing a framework of time series decomposition based on an improved Prophet algorithm, we divide the load into the product of a trend factor and a fluctuation factor. We demonstrate the trend factor having a high correlation with the general economic growth and infrastructure development. We further decompose this factor into several components to reflect its periodic and aperiodic characteristics across different time scales. By identifying the relevance of these characteristics to the temperature, we capture the temperature-sensitive part out of the total load, which accounts for the influence of immediate weather conditions on the load. By applying causal inference tools, we verify the existence of an accumulated temperature effect (ATE) in the power system and deduce the duration of such effect through computing causal intensities. Our results clarify various components of urban electricity load and provides a foundation for future load forecasting, thereby supporting safe dispatch and operation for power companies, especially during extreme weather or emergency conditions.