Acta Polytechnica CTU Proceedings https://doi.org/10.14311/APP.2022.38.0340 Acta Polytechnica CTU Proceedings 38:340–346, 2022 © 2022 The Author(s). Licensed under a CC-BY 4.0 licence Published by the Czech Technical University in Prague BIM BASED HYGROTHERMAL MODELLING OF BUILDING CONSTRUCTIONS Balázs Fürtön, Balázs Nagy∗ Budapest University of Technology and Economics, Faculty of Civil Engineering, Department of Construction Materials and Technologies, 3. Műegyetem rkp, H-1111, Budapest, Hungary ∗ corresponding author: nagy.balazs@emk.bme.hu Abstract. The construction industry is currently one of the least digitised. However, due to recent technological developments, building information modelling and management (BIM) systems are increasingly being used to increase the efficiency of constructions and sustainable use of resources and increase the energy efficiency of buildings. In the case of new constructions or building renovations, properly designed hygrothermal behaviour of building materials, elements and structures are unavoidable to achieve the objectives mentioned above. Therefore, we can ensure the comfort of our users and healthy living space. BIM is the most suitable and widespread method for digital management of building stock and related data and integration. With the possibilities offered by BIM, it is possible to integrate building physical modelling right into the architectural design process. However, with standard BIM systems, we cannot embed detailed material properties, weather, or the effects of the natural and built environment in the BIM model for heat and moisture transfer (HAM) modelling for hygrothermal performance evaluation at present. Furthermore, the data exchange process is one-way mainly and unregulated between BIM and numerical simulation tools. In the presented paper, both a Closed BIM and an Open BIM based workflow are presented to bridge the gap between BIM and hygrothermal modelling of building components and constructions, that can contribute to the design of sustainable, economic and hygrothermal optimal building constructions and also could be a basis for artificial intelligence-based optimised design solutions. Keywords: Building information modelling, heat and moisture transfer, numerical simulation. 1. Introduction The construction industry is currently one of the least digitised and among the most fragmented. However, due to recent technological developments, building in- formation modelling and management (BIM) systems are increasingly being used to increase the efficiency of constructions and sustainable use of resources and increase the energy efficiency of buildings. BIM is the most suitable and widespread method for the digital management of building stock. The fragmentation of the AEC industry [1] is present in the building physics focused area as well. Many disciplines should work together until a hygrothermally sound struc- ture is designed. In the case of new constructions or building renovations, properly designed hygrother- mal behaviour of building materials, elements, and structures are unavoidable in order to achieve the aforementioned objectives. Therefore, we can ensure the comfort of our users and healthy living space. With BIM’s potential, it is possible to integrate building physical modelling right into the architectural design process. Though BIM has different flavours, national regulations and software developers shape what it means to work with this methodology. Using standard BIM systems, we usually cannot embed ma- terial properties, weather, or the effects of the natural and built environment in the BIM model for heat and moisture transfer (HAM) modelling for hygrothermal performance evaluation at present. Furthermore, the data exchange process is one-way mainly and unregu- lated between BIM and simulation tools. Even with today’s available BIM systems, performing a HAM simulation is usually a struggle because the software developers present no clear workflow. The possibility of creating multidimensional simu- lations by using commercial software started in 2000 when the Fraunhofer Institute presented WUFI 2D [2]. In 2013, they demonstrated the potential of cou- pling multidimensional thermal bridge simulation with a whole building simulation model in 3D [3]. The use of Comsol to model the hygrothermal behaviour of building constructions is even more modest than 2D or 3D WUFI. Simplified building constructions demon- strate this software’s capabilities [4, 5]. A few papers deal with simulations of wall and slab connections in 2D [6] or hygrothermal modelling of external wall and window joints [7], or in 3D with different concrete structures to examine the effect of different materials or inhomogeneity [8, 9]. BIM can be classified into Open BIM and Closed BIM, based on the availability of the