Mid-term Load Forecasting Algorithm for Large-Scale Power Systems Based on Deep Learning Considering the Impact of Behind-the-Meter Solar PV Generation

초록

An algorithm for mid-term load forecasting (MTLF) is introduced for large-scale power systems, incorporating the influence of behind-the-meter (BTM) solar PV generation on system loads. To account for the impact of BTM solar PV generation, the installed capacity of the BTM solar PV generator is estimated and used as an input for the deep learning-based forecasting model. In the case studies, weekly peak loads in the Korea Power System are forecasted for the period from January 2022 to December 2022, employing both multiple linear regression and deep learning methodologies. The comparison of forecasted loads reveals a notable enhancement in the accuracy of MTLF when the impact of BTM solar PV generation is incorporated into the forecasting model.

키워드