LSTM recurrent neural network for energy demand forecasting

RODRIGO ALARCON; MARTIN ALARCON; GONZÁLEZ, ALEJANDRO HERNÁN; FERRAMOSCA, ANTONIO

Congreso; AADECA 2023; 2023

Recurrent Neural Networks (RNN) of the LongShort Term Memory (LSTM) type provide high accuracy inpredicting sequential models in various application domains. Asin most process control problems, their dynamics include nonmanipulatedvariables that need prediction. This paper proposesusing an LSTM neural network for energy demand forecasting,which applies to an Economic Model Predictive Control (EMPC)as a forecasting tool. For the training, data are taken froma three-phase intelligent power quality analyser located at theNational Technological University, Reconquista Regional Faculty(Santa Fe, Argentina). A recursive strategy is used to updatethe state of the neural network and the predictions made, asshown in different prediction horizons. The accuracy achievedin training the neural network is measured using the root meansquare error (RMSE) metric. Experimental results show thatthe proposed LSTM neural network has excellent generalisationcapability.