Automation in Construction 91 (2018) 166–181

Contents lists available at ScienceDirect

Automation in Construction
journal homepage: www.elsevier.com/locate/autcon

An automated IFC-based workflow for building energy performance T

simulation with Modelica
Ando Andriamamonjya,b,*, Dirk Saelensa,c, Ralf Kleinb
a
KU Leuven, Department of Civil Engineering, Building Physics Section, Kasteelpark Arenberg 40 – box 2447, Heverlee BE-3001, Belgium
b
KU Leuven, Technology Cluster Construction, Technology Campus Ghent, Gebroeders De Smetstraat 1, Ghent BE-9000, Belgium
c
EnergyVille, Thor Park 8310, Genk BE-3600, Belgium

A R T I C LE I N FO A B S T R A C T

Keywords: Steadily increasing use of Building Information Modeling (BIM) in all phases of building's lifecycle, together with
BIM more attention for openBIM and growing software support for the most recent version of the Industry
IFC4 Foundation Classes (IFC 4) have created a very promising context for an even broader application of Building
Energy efficiency Energy Performance Simulation (BEPS). At the same time, an urgent need for modeling guidelines and stan-
Building performance simulation
dardization becomes evident. A well-defined BIM-based workflow and a set of tools that fully exploit and extend
Building life cycle
the possibilities of the openBIM-technology can make the difference when it comes to reliability and cost of BEPS
to design, build and operate high-performance buildings. This paper describes the essential elements of this
integrated workflow, explains why openBIM comprises much more than just a standardized file-format and what
is achieved with the already available technology, namely the Information Delivery Manual (IDM) and a newly
developed Model View Definition. This MVD is tailored to the needs of Building Energy Performance Simulation
(BEPS) that uses the Modelica language together with a specific library (IDEAS) and can easily be adapted to
other libraries. In this project, several tools have been developed to closely integrate BEPS and IFC4. The si-
mulation engine now gets the vast majority of the required input directly from the IFC4-file. For the im-
plementation of the tools, the PYTHON language and the open source library IfcOpenShell are used. A case study
is presented, that was used for extensive tests of the proposed approach and the implemented tools. The essential
benefits of this new workflow are illustrated, and the feasibility is demonstrated. Opportunities and remaining
bottlenecks are identified to encourage further development of BIM software to fully support IFC4 as an in-
formation source for BEPS. Besides some improvements of the proprietary class structure and functionality,
enabling the export of IFC4 files based on custom MVDs is one required key feature.

1. Introduction of physical and functional characteristics of a building project [10, 11].

It is an object-oriented information model incorporating data such as
Building Energy Performance Simulation (BEPS) models of different 3D geometry, spatial relationships, geographic information, quantities
detail levels enable to improve building performance and comfort [1, and building element properties [12-14]. Also, BIM integrates and
2]. Traditional simulation engines such as EnergyPlus [3] or unifies the communication among stakeholders through a shared, and
TRNSYS [4] are often used for detailed and transient simulations. collective data model, potentially enhancing decision making and work
However, their lack of modularity and flexibility for prototyping and efficiency [11, 12]. Although collaborative, technological and legal
simulating new technologies has driven the emergence of the modular, challenges need to be addressed [15, 16], BIM adoption has the po-
equation-based simulation engine: Modelica [5, 6]. Independently of tential to improve construction coordination, productivity, cost esti-
the simulation engine type, BEPS implementation is labor-intensive, mate accuracy while reducing clashes (up to 10%), omissions, con-
time-consuming and error-prone [1, 7, 8]. A large part of the required struction time (up to 7%) and overall project cost [10, 16-19]. Recently,
information could be sourced under certain conditions directly from the use of BIM as an information source for BEPS has gained momentum
Building Information Models (BIM). [1, 8, 20-28]. Three types of solutions can be distinguished:
BIM, an emerging technology [9] in the Architecture Engineering
and Construction (AEC) industry, is defined as a digital representation 1. Integration of a simulation engine into the BIM tool or direct

*
Corresponding author.
E-mail addresses: [email protected] (A. Andriamamonjy), [email protected] (D. Saelens), [email protected] (R. Klein).

https://doi.org/10.1016/j.autcon.2018.03.019
Received 12 September 2017; Received in revised form 6 February 2018; Accepted 8 March 2018
0926-5805/ © 2018 Elsevier B.V. All rights reserved.
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

coupling using the Application Programming Interface (API) of such simulation engines and the need for a variable level of detail of the
software. BEPS models for the different life-cycle stages. Expectations concerning
2. Export of the relevant information from the BIM tool to a file using the error margins of the results must also be kept realistic. Altogether,
the gbXML format and subsequent import of that file into the si- these conditions have driven the recent use of the flexible, modular
mulation software. The Green Building XML schema (gbXML) was modeling language Modelica for energy simulations.
created for the information transfer from Autodesk REVITTM to the The successful integration of IFC and Modelica-based BEPS requires
online BEPS engine of Green Building Studio. gbXML is a flexible, a reliable and secure data exchange between BIM and BEPS tools by
open, straightforward data schema and was later adopted by several using guidelines and software certification based on specific MVDs.
software vendors as a de facto standard for the information transfer Pinheiro et al. [48, 49, 79] present a first step addressing this need
from BIM to BEPS [29]. However, the gbXML schema has limited where they propose an MVD for advanced BEPS. They also identify
geometry definition capabilities [30, 31] (e.g. only rectangular potential improvements over the current IFC schema release regarding
shapes, based on element centerline) that can lead to an over- energy simulation.
estimation of the building energy consumption [32]. A significant amount of input data is obtained directly from BIM,
3. Export of the entire BIM (or preferably the relevant parts of it) to the but the BEPS expert still needs to take additional decisions and insert
Industry Foundation Classes (IFC) format [8, 20-22, 25, 26, 28] and new parameters. This task will typically happen in the simulation en-
subsequent import of that file into the simulation software. Devel- vironment and is subject to the BEPS expert interpretation which can
oped and maintained by BuildingSMART [33], IFC was officially introduce modeling errors [50]. Thereby, the decisions should be
defined in ISO 16739 and adopted by CEN as EN ISO 16739. documented and communicated together with the results while a user-
Moreover, the BuildingSMART openBIM [33] approach comprises friendly BEPS modeling interface would limit the risk of mistakes. Apart
two other important aspects defined in ISO standards: mapping of from that, the difference between control strategy assumptions during
terms in the form of IFD (International framework for dictionaries, simulations and the proprietary ones implemented in device controllers
ISO 12006) and IDM (Information Delivery Manual, ISO 29481), a contributes to the performance gap. Controls are a crucial link between
formal process description for the implementation of an openBIM occupancy, climate data, building characteristics and system response,
workflow in organizations, for projects and as reference for software and become increasingly important in reducing energy consumption
certification, based on Model View Definitions (MVDs). Two tech- and peak electrical power demand. Thereby, it is essential to specify
nical committees (CEN/TC 442 [34] and ISO/TC 59/SC 13 [35]) are controls in BIM to avoid inconsistency between design and operation
actively developing the further standardization of openBIM. [6, 51]. Future IFC to (Modelica-)BEPS approaches need to have the
capability to translate control representations in BIM into BEPS.
Jeong et al. [23, 24] implemented an example of the solution type 1, In the frame of IEA Annex 60 activity 1.3, Remmen et al. [28, 52,
where the Autodesk REVITTM API is coupled with the Modelica Build- 53] have initiated the interfacing of IFC with Modelica. They propose a
ings library [36]. The lack of interoperability and the dependence on a semi-automated translation of system descriptions encapsulated in IFC
specific version of a proprietary software constitute the main drawback into the Modelica library AIXLIB by using SimModel as an intermediate
of this approach. data model [54]. Their methodology relies on the existing IFC into
The second type of solution can be found in commercial software, SimModel interface (used in SimergyTM) and adds the step to map
e.g. DesignBuilderTM [37] and IES-VETM [38] or the open source soft- SimModel into Modelica.
ware tools collection OpenStudio [39], supporting the use of En- An advantage of using an intermediate data model is the possibility
ergyPlus and Radiance. OpenStudio also partially supports the third to insert information at this stage that might not be available within
type solution by providing IFC-import for some objects of the building IFC. A direct mapping, on the other hand, imposes that IFC incorporates
envelope. all BEPS-related information (geometry, system, and control) before-
Most of the third type solutions focus on the geometry representa- hand. The latter approach enables a direct and automated IFC to
tion [25, 40, 41]. One of the early issues was the monolithic description Modelica translation but requires a more stringent BIM modeling pro-
of construction elements (e.g. walls, slabs, ceiling) incompatible with cess and a formal data exchange.
the one-dimensional heat transfer in simulations [22]. Bazjanac et al. The research presented in the current publication investigates the
addressed this limitation by proposing the space boundary concept [42, direct translation of BIM to BEPS and proposes an improved workflow
43]. However, the IFC standard did not include the latter concept that can perform a direct mapping of geometry, system and control
correctly before IFC version 4, which led to the development of dif- representations contained in an IFC4 file to a Modelica-based BEPS
ferent space boundary tools [44-46]. Another aspect of the third type model. The flexibility of Modelica facilitates the extension of the
solution is the translation of building systems information into BEPS. workflow to different building life cycle stages with different require-
Bazjanac and Maile [21] designed the IFC HVAC to EnergyPlus utility ments for model complexity. This work emphasizes the advantages and
which can generate an EnergyPlus HVAC input (IDF) from an IFC2×2 describes remaining bottlenecks of such a workflow.
and vice versa. However, some information still has to be added The key components of the workflow are:
manually to obtain a functional IDF file. Later, O’Donnell et al. [47]
developed SimModel to tackle the incompatibility between IFC and 1. A BIM guideline and a Model View Definition (MVD, see Section 4)
EnergyPlus input. SimModel is an intermediate data model (currently ensure that the IFC file incorporates all relevant BEPS information,
used in SimergyTM) designed to facilitate data exchange for building and allow to automatically check the file's integrity during the ex-
energy simulation. change process.
Most of the published methodologies solely focus on one specific 2. A novel, PYTHON based IFC to Modelica adapter (hereafter referred
part of the translation (e.g. geometry or HVAC) and use traditional si- to as Ifc2Modelica - Section 6) generates the building geometry, the
mulation engines (e.g. EnergyPlus) that require an amount of data only HVAC components and systems and especially the control BEPS
available at the detailed design stage. Bazjanac et al. [7] assessed the model from an IFC file compatible with the proposed BIM guideline.
use of EnergyPlus at the early design and evaluated the impact of in- The resulting BEPS model is graphically represented in the user
sufficient reliable data that can potentially lead to wrong conclusions. interface of the Modelica simulation engines (e.g. DymolaTM,
Moreover, the complexity of a detailed BEPS model is not compatible OpenModelica) to facilitate the integration of simulation related
with the need of the operational phase (regarding simulation time), e.g. parameters (e.g. type of Air model, solver, initial condition, weather
for Fault Detection and Diagnosis (FDD) or Model Predictive Control data) typically not (yet) available in BIM.
(MPC). These examples underline the lack of flexibility of traditional 3. Modelica requirements (Section 3) define how Modelica libraries

