IS IT POSSIBLE to use a SINGLE REDUCED MODEL for a NUMBER OFBUILDINGS in URBAN ENERGY SIMULATION?

Proceedings of BS2015: 14th Conference of International Building Performance Simulation Association, Hyderabad, India, Dec. 7-9, 2015

IS IT POSSIBLE TO USE A SINGLE REDUCED MODEL FOR A NUMBER OF BUILDINGS IN URBAN ENERGY SIMULATION?

Eui-Jong Kim1, Xi He2, Jean-Jacques Roux3, Kévyn Johannes3, Frédéric Kuznik3
1Department of Architectural Engineering, Inha University, Incheon, Korea
2Department of Civil Engineering, Tongji University, Shanghai, China,
3Université de Lyon, INSA-Lyon, CETHIL, Villeurbanne, France;

ABSTRACT

Roles of urban energy simulation become important for sustainable city design and energy management.

In this context, a single reduced model that can describe a number of buildings is proposed and tested in this paper. Simulation results show that the proposed single reduced model for multiple buildings can give very similar results with separate reduced models in load calculation. However, the calculation time of the single reduced model is more important
than the case of several reduced models because the PID controller used in Dymola to calculate building loads requires more iterations for a MIMO (multi inputs and multi outputs) system than a SISO (single input and single output). In conclusion, the use of a single reduced model can be a solution for urban energy simulation, but further work is needed to simplify calculation processes by an optimal solver.

INTRODUCTION

Precise urban energy simulation is a rising issue in the domain as it can provide plenty of valuable information for sustainable city design. A proper modelling of buildings, each of which plays a role as an energy node, is crucial to predict accurate city
energy consumption and to optimize energy uses in the city.

Several modelling methods of different levels of detail have been developed for this purpose. A typical method is the heuristic approach that city empirical data are analysed and a simple statistical model is adopted. In this category, many uncertainty analysis methods have been used (Morris et al., 2014). As this approach cannot directly reflect changes in materials and systems of buildings, the grey-box model approaches have been developed (Déqué et al., 2000, Plessis et al., 2011). The validity of this approach is strongly affected by the building typology selected for the model development.

Recently, a simple direct model has been developed for bottom-up city modelling. An attempt is the development of the urban simulation tool CITYSIM (Robinson et al., 2009). In this tool, lumped RC models are used for buildings in a district scale. On
the other hand, building models in detailed commercial simmulation tools have been directly used for urban energy simulation (Courchesne-Tardif et al., 2011.). This is the most detailed approach among current urban simulation methods, but cosiderable computational resources and efforts are required.

In this context, a reduced model for a single building was previously proposed that uses fewer equations and at the same time keeps the high level of detail found in a detailed building simulation model. Since each reduced model represents a single building, n reduced models must be defined and simulated for n buildings in a city considered. However, if a single reduced model can describe a number of buildings, the model can be very useful to rapidly optimize city systems, e.g. district heating networks, without requiring important computational resources.

This possibility is tested in the paper. Above all, the component models of the previous model are replaced by ones of an open source model library for wider use. Then, a number of building models are assembled and transformed into a single model, and its performance is compared to the initial reference model.

READ the PROCEDING --->