Estimación de la demanda eléctrica mediante la red neuronal recurrente tipo LSTM aplicado al estudio de un sector residencial en una micro-red

Abstract

This work is carried out with the purpose of estimating the electrical demand of a residential sector in certain periods of time using the Recurrent Neural Network known as Long-Short Term Memory (LSTM). The residential sector of study is Armenia, belonging to the city of Quito in Ecuador; its historical data corresponds to a year between September 2022 and September 2023 taken from primary 23 D. This residential sector is analyzed as a micro-grid in isolated mode, the results of the future estimates of energy consumption allow us to devise the appropriate planning to guarantee the sufficiency of distributed generation based on the new loads obtained and to observe the historical demand data in relation to the new demand data provided by the LSTM network. Knowing the necessary generation that must be supplied, it would be possible to guarantee continuous and uninterrupted energy service for users. The aforementioned estimation case is carried out using the Matlab software to visualize the demand curve in three scenarios, by hours of a day, days of a week and months of a year.