167
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

shall be implemented to be compatible with the proposed workflow. lifecycle [55] especially during the design stages. In the “preliminary
These constraints extend the applicability of the workflow to all concept design” stage (see Fig. 1.b), a simplified BIM comprises the
stages of the entire building lifecycle by implementing libraries building footprint, including the exterior walls with fenestration as a
which suit the particular needs of each building lifecycle stage. percentage of wall area. In the course of the “final concept design”, BIM
is enriched with the detailed building elements (e.g. windows, internal
This workflow is hereafter referred to as BII (BIM-IFC- walls, floors) and a simplified system description. Later in the “design
Ifc2Modelica). Prior to a more detailed presentation of these compo- development” (Fig. 1.b), BIM incorporates the detailed description of
nents in Sections 3 to 6, Section 2 provides an overview of BII and its the architectural elements, systems and controls which are subse-
application for the “detailed design” stage. In Section 7, a proof of quently finalized in the “detailed design” stage. In any of these stages, a
concept application is presented. The facility of this case study is a BEPS analysis can be conducted [56]. For instance, during the “pre-
university building with two lecture rooms and it is equipped with a liminary concept design” stage, a BEPS enables to identify the different
variable air volume air handling unit. Two scenarios are considered to design options. While during the “detailed design” stage, a BEPS model
emphasize the advantages of BII and to show also remaining limita- can be used for energy performance certification.
tions, currently preventing a more general adoption. BII applies to the entire building lifecycle but its application will be
demonstrated here for the detailed design stage (see Fig. 1.a and .b),
and the necessary changes for its use in other stages will be described.
2. Context and general approach BII can generate a Modelica BEPS model compatible with the needs
of a specific building life cycle stage. Its use starts with the definition of
This paper assumes that the involved parties (employer, architects, a Modelica library (see Fig. 1.d, step 1) that suits the needs of the
mechanical engineers, system integrators, manufacturers and facility considered building phase and complies with the constraints defined in
managers) use an openBIM workflow to manage and exchange in- Section 3. Subsequently, the Information Delivery Manual (IDM, ISO
formation throughout the entire building lifecycle. A distinction is 29481-1:2016 [57]) process is used to identify the data requirements
made between information providers (e.g. the architect or a manu- (see Fig. 1.d, step 2) to implement the BEPS models compatible with the
facturer) and the BEPS experts, who process the information for energy considered building stage (e.g. detailed design stage). IDM deliverables
simulation modeling. IFC4 constitutes the exchange format between the serve as base for a Model View Definition (MVD) (see Fig. 1.d, step 3),
project team members during design and construction, as well as for the which is a machine-readable (in XML-format) translation of the re-
as-built information transfer to the facility manager and the employer quirements and makes it possible to export the exact subset of BEPS-
(hand-over). related information. The MVD is used also to ensure a correct IFC file
The information inside BIM evolves throughout the building

a)
Life-cycle stages

Building in
Initiative Initiation Design Construction End of Life
use
Design sub-stages

b) Preliminary
Final Concept Design Detailed
Concept
Design Development Design
Design

c) Prepare BIM Export IFC Ifc2Modelica

BIM to BEPS

I2M-MVD Guidelines IFC STEP file Library template (.xml) Modelica model (.mo)
d)
Step1: Select a Step 2: Exchange Step 3: Create a BEPS-
Step 4: Propose a BIM
Modelica library (e.g., Requirements compatible MVD from
modeling guideline
IDEAS) identification existing MVDs
Information exchange

Create a library template

Library template (.xml) I2M-MVD Guidelines

Fig. 1. Integration of the proposed workflow throughout the building lifecycle. (a): life-cycle stage of a building, (b): different stages of the design, (c): the proposed BIM to BEPS
workflow, (d): Information exchange defining the steps to implement the MVD. (a) and (b) define the stage where the workflow is applied. (c) defines the workflow and (d) how the
requirements are implemented.

168
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

export. Lastly, a set of BIM modeling guidelines (see Fig. 1.d, step 4) – Table 2
listing the requirements in a human-readable format – is implemented Connector configuration required for Modelica components intended to be compatible
with BII.
and should be provided to the design team at the beginning of a project.
Section 4 demonstrates how the IDM process, the existing buildingS- Layout Connector requirements
MART MVDs and the Ifcdoc tool [58] can be used to implement custom
MVDs and guidelines, designed for a specific purpose (here: the crea- Zone components must have a bus connector; and optionally:
two opposite flow fluid connectors, and two heat connectors.
tion of Modelica models for BEPS at different lifecycle stages).
At any stage of the building lifecycle, the BEPS expert uses the ex-
changed IFC file as Ifc2Modelica input (see Fig. 1.c). A verification (and
enrichment) of the IFC file is performed to ensure a correct information Construction components must have one bus connector for
exchange (Section 5). Subsequently, Ifc2Modelica maps IFC to a Mod- external construction, otherwise two. (non-bus connectors can be
elica BEPS using a Modelica library that is compatible with the needs of added if necessary).
that stage (e.g. simplified Modelica library for the early design stage).
Section 6 presents the description of the approach behind Ifc2Modelica. System components use only fluid connectors. A signal
connector is attributed to controllable devices (e.g. fans,
dampers).
3. Modelica library requirements

The present section defines how Modelica components shall be Control components use only signal connectors.
implemented to be compatible with BII. Modelica is a non-proprietary,
multi-domain and industry capable modeling language. Modelica
models contain differential and algebraic equations, as well as discrete Internal gains and heater components use heat connectors for
ones. These equations are encapsulated in a Modelica class referred to heat transfer to a zone. A signal connector is added for a
as Modelica component which provides interfaces (Modelica con- controllable component (e.g. ideal heater), fluid connectors are
nectors) enabling interaction with other components. A component is a added for radiators.

graphically represented “object” which can be “dragged and dropped” on

the user interface of the Modelica simulation engines (e.g. in DymolaTM,
multi-domain connectors that can be used within any Modelica
OpenModelica). This facilitates model reuse, exchange, and extension
component. Components compatible with BII shall use the con-
by interconnection with other components.
nectors described in Table 1 for heat and mass transfer as well as
As a modeling principle, a Modelica component preferably re-
signal communication. bus connectors must be used to connect
presents an actual physical device with the same interface con-
construction components (e.g. walls, windows, slab) with a zone
nectors [6]. However, there are no strict rules on how BEPS components
component. heat connectors enable the link between a zone and its
need to be defined. As an example, a zone component can either in-
internal components; fluid connectors between HVAC components
tegrate the bounding walls or not; although, Modelica's flexibility al-
and signal connectors between control components.
lows to adapt one configuration to another depending on the needs.
2. Components: Components compatible with BII must be defined at the
Moreover, a Modelica connector represents the flow exchange that
building element level (e.g. a wall is modeled as a separate com-
can occur in reality between the constitutive elements of a building (see
ponent) and at the HVAC device level (e.g. dampers, fans, filters).
Table 1 for Modelica connector examples). For instance, one can as-
Table 2 shows how zone, construction, HVAC, control and internal
sume that a zone exchanges air (supply or return) with the HVAC
gain components must be represented in term of number and type of
system and exchanges heat (radiative and convective) through the
connectors used. These templates provide the interface definitions
bounding constructions. Hence, a Modelica zone component can have
(encapsulation) in which equations defining the device are im-
two fluid connectors (input and output) to exchange air, and also heat
plemented. This makes it possible to extend the use of BII to other
connectors. There are no standard rules defining the number and type
lifecycle stages by varying the complexity of the embedded equa-
of connectors for a BEPS component.
tions to fit into the considered stage.
Considering such flexibility, it is not realistic to implement an IFC to
3. Parameters: Required parameters (e.g. area, volume) need to be
Modelica interface compatible with all possible component configura-
defined under the Modelica “public” statement.
tions. Hence, the following constraints are proposed for the im-
plementation of BII:
Domain-specific components (e.g. BEPS) are gathered in a Modelica
package or library. The IEA EBC Annex 60 library, recently rebranded
1. Connectors: The Modelica Standard Library [59] defines elementary
IBPSA library (IBPSA 1 project), constitutes the core library of the main
Modelica libraries dedicated to BEPS: AIXLIB [54], Buildings [36],
Table 1 BuildingSystems [60] and IDEAS [61].
Modelica connectors used in a Modelica library to be compatible with BII. Although any library could be modified to comply with the previous
Connectors Description
constraints, the building component descriptions in IDEAS (see Table 3)
meet the constraints as mentioned earlier, while both the “Buildings”
Type:bus connector. Use: Multipurpose connector used to represent library and IDEAS satisfy the HVAC and control component require-
exchanges between zones and construction components ments. Therefore, IDEAS is used in this study.
Type:fluid connector, Flow direction: source (left), sink (right),
IDEAS allows a dynamic simulation of buildings (single or multi-
Medium: air or water. Use: Enable mass (or fluid) transfer between zones), systems, controls and also electricity grids as described in
components Baetens et al. [61]. IDEAS is best suited for the “detailed design” stage,
as the amount of information required by the current IDEAS compo-
Type:heat connector, Subtype: convective (left), radiative (right).
nents corresponds to the needs of this stage.
Use: Enable heat transfer between components
For the “early design” and the “operational” stage, simplified li-
Type:signal connector, Flow direction: source (left), sink (right), braries are preferable, since they require less input data and less cal-
Medium: signal type. Use: Enable control signal exchanges culation time.
between control components