file format struc- ture it is using. Today, Industry Foundation Classes (IFC) is the commonly accepted semantic standard to describe building-related data. It is generally sup- ported in the AEC industry. However, the framework necessary to run hygrothermal simulations does not 340 https://doi.org/10.14311/APP.2022.38.0340 https://creativecommons.org/licenses/by/4.0/ https://www.cvut.cz/en vol. 38/2022 BIM based hygrothermal modelling of building constructions Figure 1. Closed BIM workflow based on Revit. Figure 2. Open BIM workflow based on Archicad. entirely fall into this industry, which means IFC’s support is weak. It is an open, international stan- dard [10]. It promotes vendor-neutral, or agnostic, and functional capabilities across a wide range of hard- ware devices, software platforms, and interfaces for many different use cases [11], offering the possibility to develop an interconnected, accessible data flow among the built assets entire life cycle. Buildings can last for centuries; we do not have enough evidence for the longevity of digital data at this point for a time frame this long. In our research, we created and tested an Open BIM- and a Closed BIM-based workflow for creating HAM simulations to model, evaluate or optimise the hygrothermal performance of building constructions during the integrated design process. Then, we made a comparative assessment between the two possible workflows and summarised the possibilities and limi- tations of the different approaches. 2. Methodology We chose Comsol Multiphysics to run the HAM simulations because it is validated according to EN 15026:2007 Annex A [12] and showed promising re- sults according to other benchmarks, including Ham- stad [13]. The workflows for Closed BIM and Open BIM differ for various reasons. For the Closed BIM workflow solution, we used Revit 2021 for BIM and Comsol 5.6 for numerical sim- ulation to link the BIM model into the Multiphysics environment, as the schematic workflow shows in Fig- ure 1. A Livelink solution for Revit made the initial geometry and material setup relatively easy in Comsol, especially if we consider some techniques not included in the manuals but will be introduced in the present paper later. For the Open BIM workflow solution, we used Archi- cad 24 as the BIM platform. Since there is no stan- dard format between Archicad and Comsol that is reliable enough to transfer geometry, we used Rhino as an interim step to organise and clean the model. Unfortunately, the IFC format cannot be imported into Comsol directly; this is why we needed this step. We used the GeometryGym IFC addon to import the geometry to Rhino, as shown in Figure 2, respectively. An externally insulated wall corner joint consisting of solid masonry blocks and a mineral wool-based external thermal insulation composite system (ETICS) is considered for testing the workflows. 3. Numerical simulations from Closed BIM: Revit-Comsol Livelink Revit is one of the most widespread in the interna- tional BIM software landscape. Therefore, a transpar- ent workflow to perform HAM simulations from Revit models is probably needed for professionals. Comsol is a multiphysical finite element-based engine that lets users examine coupled phenomena, such as conjugated heat and moisture transfer in structures or building materials. The Comsol-Revit Livelink is a proprietary software solution to bridge the gap between BIM and numerical simulation. As such, it is a black box with a bit of hard-to-understand operation. 3.1. Geometry import The primary options for sending the BIM elements are based on spaces or 3D views. For our purposes, the latter proved to be more useful: most of the time, it is sufficient to have only a portion of the structure modelled to run the simulation; we are focusing on this scenario. 3D views can use view templates, so hiding excessive geometry by type and instance is a semi-automatic workflow. Certain geometry types, like planes, are not needed in the numerical model, and cleaning is possible in Comsol, but it is worth avoiding the issues. It is essential to set the “Parts visibility” in Revit to “Show parts” only in the 3D view used to synchronise the BIM model to Comsol to ensure that all set layers are visible in Comsol. This way, building materials are transferred to each layer accordingly (see Figure 3). 341 Balázs Fürtön, Balázs Nagy Acta Polytechnica CTU Proceedings Figure 3. The model from Revit 2021 (left) imported to Comsol 5.6 (right). Figure 4. Building material properties transferred from Revit to Comsol. Using virtual operations on the imported geometry based on BIM is possible in Comsol. It could be useful to handle modelling issues on the geometry, especially to form composite faces or domains from separated materials to improve the meshing performance or ig- nore vertices, faces or domains that are not needed for the HAM simulation, e.g. thin layers that can be handled as surface resistances. 