Predicción de la generación de energía eléctrica producida en una central hidroeléctrica utilizando redes neuronales artificiales

The application of artificial neural networks was executed due to the existence of economic losses in the distribution systems, caused by the inconvenience that exists in the prediction of the generation of electrical energy. The objective of this research is to predict the generation of electrical energy in the Illuchi 1 hydroelectric plant, using artificial neural networks through training algorithms from the Keras Python library. For this, historical data (2015 - 2019) of the flow and active power of the Illuchi 1 hydroelectric plant were obtained, provided by ELEPCO SA, which were divided, transformed and grouped for subsequent training and testing. The artificial neural network was structured by an input layer with 10 neurons, two hidden layers, two freezing layers and an output layer, for the training of the models a Python software platform was started, with the use of algorithms of training of the Keras library and for validation the optimizer RMSPROP was achieved with 10 numbers of delays and 10 neurons. It should be mentioned that when predicting electrical energy, it contributes to the National Development Plan "Toda una Vida" for the optimization of the use of natural resources.

• Autor principal: González Piedra, Joel André (author)
• Altres autors: Hidalgo Barrera, Christopher Alexander (author)
• Format: bachelorThesis
• Idioma: spa
• Publicat: 2022
• Matèries: RED NEURONAL; PREDICCIÓN; ENERGÍA ELÉCTRICA; ALGORITMOS; ELÉCTRICA; POTENCIA
• Accés en línia: http://repositorio.utc.edu.ec/handle/27000/9337