Development of Scalable Whole-Building Energy Models under Zero Commissioning and Zero Maintenance Constraints
Biography: Dimitrios-Stavros graduated from Aristotle University of Thessaloniki in 2009 with a Diploma in Mechanical Engineering. He continued his studies and graduated from University of Southampton in 2011 with an MSc in Sustainable Energy Technologies. His research projects were “Retrofit measures for residential multi-family buildings in the urban existing building stock” and “Sizing of Hybrid PV-Wind Energy Systems” for the undergraduate and postgraduate studies respectively. After graduation he worked as a mechanical engineer at e2 Architects, taking part in projects related with renewable energy, such as dynamic energy simulation of buildings, installation of green roofs, design and supervision of PV installation on residential roofs and agriculture fields and issued certificates of energy efficiency for buildings He commenced his PhD in UCD in February 2013, within the Electricity Research Centre (ERC). His research interests include measured data analysis, machine learning models, building modeling and energy efficiency.
Dimitrios-Stavros Kapetanakis is funded by the Irish Research Council and UTRC.
Project: This project proposes a novel approach, using a data-driven methodology, based on the application of predictive machine learning techniques. The approach should be capable of predicting the thermal load of any commercial building using historical BEMS data. Two machine learning approaches, Artificial Neural Networks and Support Vector Machines are being considered. Benchmarking against regression models is carried out. In order to develop a robust methodology, synthetic data from different types of commercial buildings for various climates are utilised. The most suitable approach will be selected by comparing the accuracy of the predictions. The scalability of the model will be examined by applying the technique to a number of demonstration commercial buildings located in different geographical locations.
Keywords: commercial, buildings, thermal load, cooling load, prediction, accuracy, scalability, zero commissioning, zero maintenance
Dimitrios-Stavros Kapetanakis, Eleni Mangina, Donal Finn,
Conference Paper Angers, France, 6 Mar. 2014 published on 06/03/2014