169
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

Table 3 The IDM-MVD methodology can be implemented by the following

Examples of IDEAS library components representing construction elements. steps [48, 63-66].
Component Modelica component
1. Definition of the overall process as well as roles and functions of the
IDEAS.Buildings.Components.Zone (V:volume, hZone:zone height, different actors. This step results in a process map presented in a
n50:infiltration, nSurf:number of boundary surface) Busing Process Modelling Notation (BPMN).
2. Identification of the exchange requirements (ERs) defining which
information to exchange based on use-cases. A use case defines the
scope of the process (here: whole building energy simulation models
IDEAS.Buildings.Components.InternalWall (inc:inclination,
azi:azimuth, A:surface, constructionType:construction properties) using IDEAS). The ERs are reorganized into Exchange Requirement
Models (ERMs) from a set of Exchange Concepts. The latter specifies
the commonly used (and reusable) packages of information.
3. Implementation of the MVD by representing the ERMs with specific
IDEAS.Buildings.Components. SlabOnGround (inc, azi, A) IFC entities (here IFC4), attributes and types. This means that the
Exchange Concepts are translated into MVD concepts.
4. Implementation of a BIM guideline for the end user to comply with
the defined ERs.
IDEAS.Buildings.Components.OuterWall (inc, azi, A,
constructionType) An MVD consists of a set of MVD concepts that can satisfy the ex-
change requirements previously set. The latter feature enhances MVDs
reusability since concepts in existing MVDs can be analyzed to identify
if they meet the exchange requirements for energy simulation.
IDEAS.Buildings.Components.Window (inc, azi, A, glazing)
Therefore, MVD concepts defined in official buildingSMART MVDs can
be used as building blocks for an energy simulation MVD.
Also, business rules can be applied to an (existing) MVD to adjust its
use to a specific application such as energy simulation. They can be
applied to constrain the use of specific entities, attributes, properties,
4. BIM to BEPS exchange requirements and values.
The remainder of this section presents the application of the IDM-
This section ensures that the relevant information for energy si- MVD approach to formally produce a modeling guideline and MVD
mulation is included in the BIM and can be reliably exported through compatible with energy simulation using IDEAS. It details how the re-
IFC later in the process. quirements for an IDEAS-based energy simulation are identified, fol-
Typical problems encountered in a BIM-based energy simulation lowed by an approach combining existing MVD concepts with business
are: rules to restructure the requirements into an IDEAS-compatible MVD.
At the end of the section the modeling guideline is presented.
1. Incomplete product data, incompatible and inaccurate geometry
modeling, missing information about zones and spaces 4.1. Exchange requirements
2. BEPS-incompatible BIM modeling convention (see [62])
3. Inaccuracy of the information model This step identifies the requirements for IDEAS-based energy si-
4. Information loss, due to limitations of the BIM software itself, the mulation models. This means identifying components, component
used project templates or the mapping between proprietary classes parameters, and Modelica connections necessary to build an energy
and the corresponding IFC classes simulation model.
5. Information loss during the IFC export caused by a data scheme, i.e. Table 3 shows an excerpt of the IDEAS library presenting the zone,
a Model View Definition (MVD) that is not fully compatible with the window, slab, internal and boundary wall components. An initial step is
needs of BEPS to identify the required parameters. A zone component requires, at
6. Incomplete or incorrect HVAC system modeling and missing in- least, its volume, height, n50 value and the number of bounding con-
formation about controls structions. A wall needs its azimuth, surface area, orientation, and
constructiontype. The constructiontype defines the characteristics of a
For a successful use of BIM for energy simulation all concerned construction element and must include the number of material layers,
parties have to know exactly which information they shall insert into their thickness and thermal properties (thermal conductivity, thermal
the BIM at a specific point in time of the process. A modeling guideline capacity, density, longwave and shortwave emissivity). Components
for BIM shall be provided at the project start to detail such instructions connectors are then interconnected (see Fig. 4 as an illustration) to
while a BEPS-dedicated MVD can ensure a seamless data exchange obtain a Modelica model. Walls and windows can only be connected to
between the design team members and the energy simulation tool a zone via the bus connector while ventilation and control system
through IFC. The Information Delivery Manual and Model View components need to be connected based on actual devices' connections
Definition (IDM-MVD) methodology allow to define these two key (see Fig. 9 for illustration).
components: a guideline and a MVD (see Fig. 1.c and d). The requirements identification was extended to several use-cases
IDM is an international standard (ISO 29481-1) used to identify the including single- or multi-zone building models with and without un-
Exchange Requirements (ERs) in a specific process. It ensures that the heated spaces, ideal and hydronic heating systems, Variable Air Volume
final information model is semantically meaningful, thus allowing to air conditioning and generic control systems. The models were based on
support the end application (here: BEPS). IDM also specifies when the multi-space buildings where a space is a volume enclosed by the sur-
information from a specific actor in the project is required. A Model rounding construction elements (walls, windows, roofs, slab), not ex-
View Definition (MVD), on the other hand, translates the previous re- changing air flow through the openings. A thermal zone (modeled as a
quirements into a machine-readable format enabling information ex- IDEAS zone component) consists of one or several spaces depending on
changes between the different software tools in a project. the building configuration. Thus, a zone component is instantiated
using the properties of the aggregated spaces.

170
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

Considering system modeling, since IDEAS contains mainly the base properties (thickness and layers, material properties). Specifically, rule
system components (e.g. damper, fan, duct components), a ventilation 4 enables the definition of the HVAC zoning. Rules 5 to 7 ensure that
system and especially an AHU must be represented (in BIM) with its HVAC systems are correctly described, providing unambiguity for the
constituting elements. This ensures that the methodology is generic and flow direction within each element of the system. The link between a
does not depend on the availability of manufacturer specific Modelica thermal zone and systems is also clearly defined (Rule 7). Rule 8 to 11
components (in IDEAS). Flow direction information ensure that a spe- are setting the constraints to establish a control system properly. Rule
cific connector of a BIM object is mapped into the corresponding con- 12 provides a general constraint for all the property sets that need to be
nector in a Modelica component. As an example, if the flow direction of defined.
a (BIM) fan object connector (source or sink) is not provided, there is no Finally, the required concepts and entities are combined into a
information to identify which of the object connectors will be mapped custom MVD referred to hereafter as I2M by using IfcDoc [58]. I2M-
into the sink (or source) connector of the Modelica fan component. MVD is then provided to the design team members to ensure seamless
data exchange during a project.
4.2. MVD implementation
4.3. Guideline
This step identifies and selects among existing MVD concepts those
fully or partially fulfilling the exchange requirements. New property The requirements set in the Exchange requirements must be trans-
sets and business rules are defined if necessary. lated into a modeling guideline that will be provided to the design team
This means identifying MVDs that exchange information which can members at the project start. This guideline contains rules that shall be
represent zones, HVAC and control components representation, as well thoroughly applied in order to obtain a BEPS-compatible IFC file. The
as the different properties and connections between the Modelica main rules can be summarized as follow.
components.
Among the official buildingSMART MVDs, the IFC 4 Design Transfer 1. Geometry description: Spaces must be defined and fully enclosed
View (DTV) satisfies a broad range of exchange requirements varying within constructions. The general guideline presented by Maile
from coordination planning to visual representation. A comprehensive et al. [62] can be applied. In addition, spaces belonging to one
list of DTV concepts and their definitions can be found in the official HVAC zone shall be grouped into one IfcZone entity.
IFC4 DTV documentation [67]. These concepts satisfy a large part of the 2. HVAC and control description: Equipment and devices constituting
BEPS requirements. For instance, they exchange the required entities HVAC and controls must be (logically) connected following the real
(e.g. IfcSpace, IfcZone, subtypes of IfcBuildingElement, sub- devices in the system. Flow and signal directions, as well as devices
types of IfcFlowControlElement and subtypes of IfcCon- distribution system, shall be provided (rules 4, 5, 6 and 7, Table 5),
trolElement) as well as some parameters (e.g. zone component vo- i.e., the flow direction of a BIM object connector must be provided
lume from Qto_SpaceBaseQuantities.GrossVolume). Other as source or sink.
parameters result from the combination of different concepts. For in- For Net Zero Energy Buildings, low-energy systems that are capable
stance, walls' azimuth, area and orientation can be computed from the of bidirectional flow shall be represented in such a way that one of
combined information from the Axis 2D Geometry, Product Pla- the two possible flow directions is set as the default flow direction in
cement, Body SweptSolid Geometry and space boundary con- BIM. Since IDEAS components are capable of bidirectional flow, the
cepts. Connections between building envelop components are retrieved resulting system model is capable of bidirectional flow, in-
from the space boundary concept while those of the systems and dependently of its default flow direction in BIM.
controls result from the port connectivity concept. Since DTV is 3. AHU description: An air handling unit must be represented by its
not a BEPS-specific MVD, additional property sets and business rules constituting devices such as dampers, heat recovery or fans (For
are added. As an example, Table 4 shows new property sets not defined illustration, see Fig. 8). This condition is required since IDEAS does
in DTV but required by IDEAS. Infiltration property (n50) is attached to not contain entire AHUs, but does contain generic system compo-
IfcSpace entities while parameters describing the glazing perfor- nents from which an AHU model can be built. Also, this condition
mances (SwAbsi, SwAbsDif, SwTrans) are added to IfcWindow enti- allows the workflow to be more generic.
ties. Additional property sets could be added if required. 4. Actuator: BIM objects representing a controllable device (e.g. fan,
The business rules define constraints for entities, attributes, and pumps, valves) must be connected to an actuator object (rule 9,
properties to adapt the generic DTV concepts to the needs of BEPS. Table 5). An actuator must be connected to a controller object.
Twelve constraints (see Table 5) subdivided into three main parts were 5. Subtype. The BIM object's subtype must be provided. This further
added: rules for building geometry, HVAC, and control system. Rules 1 identifies the BIM objects having the same type but different sub-
to 4 (Table 5) ensure that adjacency relationships (2nd level space types (e.g. a sensor object can have a subtype temperature, CO2,
boundaries [42]) are exported alongside with the correct construction humidity,...).
6. Property set. Standard and custom properties defined in the re-
Table 4 quirements must be provided.
Example of additional property sets added to some entities.
A similar four-step process can be followed to define the guideline
Properties Descriptions and MVD compatible with other libraries and life cycle stages. The
IfcSpace guidelines have to be available to the BIM modelers at the beginning of
n50 Infiltration value (required by IDEAS) a project alongside with the MVDs. All parties involved have to comply
IfcWindow with the guidelines and enter the required information. These specific
SwAbsi Absorbed solar radiation for each layer as function of the angle of guidelines can be integrated into a project by using a BIM execution
incidence
SwAbsDif Absorbed solar radiation for each layer as function of the angle of
plan as described in the BIM project execution planning guide (CIC-
incidence RP) [68] or in PAS 1192 [69].
SwTrans Transmitted solar radiation as function of the angle of incidence
subtypes of IfcDistributionFlow 5. Verification and enrichment
fluid Fluid used in the component: air or water
IfcSensor
input type fluid (air or water), heat flow BIM-based collaborative design involves different task-specific tools
interoperating through IFC. These tools do not have a uniform IFC4

171
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

Table 5
Summary of the business rules defined to tailor standard MVD concepts for energy simulation.

