Back to search

Technical University of Crete (TUC)

Development of methodologies for automatic thermal model generation using building information models

Abstract

dc:description

Following the recent requirement for efficient allocation of energy resources in the buildingsector, the use of Building Energy Performance (BEP) simulations became moreand more frequent during the design as well as the operation phases. However, BEPsimulation models’ generation require significant effort for set-up, limiting the potentialutilization of modeling in both phases. The substantial effort stems from the difficulty tocollect and appropriately define the input data for accurate BEP simulation modeling,which are categorized as follows: static data, containing information on the geometry,construction and actuating energy systems; and dynamic data, including factual (sensed)data, along with forecasts for pertinent parameters (e.g. weather, occupancy), used tobridge the “simulated” and “real” worlds, reducing or even mitigating uncertainties.Concerning the static data, a methodology for semi-automated BEP (known also asthermal) simulation model creation could make the BEP simulation modeling processmuch more expedient and as such lower to threshold for the use of such models. BuildingInformation Models (BIMs) are information-rich repositories that could be used tostreamline and expedite the collection of such information. The Industry FoundationClasses (IFC) BIM schema provide static building information that include geometricconfiguration and material properties, but in a form that might not be directly usablefor the generation of thermal simulation models due to the absence of 2nd-level spaceboundaries information. Even if 2nd-level space boundaries information exists, an IFCto a specific BEP simulation engine’s input data mapping must be performed. Towardsmodeling an automatic data mapping process, the modeler must have expert knowledgeof the IFC and the specific BEP simulation engine input data structures. A precise datamapping can lead to a BEP model’s geometry of high detail. Here, another recurringtheme arises that the modeler has to address during the BEP simulation models’ geometrydata generation: the decision on how accurate the model should be. For Control Design tasks, the accuracy requirements are markedly different, compared to the simulationmodel for energy auditing purposes. Control-Design processes require a model thatis able to capture the sensitivities and trends but no high accuracy is necessary. Concerningthe dynamic data, for any BEP simulation related task, the availability of senseddynamic data in a building presents an opportunity to eliminate with minimal effort theuncertainties of the BEP model. Moreover, in-situ weather measurements contributeto the development of an accurate BEP model without increasing the computationalcomplexity. These measured weather data are forwarded to the model through specificallyformatted weather files. However, it is quite common that weather data do notinclude information concerning solar radiation, but only a percentage of cloud coverage.Even if global radiation data are available, a crucial input in the simulation of building’senergy performance is the availability of both diffuse and direct solar radiationdata. Nevertheless, in most cases, measured data of diffuse and direct radiation are notavailable.In this thesis, initially a methodology is presented for (semi-) automated generationof thermal simulation models’ static data, including: a query on the building datamodel — embedded in the relative IFC file — requesting geometry-related information;a processing of the acquired data by a 2nd-level space boundary identification algorithm,the Common Boundary Intersection Projection (CBIP) algorithm; and a transformationprocess that converts the geometry information of IFC, along with the data obtainedfrom the CBIP algorithm, to an EnergyPlus and/or TRNSYS — widely used BEPsimulation engines — input file. Next, commonly the BEP simulation model’s geometryderived by applying the aforementioned methodology, is of high detail due to numeroussurfaces and thermal zones, increasing the simulation runtime, and as such not suitablefor a Control Design process. Hence, simulation speed-up approaches are investigated,that focus on geometry and/or zoning reduction, in a way that reduces complexity whilemaintaining features of the simulation. Establishing a link between the BEP simulationstatic data and the building’s sensed measurements so that the dynamic schedules canbe incorporated in the simulation, is of paramount importance. This link, especiallydesirable also for the testing and design of control strategies, is adopted, following awidely used co-simulation framework, the Buildings Control Virtual Test Bed (BCVTB).Finally, among a plethora of methods for estimating the diffuse (or direct) fraction ofglobal solar radiation, the question of the best method has been not fully settled, a taskthat this thesis tries to address.The methodology for (semi-) automated generation of thermal simulation models’ staticdata is applied to real buildings and its results are presented, highlighting the ability inhandling non-convex geometries and generating all the possible BEP simulation model’sgeometry data. The methodology facilitates significantly the overall process of energy simulation model creation from IFC geometric data. Next, the simulation speed-up approaches are evaluated with respect to accuracy and computational effort, using threetest buildings. Moreover, among a plethora of existing experiments on real buildings,one is chosen to be presented highlighting the necessity of the co-simulation setup towardseliminating the building’s uncertainties. Concerning the solar radiation models’efficiency, results of our investigation are presented, indicating that there is no methodlargely better than the others.

Degree

thesis:*
Grantor dc:publisher
Technical University of Crete (TUC)
Year dc:date
2015

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Giannakis, Georgios
  • Γιαννάκης, Γεώργιος

Subjects

dc:subject × 9

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/35981

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Giannakis, Georgios; Γιαννάκης, Γεώργιος. Development of methodologies for automatic thermal model generation using building information models. Technical University of Crete (TUC), 2015. http://hdl.handle.net/10442/hedi/35981