PRENERGET, framework for energy forecasting

• Meritxell Gomez-Omella ^[1] ; Susana Ferreiro ^[1] ; Alvaro Garcia ^[1] ; Francisco Javier Díez Trinidad ^[1]
  1. [1] TEKNIKER, Basque Research and Technology Alliance (BRTA), Iñaki Goenaga 5, Eibar, 20600, Gipuzkoa, Spain
• Localización: EESAP13 International Conference 2022, 5-6 October Donostia-San Sebastián: Akten liburua = Libro de actas = Proceedings book / Escuela Técnica Superior de Arquitectura de San Sebastián (dir. congr.), 2022, ISBN 978-84-1319-499-8, págs. 10-19
• Idioma: inglés
• Enlaces
  • Texto Completo Libro
• Resumen

  • For the optimisation of energy management, it is crucial to be able to make decisions in advance. For this decision making it is necessary to have reliable predictions. In a building, there can be different types of predictions related to energy management; demand, production, temperature, price, occupancy, etc. Machine learning algorithms are a good technology to make these predictions, but they must be adapted to each variable and context, so many algorithms are needed running in parallel. In this paper a framework that allows to facilitate the execution of these predictions is presented. The boundary conditions on which these algorithms are based change over time and the predictions become less reliable. The presented framework allows to adapt to these changes in order to maintain the reliability of the predictions.