Rule Id Proposition Allowed value

BEPS_Rule_01 A construction must be described as a set of material IfcRelAssociatesMaterial.RelatingMaterial = IfcMaterialLayerSet or IfcMaterialLayerSetUsage or

layers. IfcMaterialLayer
BEPS_Rule_02 Material thermal properties must be provided. EXISTS IfcMaterialProperties.Properties=TRUE; including Pset_MaterialCommon, Pset_MaterialThermal
and Pset_MaterialOptical
BEPS_Rule_03 Space boundaries must be defined. EXISTS IfcSpace.BoundedBy=TRUE
BEPS_Rule_04 Spaces belonging to the same HVAC zone must be EXISTS IfcSpace. HasAssignments. RelatingGroup = IfcZone
grouped in a zone.
BEPS_Rule_05 Ports must have a specified flow direction. IfcDistributionPort.FlowDirection=SOURCE or SINK
BEPS_Rule_06 The distribution system (e.g. Supply air, return air, EXISTS IfcDistributionFlowElement. HasAssignments. RelatingGroup IfcSystem
exhaust air) to which an object belongs should be
provided.
BEPS_Rule_07 If exists, space heater and air terminals must be IF EXISTS IfcSpaceHeater, it must be related to an IfcSpace through the spatial containment concept. IF
related to a space. EXISTS IfcAirTerminal, it must be related to an IfcSpace through the spatial containment concept.
BEPS_Rule_08 A set of sensors (Temperature, CO2 and Humidity if IfcSpace.ContainsElement.RelatedElement INCLUDES IfcSensor. PredefinedType.CO2SENSOR, IfcSensor.
required) must be contained in a space. PredefinedType. TEMPERATURESENSOR, IfcSensor. PredefinedType.HUMIDITYSENSOR.
BEPS_Rule_09 An Actuator must be connected to a controller IfcActuator must be connected to a IfcController through the Port Connectivity concept [67].
BEPS_Rule_10 A controllable HVAC equipment must host an IfcActuator must be connected to a controllable subtype of IfcDistributionFlowElement through the
actuator object. Element Connectivity concept [67].
BEPS_Rule_11 Type of space heater, sensor, controller and damper EXISTS IfcSpaceHeater.PredefinedType=TRUE; EXISTS IfcController.PredefinedType=TRUE; EXISTS
must be provided. IfcSensor.PredefinedType=TRUE; EXISTS IfcDamper.PredefinedType=TRUE.
BEPS_Rule_12 Property set that must be provided (e.g. Table 4)

I2M MVD- Ifc2Modelica

compliant IFC ?
Topology Mapping
Verification
BEPS model generation process

yes generation process

No

Enrichment

IFC input I2M-MVD Spreadsheet file Spreadsheet file I2M-compliant Topology Library template Modelica BEPS model
report Report updated IFC (.mo)

Fig. 2. BEPS generation process. Verification and enrichment and direct translation process (Ifc2Modelica).

MVD compliance nor support custom MVDs, increasing the risk of in- 1. Object Definition: Property Sets, Quantity Sets
formation loss during information exchanges. Therefore, even though 2. Object Association: Material Layer Set, Material Layer Set
the modeling guideline in the previous section is thoroughly applied, Usage
data loss might occur during the exchange processes making the re- 3. Connectivity: Port Connectivity, element connectivity
sulting IFC file incompatible for energy simulation [70, 71]. In this
section a two-step PYTHON based approach is proposed that takes an The ability to instantiate property sets and Material layer makes it
IFC file as input, performs verification and enrichment to ensure an possible to update custom and material properties. Similarly, entities
I2M-compliant IFC and produces an accordingly modified output IFC that have disconnected ports (IfcDistributionPort) are listed and
file (see Fig. 2). can be connected through this enrichment process by providing in the
spreadsheet the port to which the disconnected one shall be attached.
An I2M-compliant IFC file results from this process. It is important to
5.1. Verification
note that this is a temporary step which is introduced to overcome
current interoperability limitations. Hoping that those issues will be
The verification process assesses whether the (IFC) file is fully I2M-
fixed in future versions of the commercial modeling software, this step
compliant or not. It compares attributes and properties of the entities in
can be skipped as soon as a custom MVD (here: I2M) can be used di-
the file with I2M concepts and business rules (Table 5). A non-com-
rectly during the IFC-export of that software.
pliant IFC file will result in a request to enrich the file. At the same time,
the developed tool generates a spreadsheet containing the description
of the missing information enabling the user to fill the missing values
6. Ifc2Modelica: model generation
into the corresponding cells before the enrichment process is started.
At this stage (see Figs. 1.c and 2), the IFC file contains all the re-
5.2. Enrichment levant information for BEPS modeling. The present section details the
translation of this IFC file to a Modelica-based energy simulation model
This process can add semantic information to the IFC file and uses using Ifc2Modelica. The tool that was developed in this research project
the previous spreadsheet as input to enrich the IFC file with three types uses the PYTHON library IfcOpenShell [72] to handle IFC4 and requires
of concepts: an I2M-compatible IFC and the I2M-MVD (as a .mvdxml) as input.

172
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

a) Wall 1 Wall 1-1 Wall 1-2 b) w IfcWallStandardCase C IfcController

Temperature Sensor w w
S IfcSpace IfcZone

Space 1 Space22
Space Se S1 w S2 Space boundary
H IfcSpaceHeater
Spatial containment
Zone 1
A IfcActuator Element connectivity
C A H
Controller Actuator Ideal Port Nesting +
heater Wall 2
Port Connectivity
Se IfcSensor
c) wi d)
we Zone node: IfcSpace/
we wi we Z Controller node:
IfcZone C
IfcController
Z
Se Internal wall node:
Z Z wi
IfcWallStandardCase O Occupant node
Se
C

External wall node: Sensor node:

we Se
C

IfcWallStandardCase IfcSensor
A O
H Actuator node: Heater node:
A A H
O O IfcActuator IfcSpaceHeater
H

Fig. 3. Topology generation process. (a) schematic view of the original BIM. (b) IFC structure of the BIM in (a). (c) Topology generated by Ifc2Modelica where the two spaces are merged.
(d) Topology generated by Ifc2Modelica where each space is assumed as a zone.

Ifc2Modelica consists of two steps: topology generation and mapping component. A zone entity (IfcZone) is translated into a zone node
process (see Fig. 2). with the characteristics of the aggregated spaces (e.g. spaces 1 and 2 in
The two-space building (spaces 1 and 2) depicted in Fig. 3.a serves Fig. 3.b), e.g. the zone volume is equal to the sum of the volumes of all
as an example to illustrate the process. A thermostat controlled heater aggregated spaces.
conditions the space 1. For simplicity, only walls 1 and 2 are depicted in In the case where the spaces (IfcSpace) are not related to a zone, a
the IFC representation of the building in Fig. 3.b and mentioned in the space entity (IfcSpace) corresponds to a zone node (Fig. 3.d). Op-
description hereafter, since all other walls follow the same rules. tionally, for each zone node an occupancy node (O) can be created.
Construction nodes. Construction node types include external
6.1. Topology generation (We) and internal walls (Wi), windows (Win) and slabs (S). Each node
should correspond to a Modelica component (e.g. in Table 3).
A Modelica model structure consists of a set of (1) Modelica com- As a first translation rule, construction entities (e.g.
ponents where each component has a set of (2) Modelica connectors. IfcWallStandardCase, IfcSlab) spanning several spaces are split
These connectors are then (3) connected to other components' con- into several nodes in accordance with the space boundary con-
nectors, thus forming a BEPS model. cept [42]. As an example, wall 1 in Fig. 3.a spans space 1 and 2 and is
IFC has a different structure which prevents its direct mapping into split into five parts. If they are adjacent to an open space (e.g. wall 1-1
Modelica. As an example, the space boundary concept defining the and wall 1-2), the parts bounding a space are translated into external
logical connection between a space (IfcSpace) and walls nodes (We). Otherwise, they are translated into internal nodes (Wi)
(IfcWallStandardCase) is constituted of several abstract entities, (e.g. wall 2). The distinction between internal and external construction
which do not have real physical meanings. A direct mapping implies is determined based on the IfcRelSpaceBoundary.InternalOr-
that each abstract entity is represented as a Modelica component which ExternalBoundary property and/or by identifying the different
is both impractical and difficult to implement. overlaping boundary areas computed from IfcRelSpaceBoundar-
In addition to that, some IFC information is not explicitly available y.ConnectionGeometry.
and needs to be computed from a combination of properties and con- Moreover, slab nodes (S) and window nodes (Win) correspond re-
cepts. spectively to: slab entities (Ifcslab) constituting the floor of a space
The topology generation is a rule based method that transforms the located on the lowest floor and window entities (IfcWindow).
IFC model (e.g. Fig. 3.b) into a topology (e.g. Fig. 3.c or Fig. 3.d) The boundary area (#A# in Table 3) is computed from the If-
compatible with a direct mapping to a Modelica model (the Modelica cRelSpaceBoundary.ConnectionGeometry. Walls and windows
library used must comply with the constraints in Section 3). are assumed vertical (inclination=90°.), while their orientation (azi-
The topology consists of three elements: Nodes, Node connectors muth) is computed based on the angle between the (true) north and a
and Edges. The first two transform IFC entities into objects that mirror vector normal to the base surface of the bounded space. This normal
Modelica components while the edges represent the Modelica con- vector is computed from the Axis 2D Geometry and Product Pla-
nections. cement concepts. Also, slabs are assumed horizontals (inclination=0°.
or 180°.) with an azimuth fixed to zero. Moreover, construction prop-
erties come from the Material Layer Set or Material Layer Set
6.1.1. Nodes
Usage concepts. They are formatted into IDEAS specific definitions.
A node can be mapped into a Modelica component and is an image
System nodes. The entities that are subtype of the
of an IFC entity. It has the same attributes, and concepts as the entity it
IfcDistributionFlowElement entity are converted into system
represents. However, based on different criteria presented in this sec-
nodes (St). S defines the entity type (e.g. H for an heater entity,
tion, the translation from entity to node can be one-to-one, one-to-many
IfcSpaceHeater) and the subscript t its “predefined type” (e.g.
or many-to-one.
RADIATOR). Therefore, (HR) represents a radiator node.
This section focusses on how IFC data is used to implement the four
An exception to the above rule applies to the entities that are a
types of nodes (Zone, Construction, System, and Control).
subtype of IfcFlowSegment or IfcFlowFitting entity. These
Zone nodes. A zone node (Z) corresponds to a Modelica zone

