Electricity Demand Forecasting for the BCS Region of the Mexican SEN Using Artificial Neural Networks

In this article, electricity demand forecasting is performed by an Artificial Neural Network (ANN). The ANN is created from two years of data (17520 hours of operation/supply) of the Mexican National Electric System (SEN), obtained from the system operator (National Energy Control Center, CENACE). In order to increase the accuracy of the network, an exhaustive filtering of atypical values of the raw data was performed. The developed forecasting method was applied to the region of Baja California Sur (BCS), a province of small demand compared to the other areas of the Mexican SEN. The 14-year electricity demand of the region was estimated using ANN. The results obtained are evaluated showing an acceptable behavior of the ANN.