2021 RoEduNet Conference: Networking in Education and Research

Deep Learning for Forecasting the Energy Consumption in Public Buildings

5 Nov 2021, 16:20

20m

Virtual Room A

Paper presentation Technologies for Future Internet Technologies for Future Internet

Speakers

Dr Viorica CHIFU (Technical University of Cluj-Napoca)Dr Cristina Bianca POP (Technical University of Cluj-Napoca)Dr Emil Stefan CHIFU (Technical University of Cluj-Napoca)

Description

In this paper we propose a deep learning based method to forecast the energy consumption in public buildings, based on past measurements. The method integrates two neural networks, namely a Feed-Forward Neural Network and a Long-
Short Term Memory Network. Our approach consists of three main steps: data processing, training and validation, and finally the forecasting step. We validated the method on a data set consisting of measurements taken every half an hour from the main building of the National Archives of the United Kingdom, in Kew.

Presentation materials

RoeduNet-Deep-Learning.pdf