173
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

entities are aggregated into one node instance or into the neighboring representation similar to a Modelica component (see Fig. 3). Edges are
node instance(s) of the subtype IfcDistributionFlowElement. the last building block of the topology and represent the Modelica
This will reduce the number of nodes and hereby simplify the energy connections, they contain the information about the different pairs of
simulation model. connectors that need to be linked. Thus, the association of those three
Control nodes. Control nodes (Ct) represent entities that are a concepts (node, connector and edge) results in a Modelica compatible
subtype of the IfcDistributionControlElement entity. C defines topology (e.g. Fig. 3.c and d).
the entity type (e.g. Se for a sensor entity, IfcSensor) and the subscript t An edge originates from the conversion of an IFC connectivity
its subtype (e.g. TEMPERATURE). Therefore, (SeT) represents a tem- concept based on one of the following rules:
perature sensor node.
1. A space boundary concept, between a zone node and a con-
6.1.2. Node connectors struction node, links their bus connectors.
Connectors are additional attributes attached to nodes, derived from 2. A spatial containment concept, between a zone node and any
the entity type or connectivity concepts. Node connectors represent an node with heat connector, links their heat connectors.
analogy to Modelica connectors (see Table 1). This section defines four 3. A spatial containment concept, between a zone node and an air
types of connectors (bus, heat, fluid, signal) and details how IFC in- terminal node (or sensor node), links their fluid connectors with
formation is used to attribute these connectors to nodes. opposite flow direction.
bus connector. It enables heat transfer between zones and con- 4. A port connectivity concept, between any pair of system (or
struction elements. It is attributed to zone and construction nodes. An control) nodes, links their fluid (or signal) connectors with opposite
internal node (Wi) has two bus connectors while other construction flow direction.
nodes and the zone node have only one. 5. An element connectivity concept, between a system and a
heat connector. It enables convective and radiative heat transfer control node, links their signal connectors.
between a zone and its “internal components” namely occupants, sensors
and heaters (IfcSpaceHeater). Zone, occupant, sensors (heat flow 6.2. Mapping process
temperature) and heater nodes (Z) have heat connectors.
fluid connector. It enables fluid transfer between building system The Mapping process converts a topology into a BEPS Modelica
nodes. A fluid connector is characterized by its flow direction (source or model. However, as a precondition, each node must have a corre-
sink), fluid type (e.g. air or water) and a group parameter, e.g. “supply sponding Modelica component known beforehand. This information is
air” or “return air”. An additional requirement is, that each fluid con- stored in a file hereafter referred to as library template, implemented at
nector belonging to the same node must have a unique set of char- the same time as Section 4. As a library template example, Table 6
acteristics (e.g. a node with two fluid connectors, both having the same presents the one used for the example case in Fig. 3.
fluid type and flow direction, is not permitted). This will ensure a The following two sections provide a more detailed description of
correct mapping of a fluid connector into a Modelica fluid connector, the library template and the mapping process.
especially for components with multiple connectors (e.g. heat recovery
system). 6.2.1. Library template
fluid connectors are attributed to zone and system nodes. A zone node The library template contains a set of component templates which
has two fluid connectors with opposite flow direction (source and sink). The defines the analogy between a Modelica component and a node. A
number of fluid connectors associated with a system node depends on the component template contains five fields with information related to the
number of IfcDistributionPort associated to it. Subsequently, the flow Modelica component: node, function, input, connector and annotation (see
direction is deduced from IfcDistributionPort.FlowDirection Table 6).
while the group to which it belongs originates from The “node” field defines which node corresponds to the Modelica
IfcDistributionPort.HasAssignments (Group Assignment component (e.g. Table 6, columns 1 and 2). The “function” field defines
Concept). The fluid type is either deduced from the entity type (e.g. a node the component path used to instantiate the component, while the
instance representing an IfcAirToAirHeatRecovery entity has air as “input” field contains the names of the parameters (between #) that are
fluid type) or from the additional property set in Table 4. As an exception to required to instantiate the component (e.g. Table 6, third column).
these rules, an air terminal node (IfcAirTerminal) is itself represented as Those parameters have to be named in accordance with the corre-
a fluid connector. A fluid (sink) connector is also attributed to a sensor node. sponding property in the (property set) concept. The “connector”
signal connector. It enables the exchange of control signals be- field describes the different Modelica connectors attached to the com-
tween control system devices. A signal connector is characterized by its ponent (e.g. Table 6, third column) according to Table 1. Especially, it
signal direction (source or sink) and type (e.g. temperature, CO2, con- contains the flow direction, fluid type and group of the Modelica fluid
troller). The latter defines the type of signal sent or expected by the connectors. The “annotation” field provides the position information to
connector (e.g. a source-temperature connector sends a temperature the Modelica simulation engine for the graphical layout by providing
signal while a sink-controller connector expects a signal from a specific values for x0, y0, xw and yw.
type of controller).
As a rule, an actuator node (A) has two signal connectors with op- 6.2.2. Mapping
posite flow while a sensor node (Sei) has a (source) flow connector. The mapping process instantiates a Modelica component for each
Also, a signal connector is attributed to a controllable system node re- node in the topology and converts the edges into Modelica connections
lated to an actuator (element connectivity concept). For the re- (Modelica statement connect). This process consists of four steps: tem-
maining control nodes, the number of signal connectors depends on the plate identification, connector comparison, instantiation and connection
number of IfcDistributionPort associated to the node. (see Fig. 4).
Connectors flow direction is deduced from The template identification searches in the library template (see
IfcDistributionPort.FlowDirection while the type is de- Table 6) the Modelica component corresponding to a node. Example: a
termined from the node type connected to the port connectivity zone node (Z) corresponds to the IDEAS.Buildings.Compo-
concept. nents.Zone component.
The connector comparison identifies whether the identified Modelica
6.1.3. Edges component has the same connectors (in terms of type and flow direc-
The combination of a node and its connectors results in a tion) as those defined on the node. If not, the component cannot be

174
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

Table 6
Example of a library template providing the correspondence between Modelica components and nodes. Column 1: correspondence between node and entity. Column 2: Modelica
component. Column 3: function and input of the Modelica component, column 4: description of the different connectors attached to the Modelica component.

Nodes: Entity Component Modelica function Connectors

: IfcSpace, IfcZone function: IDEAS.Buildings.Components.Zone(), input:V=#GrossVolume#(volume), bus, fluid1(air, source),

hzone=#Height#(height), n50=#n50#, nsurf=#nSurf#(number of adjacent surfaces). fluid2(air, sink), heat
(radiative, convective).

function: IDEAS.Buildings.Components.InternalWall(), input:inc=#inc#(inclination), bus1, bus2.

: IfcWallStandardCase,
azi=#azi#(azimuth), A=#A#(area), constructionType=#constructionType#
IfcSlab, IfcDoor (construction properties). These parameters are internally computed.

function: IDEAS.Buildings.Components.SlabOnGround(), input:#inc#, #azi#, #A#. bus.

: IfcSlab

function: IDEAS.Buildings.Components.OuterWall(), input:#inc#, #azi#, #A#, bus.

: IfcRoof,
#constructionType#.
IfcWallStandardCase,
IfcSlab

function:IDEAS.Buildings.Components.Window(), input:#inc#, #azi#, #A#, #glazing#. bus.

: IfcWindow

: IfcActuator function:Modelica.Blocks.Routing.RealPassThrough(). signal (source), signal

(sink).

function: IDEAS.Controls.Continuous.LimPID, input: k=#ProportionalConstant#, signal (source), signal

:IfcController (type:
Ti=#IntegralConstant#, Td=#DerivativeConstant#. (sink).
proportional)

function: Modelica.Thermal.HeatTransfer.Sensors.TemperatureSensor. signal (source), heat

: IfcSensor (type: Temperature,
(sink).
input:heat flow)

function: User defined. heat, signal (sink)

: IfcSpaceHeater (type: ideal)

Annotation: annotation (Placement(transformation(extent=x0,y0,xw,yw)));

instantiated. The instantiation step retrieves the values of the parameters Section 7.1 is used as a test facility.
from the topology, provides a name and instantiates the component.
Example: IDEAS.Buildings.Components.Zone zone_1
(nSurf=5, V=150, hZone= 2.7, A=55, n50=2) annotation 7.1. Test facility description
(Placement(transformation(extent=10,10,25,25))); re-
presents an instance of a zone component. In this step the component is The facility, part of the Technology Campus Ghent of KU Leuven
also positioned in the Modelica graphical layout by providing a value to (Belgium), is a Near Zero Energy Building (nZEB) and comprises four
annotation parameters x0, y0, xw and yw. spaces: the two lecture rooms E120 and E220 (for up to 80 students
The connection step connects the different instantiated components, each), staircase and technical room (see Fig. 5). The box-shaped rooms
using the information contained in the edges. Fig. 4 summarizes the (E120 and E220) are thermally insulated not only from the exterior but
mapping process and presents the resulting Modelica model. also from one another and from the staircases. This enables the use of
simple uncoupled simulation models, due to the neglectable heat flow
in between the different spaces. The external walls of E120 are designed
7. Case study with bricks (high thermal mass) while the external walls of E220 are a
timber frame structure (low thermal mass). The two lecture rooms have
This section illustrates the benefits of the proposed approach using an identical window area. High performance triple glazing is used and
the developed tools and guidelines (BII). A comparison is made with the south oriented windows are equipped with blinds controlled by the
another BIM to BEPS approach, where the need for a formal informa- BMS. All windows are motorized and can be opened and closed auto-
tion exchange between design representation and BEPS is neglected matically, controlled by the BMS (e.g. to test different night ventilation
(ad-hoc BIM to BEPS). To emphasize the advantages of BII, two sce- strategies).
narios are presented in Section 7.2. The differences in term of resulting An air handling unit (MenergaTM AdsSolair 56) ensures the indoor
BEPS are underlined and discussed. The building described in comfort and air quality in E220 and E120. Fig. 6 shows the ventilation

