Hybrid Model For The Next Hourly Electricity Load Demand Forecasting Based on Clustering and Weather Data

This research presents about short term load demand customer-prediction model for the supply system based on clustering used to secure the electricity load an industrial customer. The customer-based load hybrid using Part Swarm Optimization Backpropagation prediction method is built on forecasting generation capacity and demands in the next 1 hours ahead. To sustain the forecast model results, the daily clustering and weather forecasts supplied by local authorities, are incorporated in our hybrid model. The model's simulation was tested by calculating the Mean Absolute Percent Error (MAPE) value 0.01% for the electricity load demand forecasted data business rate and 0.005% for the industry rates.