3.2. Building materials We tested various scenarios of how the Livelink can be established: the model needs to be created from native Revit elements (walls, slabs, roofs, etc.), im- ported geometry is not supported as of writing this paper. Using these elements allows us to assign build- ing materials to each geometry. Building materials can be managed in the Material Browser. The necessary building physics parameters are listed in the “Thermal” tab: thermal conductivity, specific heat, and density. Vapour permeability or vapour resistance factor, as well as liquid transport coefficient (or as it is called in Comsol “Diffusion coefficient”) and moisture storage function (“Water content” in Comsol) that necessary for moisture transport calculations, cannot be added to the materials in Revit. There is an option to sync only the geometry without explicitly syncing material data. The Comsol Geometry import workflow has the Livelink for Revit, which can be used instead of man- ually importing from CAD formats (STEP provides a stable workflow that can be exported from Revit natively). Parts should be created in Revit before synchronising the selected view to access each wall layer and its properties. The Livelink process transfers the material proper- ties, and domain selections are created according to parts and materials, which speeds up the preprocess- ing stage of the simulation work. Boundary conditions are not yet stored in the Revit model at an element level. The parameters required to run a hygrother- mal evaluation are not present in the Revit model. They cannot be passed (Figure 4); therefore, we ex- tended them from [14] to be able to perform the HAM simulation. 342 vol. 38/2022 BIM based hygrothermal modelling of building constructions Figure 5. Meshed element originated from BIM. Figure 6. Temperature distribution (left) and relative humidity distribution (right) from HAM simulation based on BIM model. 3.3. Boundary conditions Rooms/Zones and analytical spaces can be present in the model, and the Comsol Livelink sync now works with Revit Rooms. These are abstract, non-existent elements; BIM uses them to specify how space is occupied: hourly schedules often describe the loads, lighting requirements, occupant number, required in- ternal temperature or relative humidity. It could allow the generation of boundary conditions for the inter- nal surfaces of the numerical model. As the model is referenced to real-world coordinates, solar radia- tion could be added to the external boundary condi- tions apart from external temperature and relative humidity. Small scale models often do not include broader context, but at a façade- or whole building level, numerical simulation shading from neighbouring buildings is an important aspect of the design. During our study, boundary conditions were added according to [12, 14] in Comsol as heat flux and moisture flux boundary conditions. 3.4. Meshing The tested insulated wall corner joint has meshed in Comsol after the import. The model is meshed using a physics-controlled meshing sequence and extra-fine element size in the software that gave us an average element quality of skewness of 0.6627 without errors and warnings (Figure 5). The different thicknesses of the layers in the model affect the built-in meshing functions since there are considerably significant differ- ences between the thicknesses of the domains included in the model. The element quality can be improved with manual mesh refinement, solution-based adaptive meshing, or previously mentioned virtual operations. 3.5. Physics and results We set the physics in Comsol 5.6 to Heat and Moisture Transport in Building Materials, which models the HAM transfer according to [12]. A stationary solver was deployed for the coupled equations, geometric multigrid for the moisture and smoothed aggregation geometric multigrid for heat transfer. The results using the extended material properties and bound- ary conditions are shown in Figure 6. Since we put mineral wool thermal insulation on our masonry wall, the relative humidity distribution is high since it lets the vapour through. Besides analysing temperature and relative humidity distribution, HAM analysis also 343 Balázs Fürtön, Balázs Nagy Acta Polytechnica CTU Proceedings Figure 7. Stages of model conversion from Archicad to Comsol. allows us to check the temperature factor at internal surface points or evaluate the thermal transmittance and the moisture transmittance of building compo- nents. The structure’s heat flux or moisture fluxes can also be visualised for surfaces, domains and sections, respectively. 