175
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

Fig. 4. Overview of the mapping process: the translation of a topology (a) into a Modelica BEPS model (c). Step 1: For each node in the topology (a), identification of the corresponding
component template in Table 6. Step 2: Comparison of the connectors attached to the node with those attached to the component template. Step 3: Instantiation of the components. Step
4: Connection of all the instantiated components, resulting in a Modelica model shown in (c).

system scheme including the internal elements of the AHU and the TEA
sensors. The AHU is capable of delivering a maximum air flow rate of
5100 m3/h and an air flow of 4000 m3/h in normal conditions. It has a TEA
DEA-BP FNRA FRA TRA
DEA
coefficient of power efficiency of 72% (EN 13053:2012) and possesses a
double heat exchanger (HR in Fig. 6) in polypropylene and an Indirect IEC
Evaporative Cooling (IEC) system. The latter has a cooling capacity of TRA
TOA
DOA HR DRC FSA HCE120
21.1 kW and allows a significant cooling (14 K). The heat recovery
system can reach an efficiency superior to 75%. The AHU contains two
energy-efficient fans (supply air fan: FNSA and return air fan: FNRA), a FOA
supply air filter (FSA, FOA) and a return air filter (FRA). Five internal FNSA TSA
dampers (see Fig. 6): (1) Exhaust Air (EA) bypass (DEA−BP), (2) exhaust TOA DOA-BP
HCE220
air (DEA), (3) Outside Air (OA) (DOA), (4) outside air bypass (DOA−BP)
HR: Heat Recovery F: Filter SA: Supply Air BP: ByPass
and (5) recirculation (DRC) are present. A wood pellet boiler supplies
IEC: Evaporative cooling FN: Fan RA: Return Air R:Return
hot water to the heating coils (HCE120 and HCE220). Air flow rates to the
HC: Heating Coil D: damper EA: Exhaust Air S: Supply
lecture rooms are regulated by (4) four Variable Air Volume (V
AVS−E120, V AVR−E120, V AVS−E220, V AVR−E220). Temperature sensor
The AHU has integrated temperature and humidity sensors placed at Fig. 6. Overview of the components constituting the air handling unit.
the air inlets and outlets and inside the heat recovery system (see
Fig. 6). In addition, temperature (T), humidity (RH), CO2, Volatile
Organic Compound (VOC) and motion detection sensors are present in
both lecture rooms. Supply and return air volume flow is measured by
four venturi tubes equipped with differential pressure sensors.
At hand-over, the HVAC engineer and the architect provided a BIM
representing the building envelope and the HVAC system (see Fig. 7.a
and b). The models were created before the start of the present study
and are not compliant to the guidelines defined in Section 4. Also,
properties (thermal and physical) and ventilation system descriptions
were provided in separated PDF files or spreadsheets and the objects in
these models do not contain the required parameter values for BEPS.

Fig. 7. (a) Building envelope representation, (b) Ventilation system representation

7.2. Scenarios
“detailed design” stage. The first scenario considers a BIM implemented
Two scenarios are considered to compare the “ad-hoc” BIM to BEPS without a formal definition of the exchange between design and BEPS,
approach with BII. For both, it is assumed that the project is in the or otherwise said: without prior knowledge of the guideline described

Fig. 5. a) Picture showing the southern facade of the facility. b) Section view showing the different spaces (technical room, staircase, E120 and E220). c) Floor plan.

176
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

as “confidential” and is not provided to the client [5]. The current lack
of standardized, open formats further complicates this information ex-
change. Scenario 2 assumes full cooperation to fulfill the requirements.
The different parties shall use IFC4 compatible BIM tools to ensure a
seamless data exchange.
Building Information Model. The envelope representation does
not visually differ from the original model in Fig. 7.a since the mod-
ifications consist of defining the rooms as IfcSpace objects, and the
integration of semantic data such as material properties and property
sets (e.g. Table 4).
The model for the ventilation system for scenario 2 is shown in
Fig. 8. In compliance with the requirements defined in Section 4, the
AHU and the ductsystem is represented by its constituting elements
(e.g. dampers, fans, valves). More specifically, Fig. 8.b provides a closer
view of the AHU model. Such description is obtained by (1) identifying
the AHU internal components, brand, type and all relevant character-
istics, (2) retrieving or creating the corresponding BIM objects (here:
REVIT Family) and (3) assembling and connecting these objects ac-
cording to the rules 5 to 7 in Table 5 and the scheme in Fig. 6.
Similarly, BIM objects representing actuators (blue boxes in Fig. 8.a
and b) are implemented and attached to controllable devices (fans,
dampers, valves) to comply with the rule 10 in Table 5. These actuators
are connected to controllers (depicted as brown boxes) (rule 9 in
Table 5), that send signals to the controllable devices and receive sig-
nals from sensors or controllers. As an example, a VAV controller uses
the signal from the CO2 and temperature sensors of a lecture room to
generate the control signal of the VAV box.
In the following step an IFC4 file is exported. However, due to
limitations of the used BIM tool, the direct use of a custom MVD, i.e.
I2M-MVD during the export is not yet possible. Instead, the IFC4 Design
Transfer View (DTV), in combination with the BIM tool's capability to
export user-defined property sets, is used.
Fig. 8. a) Detailed representation of the HVAC system, b) Closer view of the Air Handling BEPS Model. Fig. 9 shows the final Modelica BEPS model that was
Unit representation. obtained by using the BIM presented in Fig. 8 as Ifc2Modelica input.
The BEPS model comprises the envelope (Fig. 8.b), the ventilation
in Section 4. The BIM in Fig. 7 is directly used as Ifc2Modelica input. system (Fig. 8.a) and the control model (Fig. 8.c).
The second scenario thoroughly applies BII (including guideline, tools It was found that despite the fact that the previously established
and MVD). An example of BII application alongside with a comparison modeling requirements were met, the connections between controllers
of the generated models from both situations are provided hereafter. and actuators, actuators and system devices were not exported as de-
fined in (rules 7 and 8 in Table 5). Also, material properties were
7.2.1. Scenario 1 missing. These issues are due to the fact that the IFC file was created
The BIM in Fig. 7.b represents the AHU as a simple box. Its internal according to the DTV-MVD and some characteristics specific to the I2M-
component descriptions are provided as data-sheets (or pdf-files) while MVD were not exported, thus again emphasizing the need for a custom
control strategy information is unavailable. The properties of the MVD compatible BIM authoring tool. Also, loss of semantic translation
building elements are supplied in data-sheets and they are not in- [70, 73] explains why the material properties were not exported. These
tegrated into the BIM. An IFC file is exported from the REVITTM model issues depend on the capabilities of BIM tools to correctly translate
and used as Ifc2Modelica input. The verification process (Section 5) proprietary models into IFC4 models. Nonetheless, the newly developed
checks I2M-MVD compliance and generates a spreadsheet prompting verification tool (see Section 5) provides for now a solution allowing to
the modeler to input the missing values for the properties (material integrate the missing information into the IFC file.
properties). This manual data transfer increases the risk of errors. Also, The Model generation process specified in Section 6 starts by cate-
information such as AHU component description and control strategies gorizing entities as nodes or edges according to the rules in
cannot be inserted through this verification process. Therefore, the Sections 6.1.1 and 6.1.3. The topology is mapped into a Modelica BEPS
BEPS generated from this BIM consists only of the building envelop using the library template defined in Tables 6 and 7. Table 7 describes
model described in Fig. 9.b since there is not enough information to the system components whose detailed descriptions and equations are
generate the models of the control system and the AHU. described in IDEAS documentation [74]. In addition, some “user-de-
fined” components are used that are not explicitly defined in IDEAS, but
7.2.2. Scenario 2 implemented using several components from IDEAS as a base.
This scenario applies BII using the BIM execution plan proposed Since each room was defined as an IfcSpace and none of them are
in [68] as framework. Energy analysis is defined as one of the BIM uses aggregated in an IfcZone, Fig. 9.b presents the envelope model re-
and the team members have to adhere to the modeling guidelines de- presenting each room (E220, E120, technical room, and staircase) in-
veloped in Section 4. More specifically, the AHU must be modeled dividually as an IDEAS zone component. Fig. 9.a and d represents the
down to its constituting elements, control strategy information and HVAC system and AHU models respectively. If the AHU had not been
material properties must be integrated into the BIM. For this scenario to designed with its constituting elements, the model in Fig. 9.d would
be valid, close collaboration between all involved parties is essential. have been represented by only one Modelica component. This would
Ideally, the manufacturers shall provide BIM objects of their products have required manual modeling of the AHU or the availability of a
including HVAC control strategies. Often this information is considered component describing this specific AHU. However, representing all

177
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

DEA
FRA
FN
IEC
HR DRC
FOA D FSA

DOA-BP
FNS
d) AHU

a) HVAC system BOILER

IEC

DBP

E120 E220
DOA

FN

Technic Stairca
VAV
al room se
E120

VAV
b) Building Envelope
E220
c) Control
Fig. 9. Resulting BEPS Model. a) ventilation system model, b) building envelope, c) Control and d) Air Handling Unit.

