Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/148208
|Análisis del consumo de la energía de un edificio mediante técnicas predictivas de Deep Learning
|García de la Chica Herrera, Ángel
|Consumption reduction and increased energy efficiency are essential to combat climate change. This project analyses the use of Deep Learning technologies to predict the electrical energy consumption of buildings. In this way, we will be able to improve efficiency and reduce electricity consumption. Firstly, the project introduces the different Deep Learning systems and the different approaches that currently exist to predict energy consumption. Secondly, an analysis of electricity consumption data from multiple buildings is carried out together with historical meteorological data from the meteorological station closest to the building. For this purpose, normalisation techniques, discretisation and treatment of outliers and missing values are applied. On the other hand, a correlation analysis of the attributes is carried out to reduce the dimensionality by means of PCA (Principal Component Analysis). Then, different models are implemented using two different concepts. On the one hand, artificial neural networks and on the other hand, LightGBM (Light Gradient Boosting Machine) decision trees. Finally, the results of the different models are compared using the mean square error (MSE). This work concludes that it is possible to predict the electrical energy consumption of buildings from meteorological data and building metadata. Moreover, the models implemented for a single building do not require much capacity and computational time with good accuracies.
|Appears in Collections:
|Bachelor thesis, research projects, etc.
Files in This Item:
|Memoria del TFG
|Presentación del TFG