4. Numerical simulation with Open BIM: Archicad-Comsol Relying on open standards provides accessibility to and control over simulations’ input and output data. Finding errors or workflow issues is easier if we can track the path of each piece of information stored initially in the model. In BIM, the only widespread open standard for comprehensive communication of stored data is IFC. IFC is an international semantic standard maintained by BuildingSMART, with ex- tensive technical specifications covering the materials’ necessary building physics properties. 4.1. Geometry As there is no direct support for neither Archicad’s proprietary data model nor the openly available IFC format, the Open BIM workflow is more complicated because we need an intermediary step to convert the geometry into formats that the chosen Multiphysics platform (Comsol) can handle. For this reason, we decided to choose Rhinoceros 7 because the STEP export from it towards Comsol proved to be robust, and it has an extensive ecosystem to extend its capa- bilities. There are published examples of the Rhino to Comsol workflow previously used for building envelope analysis [15]. The GeometryGym addon lets users import and manipulate the IFC files to Rhino, then export it as STEP towards Comsol. The problem is that although material level data can be stored (even extended) in Archicad and Rhino can display it; we couldn’t find a way to attach non-geometric information to the STEP geometry apart from layer information (Fig- ure 7). Unfortunately, building material properties are lost at the export stage from Rhino to Comsol. In order to get accurate representation, it is advised to export the geometry as non-parametric BREP objects, exploding composite structures to their layers in the IFC Translation settings to ensure that everything is correct. With parametric export, the pipeline can suffer down the line, as the import capabilities of other applications are externalities from the initial export’s perspective. 4.2. Building materials This means that setting up all the properties in Archi- cad for the sake of hygrothermal evaluation is not feasible (for quantity take-offs, mass calculation, cus- tom GDL scripting they are useful to include, as the REQUEST command can be used to obtain building physics parameters and do calculations in a custom library part). With recent releases, it is possible to assign custom physical properties to building materi- als, but they are constant values only, as the default Physical properties (Figure 8). These values are ex- ported in the IFC file, but we did not take care of it since they cannot be used for the last step of the conversion process. Material parameters need to be done from scratch in Comsol, which isn’t much work. The problem occurs when we need to assign materials to the objects (this process is automatic with the help of the Livelink), it is dependent on the Import workflow how selections are stored. Colours and layers are the only attributes that can be converted to domain selections. 4.3. Meshing, physics and results These processes are exactly the same and should be handled as introduced in Sections 3.3, 3.4 and 3.5, respectively. Since the geometry, mesh, and boundary conditions are the same as described in the Closed BIM process, and simulation results are also the same. 5. Conclusions In general, it can be stated that there isn’t a satisfy- ing workflow present in the current market. Neither Closed BIM nor Open BIM solutions can provide the necessary granularity in building physics character- istics that we would require and are still far from the detailed HAM modelling possible in Comsol [14]. However, we succeeded and performed simple HAM simulations based on BIM models both from Closed BIM and Open BIM processes. The conclusions are summarised in Table 1. 344 vol. 38/2022 BIM based hygrothermal modelling of building constructions Figure 8. Building material catalogue in Archicad with the default available physical parameters. Process Closed BIM-based HAM Open BIM based HAM Passing geometry from BIM to HAM No loss of detail Passing building physics parameters from BIM to HAM Limited, depending on the input data. With simpler geometry and parts created, lossless connection No interface is provided for IFC from HAM, although it is a STEP based file format Creating domains for faster HAM model preprocessing Part- and material based selections help the work Layer or colour based indirect sorting is possible Boundary conditions No interface is provided as of now; manual work needed Meshing Input can be meshed without errors, if sufficiently clean Feedback to BIM from HAM Early-stage development for massing – geometry control Not implemented Table 1. Summary of the possibilities with BIM to HAM workflows. Acknowledgements Supported by the ÚNKP-21-4 New National Excellence Program of the Ministry for Innovation and Technology from the source of the National Research, Development and Innovation Fund. Project FK_128663 has been im- plemented with the support provided from the National Research, Development and Innovation Fund of Hungary, financed under the FK_18 funding scheme. References [1] D. Hall, A. Algiers, T. Lehtinen, et al. The role of Integrated Project Delivery elements in adoption of integral innovations. In Proceeding of EPOC 2014 Conferenence. 