AHUs on the market as a Modelica component seems unrealistic and the However, the collaboration within the design team, called BIM-
modular approach based on a detailed BIM representation is preferred. based Collaboration Network (BbCN) in [15], is key to fulfill the set of
Fig. 9.c presents the control system where actuator nodes are con- requirements in Section 4, especially to be able to integrate ventilation
verted into “Realpassthrough” components (see IfcActuator in and control system data. This partnership creates significant challenges
Table 6) and serve as a link between controllers and system devices. for sharing data and resources, influencing at the same time the deci-
Controller nodes are mapped into Single-Input Single-output (or Mul- sion-making process which involves members of several organizations
tiple-Input Multiple-output) components (see IfcController in from different disciplines and backgrounds, likely working in various
Table 6). The IFC4 standard defines five controller subtypes (Floating, locations. In addition, legal risks associated with liabilities, BIM in-
Multiposition, Programmable, Proportional, and Two-position) which tellectual property rights, and managerial difficulties need to be con-
are then mapped into five different control nodes as described in sidered. The challenges mentioned earlier are obstacles in harnessing
Section 6.1.1. These standardized subtypes are not sufficient to define BIM capabilities and constitute an active topic within the research
all existing types of a control system. However, the possibility of es- community [12, 15, 16]. As an example, Oraee et al. [15] identify and
tablishing user-defined controller types and property sets makes IFC explore the influential factors towards a successful collaboration while
capable of representing any control systems. Mignone et al. [75] provide an early solution to address BbCN colla-
boration issues.
8. Discussions In addition to the social collaboration difficulties, technology lim-
itations such as incompatibility of BIM tools with custom MVDs and
8.1. Concerning the case study information loss during object translation currently limit the possibi-
lities of direct BIM to BEPS coupling. These technological issues were
The contrast between scenario 1 and scenario 2 emphasizes the differ- already reported in the literature [16] as being among the major ob-
ence between current practice and a well-coordinated BIM-based workflow stacles to the full adoption of BIM in general and especially in the
(BII). While scenario 1 limits the added value of BII to the envelope model, translation of BIM to BEPS. In this context, Pinheiro et al. [48, 49, 79]
scenario 2 demonstrates the BIM capabilities to carry BEPS-related in- present a first step towards the support of custom MVDs. However,
formation as well as the capability of BII to generate a complete BEPS model. temporary solutions such as the one defined in Section 5 are still

178
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

Table 7
Additional library template describing the correspondence between system nodes and Modelica components.

Nodes: Entity Component Modelica function Connectors

: IfcAirTerminal function: Modelica.Fluid.Interfaces.FluidPort_a() or fuid (sink) or fluid (source).

Modelica.Fluid.Interfaces.FluidPort_b().

: IfcDamper, IfcValve function: IDEAS.Fluid.Actuators.Valves.TwoWayLinear(), input: #m_flow_nominal#= fluid (source), fuid (sink),
IfcDamper.NominalAirFlowRate, #dpValve_nominal#= signal (sink).
IfcDamper.OpenPressureDrop.

: IfcFan, IfcPump function: IDEAS.Fluid.Movers.FlowControlled_m_flow(), input: #m_flow_nominal#= fluid (source), fuid (sink).
IfcFan.NominalAirFlowRate, #dp_nominal#=
IfcFan.NominalTotalPressure,#per# from IfcFan.PressureCurve and
IfcFan.EfficiencyCurve.

function: IDEAS.Fluid.HeatExchangers.ConstantEffectiveness (), input: #eps#= fluid (source, supply air

: IfcAirToAirHeatRecovery,
IfcAirToAirHeatRecovery.TotalEffectiveness, #m1_flow_nominal #= system), fuid (sink, outside
IfcCoil IfcAirToAirHeatRecovery.PrimaryAirflowRateRange, #m2_flow_nominal #= air system), fluid (source,
IfcAirToAirHeatRecovery.SecondaryAirflowRateRange. exhaust air system), fuid
(sink, return air system).

: IfcEvaporativeCooler User-defined. fluid (source), fuid (sink),

signal (sink).

: IfcDuctSegment, IfcPipe function: IDEAS.Fluid.FixedResistances.PressureDrop(), input: #m_flow_nominal#= f fluid (source), fuid (sink).
(Qto_DuctSegmentBaseQuantities, Pset_DuctSegmentTypeCommon),
#dp_nominal#= f(Qto_DuctSegmentBaseQuantities,
Pset_DuctSegmentTypeCommon)

: IfcBoiler User defined. fluid (source), fuid (sink).

required while waiting for the necessary improvement of BIM tools. open-source languages such as PYTHON and Modelica. Free and open
In scenario 2, the use of “user-defined” components emphasizes that BEPS dedicated Modelica libraries include components for modeling
a specific library might not include all the appropriate components for a electrical, thermal, fluid and control systems. Although BII does not
specific device. Existing library components are often based on generic provide a specific user interface, it combines the advantage of a 3D
equations and do not embed manufacturer specific features. This em- modeling environment (provided by a BIM tool) with the graphical -
phasizes the importance of online, collaborative modeling processes component oriented - Modelica model representation. The latter facil-
through a GIT-based platform adopted by some Modelica libraries (e.g. itates a subsequent modification of the resulting BEPS by manual or
IDEAS and Buildings). Manufacturers and BEPS experts will then have “drag and drop” editing of components. Furthermore, since BII connects
the opportunity to submit new components or refine existing ones, thus BIM to a Modelica library, the model accuracy essentially relies on the
contributing to the fast expansion and growth of such a publicly library's validity. A thorough accuracy assessment of the BEPS model of
available library. The constraints presented in Section 3 ensure the the case study was already performed but is beyond the scope of this
compatibility of the Modelica components with BII. It can be envisioned paper and will be published soon. BII is an open BIM to BEPS workflow
in the future to integrate equations into BIM objects and directly which fulfills the current needs towards energy efficient building
transfer them into Modelica components, using the formal definition of modeling and can be used as a base for the creation of alternatives to
functions as described in the German guideline VDI 3805 [76] and the commercial BEPS tools.
current drafts of ISO 16757 [77]. This approach is currently under In this study BII is applied at the “detailed design” stage. Guidelines,
development and depends on the publication of all relevant parts of ISO MVD and library template compatible with this stage were developed.
16757 to be fully implemented. However, BII is compatible with any other building life cycle stage,
necessitating first a thorough application of the exchange process
8.2. General discussion (Section 4) to define the guidelines, MVD necessary; as well as the
definition of a compatible Modelica library and library template. This
The resolution of the previously mentioned issues would allow a process (Section 4) is performed only once and need to be updated only
direct, seamless and fully automated translation from BIM to Modelica- when a major modification of the Modelica library occurs. BII can then
based BEPS models. In such case, BII fits the future view of building be applied to any project at the same stage having a similar building
energy simulation described by Wetter, in [6], who envisions an “at the and system typology.
press of a button” whole building energy simulation analysis. He also
emphasizes the need to integrate control strategy specifications into 9. Conclusions
BIM to avoid inconsistencies between design and operation. BII presents
a successful attempt to do this. However, as noted in scenario 2, the The use of BIM to automate the creation of BEPS models has gained
description of control strategies in IFC needs to be enriched to enable a momentum. However, most of the developed methodologies focus on
broader control definition, as already reported by Benndorf et al. [51]. the geometry or system aspects and often rely on additional data
Apart from the previous aspect, BEPS should ideally be open source, sources other than BIM to obtain a functional BEPS model. Most of them
user friendly, sufficiently accurate, capable of modeling buildings, are only compatible with the “detailed design” stage.
HVAC and controls [78]. BII fulfills those conditions since it relies on This paper investigates the direct translation of geometry, system

