Diseño de Red Neuronal Artificial para la Predicción de la Demanda Eléctrica

Design of an Artificial Neural Network for Load Consumption Forecasting

Abstract

The commercial store "El Machetazo" faced difficulties in predicting the electric demand curve, which negatively impacted the management of its photovoltaic solar system and compliance with the energy-saving plan. This work aimed to develop a predictive tool based on artificial neural networks to optimize the energy management of the store. To achieve this, a Long Short-Term Memory type neural network model was implemented using historical electricity consumption data. The methodology included data collection and preprocessing, model design and training, and evaluation of its accuracy. The results indicated that the model accurately predicted electrical demand with a mean absolute error of 0.02, enabling more efficient management of the photovoltaic solar system and significantly contributing to achieving energy goals. This advancement not only improved operational efficiency but also promoted sustainable use of energy.