REDUCED ORDER MODELLING OF THE THERMAL BEHAVIOUR OF AN OFFICE SPACE

Marco Geron, R.F.D Monaghan, M. Keane

Research output: Contribution to conferencePaper

Abstract

Reduced Order Models (ROMs) have proven to be a valid and efficient approach to model the thermal behaviour of building zones. The main issues associated with the use of zonal/lumped models are how to (1) divide the domain (lumps) and (2) evaluate the pa- rameters which characterise the lump-to-lump exchange of energy and momentum. The object of this research is to develop a methodology for the generation of ROMs from CFD models. The lumps of the ROM and their average property values are automatically ex- tracted from the CFD models through user defined constraints. This methodology has been applied to validated CFD models of a zone of the Environmental Research Insti- tute (ERI) Building in University College Cork (UCC). The ROM predicts temperature distribution in the domain with an average error lower than 2%. It is computationally efficient with an execution time of 3.45 seconds. Future steps in this research will be the development of the procedure to automatically extract the parameters which define lump-to-lump energy and momentum exchange. At the moment these parameters are evaluated through the minimisation of a cost function. The ROMs will also be utilised to predict the transient thermal behaviour of the building zone.
Original languageEnglish
DOIs
Publication statusPublished - 04 Sep 2013
EventCISBAT : Cleantech for Smart Cities & Buildings - From Nano to Urban Scale - Lausanne, Switzerland
Duration: 04 Sep 201306 Sep 2013

Conference

ConferenceCISBAT
CountrySwitzerland
CityLausanne
Period04/09/201306/09/2013

Fingerprint Dive into the research topics of 'REDUCED ORDER MODELLING OF THE THERMAL BEHAVIOUR OF AN OFFICE SPACE'. Together they form a unique fingerprint.

  • Cite this

    Geron, M., Monaghan, R. F. D., & Keane, M. (2013). REDUCED ORDER MODELLING OF THE THERMAL BEHAVIOUR OF AN OFFICE SPACE. Paper presented at CISBAT , Lausanne, Switzerland. https://doi.org/10.5075/epfl-infoscience-190600