A Hybrid Approach for Weather-based Power Interruption Forecasting using Multilayer Perceptrons and Parametric Models

This paper investigates the impact of various weather conditions on the reliability performance of power distribution networks. Especially, a hybrid approach based on multilayer perceptrons (MLPs) and parametric models is proposed to forecast the daily numbers of sustained and momentary power interruptions in the distribution management area using chronological weather data. First, the parametric regression models are implemented to analyze the relationship between power interruptions and different weather characteristics including temperature, wind speed, rain precipitation, air pressure, and lightning. The selected weather characteristics and corresponding parametric models are then integrated as inputs to formulate a MLP neural network model to forecast the daily numbers of power interruptions. In addition, a modified extreme learning machine (ELM) based hierarchical learning algorithm is introduced for training the formulated forecasting model. Finally, the real power interruption data collected from a Florida electric utility is used to verify the applicability and effectiveness of the proposed hybrid approach.