Hourly Load Forecasting Model Based on Real-Time Meteorological Analysis

Accurate short-time load forecasting is vital for power system planning, scheduling, and operating. Therefore, a new hourly load forecasting model is designed basing on BP Neural Network in the literature, which takes consideration of the load information of previous hour and a variety of real-time meteorological factors including temperature, humidity, wind speed, air pressure and visibility. Based on the analysis of real-time meteorological factors on load, the correlation between load and real-time meteorological factors is concluded. Compared to method of BP neural network with the consideration of daily characteristic meteorological factors, the results show that taking real-time weather factors as input improves the accuracy of the load forecasting.