Generalized Regression Neural Network For Long-Term Electricity Load Forecasting

Abstract

The availability of electricity demand is very high. Many households and industrial equipment are using electricity as the source energy. The reliability of the power system in saving the budget is very much needed. This can be succeeded by doing good and proper operation planning. The important step of the electric power system operation planning is to predict load electricity. The load forecasting can support the corporations of electricity to assign the cost and power generation. Long-term forecasting is a technique of predicting periods for more than one year. The old data will be a guide to solve the issues. In this research, the concept of generalized regression neural network (GRNN) is to predict long-term electricity load. The advantage of the GRNN method can estimate the absolute function between input and output data sets directly from training data. The research was compared to the results of the actual data, Feed Forward Backpropagation Neural Network (FFBNN), Cascade Forward Backpropagation Neural Network (CFBNN) and Generalized Regression Neural Network (GRNN). The results of the study will be measured and validated using the Mean Absolute Deviation (MAD) and Mean Absolute Percentage Error (MAPE) methods.