2014. 20 p. [2022-03-15]. https://epossociety.org/epoc-2014 [2] A. Holm, H. M. Künzel. Two-dimensional transient heat and moisture simulations of rising damp with WUFI 2d. In Proceedings of the 12th International Brick/Block Masonry Conference, vol. 2. 2000. 8 p. [2022-03-15]. https://wufi.de/en/literature/ [3] F. Antretter, M. Pazold, J. Radon, H. Künzel. Kopplung von dynamischer Wärmebrückenberechnung mit hygrothermischer Gebäudesimulation. Bauphysik 35(3):181–192, 2013. https://doi.org/10.1002/bapi.201310068 [4] A. W. M. van Schijndel. Multiscale heat, air and moisture modelling and simulation. In G. Petrone, G. Cammarata (eds.), Recent Advances in Modelling and Simulation, chap. 41. 2005. https://doi.org/10.5772/5963 [5] M. Janetti, F. Ochs, W. Feist. 3D simulation of heat and moisture diffusion in constructions, 2011. [2022-03-15]. https://www.comsol.com/paper/3d- simulation-of-heat-and-moisture-diffusion-in- constructions-10921 345 https://epossociety.org/epoc-2014 https://wufi.de/en/literature/ https://doi.org/10.1002/bapi.201310068 https://doi.org/10.5772/5963 https://www.comsol.com/paper/3d-simulation-of-heat-and-moisture-diffusion-in-constructions-10921 https://www.comsol.com/paper/3d-simulation-of-heat-and-moisture-diffusion-in-constructions-10921 https://www.comsol.com/paper/3d-simulation-of-heat-and-moisture-diffusion-in-constructions-10921 Balázs Fürtön, Balázs Nagy Acta Polytechnica CTU Proceedings [6] J. Berger, W. Mazuroski, N. Mendes, et al. 2D whole-building hygrothermal simulation analysis based on a PGD reduced order model. Energy and Buildings 112:49–61, 2016. https://doi.org/10.1016/j.enbuild.2015.11.023 [7] B. Nagy, C. Tömböly. Hygrothermal analysis of window construcion joints. In Zbornik radova Građevinskog fakulteta 34, pp. 91–100. 2018. https: //doi.org/10.14415/konferencijaGFS2018.008 [8] F. Colangelo, G. De Luca, C. Ferone, A. Mauro. Experimental and numerical analysis of thermal and hygrometric characteristics of building structures employing recycled plastic aggregates and geopolymer concrete. Energies 6(11):6077–6101, 2013. https://doi.org/10.3390/en6116077 [9] P. Iodice, N. Massarotti, A. Mauro. Effects of inhomogeneities on heat and mass transport phenomena in thermal bridges. Energies 9(3):126, 2016. https://doi.org/10.3390/en9030126 [10] International Organisation for Standardisation. ISO 16739:2018. Industry Foundation Classes (IFC) for data sharing in the construction and facility management industries. [11] BuildingSMART: Industry Foundation Classes. [2022-03-15]. https://www.buildingsmart.org/standards/bsi- standards/industry-foundation-classes/ [12] European Committee for Standardisation. EN 15026:2007. Hygrothermal performance of building components and building elements – Assessment of moisture transfer by numerical simulation. [13] J. I. Knarud, S. Geving. Implementation and benchmarking of a 3D hygrothermal model in the COMSOL multiphysics software. Energy Procedia 78:3440–3445, 2015. https://doi.org/10.1016/j.egypro.2015.12.327 [14] B. Nagy, G. Stocker. Numerical analysis of thermal and moisture bridges in insulation filled masonry walls and corner joints. Periodica Polytechnica Civil Engineering 63(2):446–455, 2019. https://doi.org/10.3311/PPci.13593 [15] J. Hand. Optimising the performance of complex building facades, 2015. In Comsol News. [2022-03-15]. https://www.comsol.com/story/download/309771/ Newtecnic_CN2015.pdf 346 https://doi.org/10.1016/j.enbuild.2015.11.023 https://doi.org/10.14415/konferencijaGFS2018.008 https://doi.org/10.14415/konferencijaGFS2018.008 https://doi.org/10.3390/en6116077 https://doi.org/10.3390/en9030126 https://www.buildingsmart.org/standards/bsi-standards/industry-foundation-classes/ https://www.buildingsmart.org/standards/bsi-standards/industry-foundation-classes/ https://doi.org/10.1016/j.egypro.2015.12.327 https://doi.org/10.3311/PPci.13593 https://www.comsol.com/story/download/309771/Newtecnic_CN2015.pdf https://www.comsol.com/story/download/309771/Newtecnic_CN2015.pdf Acta Polytechnica CTU Proceedings 38:340–346, 2022 1 Introduction 2 Methodology 3 Numerical simulations from Closed BIM: Revit-Comsol Livelink 3.1 Geometry import 3.2 Building materials 3.3 Boundary conditions 3.4 Meshing 3.5 Physics and results 4 Numerical simulation with Open BIM: Archicad-Comsol 4.1 Geometry 4.2 Building materials 4.3 Meshing, physics and results 5 Conclusions Acknowledgements References