Load characteristics analysis of regional distribution network during summer tourism peak period considering real-time meteorological coupling

The intricate stochastic and periodic fluctuations in power load underscore the challenge of accurate forecasting. Inaccuracy in forecasting disrupts grid dispatch and operation, leading to supply-demand imbalances and shortages. Thus, we designed a real-time meteorological-coupled load analysis for summer peak tourism in regional distribution networks. Focusing on meteorological factors, especially human comfort indices (heat, humidity, rain, wind), we used K-fold cross-validation to construct datasets and isolate meteorologically sensitive loads. Employing a hybrid firefly-K-means algorithm for load clustering, we introduced a load differential matching method with coupling coefficients to quantify load interactions. Experiments revealed a prediction error $\leq 0.05$, faster extreme value identification, and a gradual increase in algorithm time from 5 to 30 iterations.