179
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

and control representations encapsulated in an IFC4 file into a performance simulation, CIB-W78 25th International Conference on Information
Modelica-based BEPS model. The applicability of this translation Technology in Construction, Santiago, Chile July 15-17, 2008, 2008 https://eetd.
lbl.gov/sites/all/files/publications/919e.pdf [Accessed on: 2018-01-31].
throughout the different life cycle stages is explored by taking ad- [9] D. Ilter, E. Ergen, BIM for building refurbishment and maintenance: current status
vantage of Modelica's flexibility. and research directions, Struct. Surv. 33 (3) (2015) 228–256, http://dx.doi.org/10.
To this end, a BIM to BEPS workflow (BII), applicable throughout 1108/SS-02-2015-0008.
[10] X. Zhao, A scientometric review of global BIM research: analysis and visualization,
the entire building lifecycle, was developed. BII defines the constraints Autom. Constr. 80 (2017) 37–47, http://dx.doi.org/10.1016/j.autcon.2017.04.002.
for Modelica library components to be compatible with BII at the [11] T. El-diraby, T. Krijnen, M. Papagelis, BIM-based collaborative design and socio-
considered stage. It also includes a formal specification of the data technical analytics of green buildings, Autom. Constr. 82 (2017) 59–74, http://dx.
doi.org/10.1016/j.autcon.2017.06.004.
exchange process which provides the essential tools (guideline and [12] Y. Liu, S.V. Nederveen, M. Hertogh, Understanding effects of BIM on collaborative
MVD) to ensure that the required information is contained in the BIM design and construction: an empirical study in China, Int. J. Proj. Manag. 35 (2017)
and correctly exported to IFC. A PYTHON-based tool (Ifc2Modelica) 686–698, http://dx.doi.org/10.1016/j.ijproman.2016.06.007.
[13] A. Ghaffarianhoseini, J. Tookey, A. Ghaffarianhoseini, N. Naismith, S. Azhar,
capable of translating the geometry, system and control information
O. Efimova, K. Raahemifar, Building Information Modelling (BIM) uptake: clear
contained in IFC into a Modelica-based BEPS was developed. benefits, understanding its implementation, risks and challenges, Renew. Sust.
Ifc2Modelica generates Modelica models with a convenient graphical Energ. Rev. 75 (2017) 1046–1053, http://dx.doi.org/10.1016/j.rser.2016.11.083.
layout representation that facilitates modifications of the model. [14] H.-y. Chong, C.-y. Lee, X. Wang, A mixed review of the adoption of Building
Information Modelling (BIM) for sustainability, J. Clean. Prod. 142 (2017)
Two scenarios comparing a typical BIM approach with BII empha- 4114–4126, http://dx.doi.org/10.1016/j.jclepro.2016.09.222.
sized the advantages of having a formal, machine readable specification [15] M. Oraee, M.R. Hosseini, E. Papadonikolaki, R. Palliyaguru, M. Arashpour,
of the data exchange and a direct mapping tool. This study demon- Collaboration in BIM-based construction networks: a bibliometric-qualitative lit-
erature review, Int. J. Proj. Manag. 35 (7) (2017) 1288–1301, http://dx.doi.org/10.
strates that the IFC standard, through the possibility to define custom 1016/j.ijproman.2017.07.001.
property sets, business rules, and MVDs is flexible enough to carry the [16] C. Sun, S. Jiang, M.J. Skibniewski, Q. Man, A literature review of the factors lim-
information for BEPS modeling. Although improvements will be needed iting the application of BIM in the construction industry, Technol. Econ. Dev. Econ.
23 (5) (2017) 764–779, http://dx.doi.org/10.3846/20294913.2015.1087071.
in some areas [48, 51], an IFC4 file can be used as the main data source [17] D. Bryde, M. Broquetas, J.M. Volm, The project benefits of Building Information
to automatically and directly translate the contained geometry, system Modelling (BIM), Int. J. Proj. Manag. 31 (7) (2013) 971–980, http://dx.doi.org/10.
and control representations into a Modelica-based BEPS model. The 1016/j.ijproman.2012.12.001.
[18] S. Azhar, Building Information Modeling (BIM): trends, benefits, risks, and chal-
capabilities of IFC4 combined with the developed direct translation lenges for the AEC industry, Leadersh. Manag. Eng. 11 (3) (2011) 241–252, http://
strategy (Ifc2Modelica) and Modelica's flexibility constitute a flexible dx.doi.org/10.1061/(ASCE)LM.1943-5630.0000127.
and adaptable BEPS workflow and enable to avoid switching between [19] C. Eastman, P. Teicholz, R. Sacks, K. Liston, BIM Handbook: A Guide to Building
Information Modeling for Owners, Managers, Designers, Engineers and Contractors,
several BEPS tools throughout the different stages of the building's
John Wiley & Sons, 2011 ISBN 1118021.9X, 9781118021.99.
lifecycle. [20] A. Andriamamonjy, R. Klein, D. Saelens, IFC-assisted building energy performance
Nonetheless, close collaboration and commitment of all stake- simulation implementation. Development of a python package, Proceedings of the
holders are crucial to ensure that the relevant information is integrated 3rd IBPSA-England Conference BSO 2016, Great North Museum, Newcastle, 12th-
14th September 2016, 2016 http://www.ibpsa.org/proceedings/BSO2016/p1094.
correctly into the BIM. Technological challenges, such as the lack of pdf no. 1094 [Accessed on: 2018-01-31].
custom MVD support, and limited compliance with IFC4 still restrict to [21] V. Bazjanac, T. Maile, IFC HVAC interface to EnergyPlus - a case of expanded in-
some extent the use of BII. Considering the remaining barriers limiting teroperability for energy simulation, IBPSA-USA National Conference, SimBuild
2004, 2004 http://escholarship.org/uc/item/5zb4j9nb [Accessed on: 2018-01-31].
BII's use, future research should further develop a BEPS-specific [22] R.J. Hitchcock, J. Wong, Transforming IFC architectural view BIMS for energy si-
strategy that improves the collaboration of all concerned parties. mulation, 12th Conference of International Building Performance Simulation
Association, Sydney, 14–16 November 2011, 2011 http://www.ibpsa.org/
proceedings/BS2011/P_1391.pdf no. 1391 [Accessed on: 2018-01-31].
Acknowledgments [23] J.B. Kim, W. Jeong, M.J. Clayton, J.S. Haberl, W. Yan, Developing a physical BIM
library for building thermal energy simulation, Autom. Constr. 50 (2014) 16–28,
This project received funding from KU Leuven, the European Union, http://dx.doi.org/10.1016/j.autcon.2014.10.011.
[24] W. Jeong, J.B. Kim, M.J. Clayton, J.S. Haberl, W. Yan, Translating Building
the European Regional Development Fund ERDF, Flanders Innovation &
Information Modeling to Building Energy Modeling using model view definition,
Entrepreneurship and the Province of Limburg. Sci. World J. (2014) 21., http://dx.doi.org/10.1155/2014/638276.
[25] I.J. Ramaji, J.I. Messner, R.M. Leicht, Leveraging Building Information Models in
IFC to perform energy analysis in Openstudio, ASHRAE and IBPSA-USA SimBuild
References
2016, 2016 http://ibpsa-_usa.org/index.php/ibpusa/article/view/365 [Accessed
on: 2018-01-31].
[1] H. Kim, Z. Shen, I. Kim, K. Kim, A. Stumpf, J. Yu, BIM IFC information mapping to [26] G. Reynders, A. Andriamamonjy, R. Klein, D. Saelens, IFC-Modelica tool facilitating
building energy analysis (BEA) model with manually extended material informa- model complexity selection for building energy simulation, Proceedings of the 15th
tion, Autom. Constr. 68 (2016) 183–193, http://dx.doi.org/10.1016/j.autcon.2016. International conference of the International Building Performance Simulation
04.002. Association (IBPSA), BS2017, 2017 http://www.ibpsa.org/proceedings/BS2017/
[2] T. Maile, M. Fischer, V. Bazjanac, Building Energy Performance Simulation Tools - a BS2017_621.pdf no. 621 [Accessed on: 2018-01-31].
Life-Cycle and Interoperable Perspective, Tech. rep. CIFE (Center for Integrated [27] V. Dimitriou, S. Firth, T. Hassan, F. Fouchal, BIM enabled building energy model-
Facility Engineering), Civil and Environmental Engineering Dept, Stanford ling: development and verification of a GBXML to IDF conversion method,
University, Stanford University, 2007, https://stacks.stanford.edu/file/ Proceedings of the 3rd IBPSA-England Conference BSO 2016, Great North Museum,
druid:ww813hh4225/WP107.pdf , Accessed date: 31 January 2018. Newcastle, 12th-14th September 2016. 2016 http://www.ibpsa.org/proceedings/
[3] D.B. Crawley, L.K. Lawrie, F.C. Winkelmann, W.F. Buhl, Y.J. Huang, C.O. Pedersen, BSO2016/p1126.pdf no. 1126 [Accessed on: 2018-01-31].
R.K. Strand, R.J. Liesen, D.E. Fisher, M.J. Witte, J. Glazer, EnergyPlus: creating a [28] P. Remmen, J. Cao, S. Ebertshäuser, J. Frisch, M. Lauster, T. Maile, D. Müller,
new-generation building energy simulation program, Energy Build. 33 (2001) C.V. Treeck, An open framework for integrated BIM-based building performance
319–331, http://dx.doi.org/10.1016/S0378-7788(00)00114-6. simulation using Modelica, 14th International Conference of the International
[4] Thermal Energy System Specialists LLC, TRNSYS, a transient system simulation Building Performance Simulation Association (IBPSA), BS2015, 2015 http://www.
program, http://www.trnsys.com/, (2011) , Accessed date: 31 January 2018. ibpsa.org/proceedings/BS2015/p2384.pdf no. 2384 [Accessed on: 2018-01-31].
[5] M. Wetter, Modelica-based modelling and simulation to support research and de- [29] Y.N. Bahar, C. Pere, J. Landrieu, C. Nicolle, A thermal simulation tool for building
velopment in building energy and control systems, J. Build. Perform. Simul. 2 and its interoperability through the Building Information Modeling (BIM) Platform,
(2009) 143–161, http://dx.doi.org/10.1080/19401490902818259 ISBN Buildings 3 (2013) 380–398, http://dx.doi.org/10.3390/buildings3020380.
1940149090281. [30] Y. Arayici, T. Fernando, V. Munoz, M. Bassanino, Interoperability specification
[6] J.L.M. Hensen, R. Lamberts, Building Performance Simulation for Design and development for integrated BIM use in performance based design, Autom. Constr.
Operation, Spon Press, 2011 ISBN 9780415474146. 85 (2018) 167–181, http://dx.doi.org/10.1016/j.autcon.2017.10.018.
[7] V. Bazjanac, T. Maile, C. Rose, J. O’Donnell, E. Morrissey, B.R. Welle, An assesment [31] B. Dong, K.P. Lam, Y.C. Huang, A comparative study of the IFC and gbXML in-
of the use of building energy performance simulation in early design, 12th formational infrastructures for data exchange in computational design support
International Conference of the International Building Performance Simulation environments Geometry information, International Conference of the International
Association (IBPSA), BS2011, 2011 http://ibpsa.org/proceedings/BS2011/P_1531. Building Performance Simulation Association (IBPSA), BS2007, 2007 http://www.
pdf no. 1531 [Accessed on: 2018-01-31]. ibpsa.org/proceedings/BS2007/p363_final.pdf [Accessed on: 2018-01-31].
[8] V. Bazjanac, IFC BIM-based methodology for semi-automated building energy [32] V. Bazjanac, T. Maile, C. Nytsch-Geusen, Generation of building geometry for

180
A. Andriamamonjy et al. Automation in Construction 91 (2018) 166–181

energy performance simulation using Modelica, Proceedings of the CESBP Central [56] United States General Services Administration, BIM Guide 05: Energy Performance,
European Symposium on Building Physics / BauSIM 2016, 2016 http://www.ibpsa. Tech. rep. 2015 www.gsa.gov/bim [Accessed on: 2018-01-31].
org/proceedings/bausimPapers/2016/D-_01-_1.pdf [Accessed on: 2018-01-31]. [57] International Organization for Standardization, ISO 29481-1: Building information
[33] buildingSMART International, buildingSMART, http://www.buildingsmart.org/, models - Information delivery manual - Part 1: Methodology and format, https://
(2017) , Accessed date: 31 January 2018. www.iso.org/standard/60553.html, (2018) , Accessed date: 31 January 2018.
[34] European Committee for Standardization (CEN), CEN/TC 442 - Building [58] T. Chipman, ifcDoc Tool Summary, http://www.buildingsmart-tech.org/
Information Modelling (BIM), https://standards.cen.eu/BP/1991542.pdf, (2017). specifications/specification-tools/ifcdoc-tool, (2012) , Accessed date: 31 January
[35] International Organization for Standardization, ISO/TC59 Buildings and civil en- 2018.
gineering works, https://committee.iso.org/home/tc59, (2018) , Accessed date: 31 [59] P. Fritzson, Introduction to Modeling and Simulation of Technical and Physical
January 2018. Systems with Modelica, John Wiley & Sons., 2011, pp. 157–168, http://dx.doi.org/
[36] M. Wetter, W. Zuo, T.S. Nouidui, X. Pang, Modelica Buildings library, J. Build. 10.1002/9781118094259 ISBN 9781118094259.
Perform. Simul. 7 (4) (2014) 253–270, http://dx.doi.org/10.1080/19401493.2013. [60] C. Nytsch-Geusen, J. Huber, M. Ljubijankic, J. Rädler, Modelica BuildingSystems
765506. âĹŠ eine Modellbibliothek zur Simulation komplexer energietechnischer
[37] DesignBuilder Software Ltd, DesignBuilder, https://www.designbuilder.co.uk/, Gebäudesysteme, Bauphysik 35 (1) (2013) 21.-29, http://dx.doi.org/10.1002/bapi.
(2018) , Accessed date: 31 January 2018. 201310045.
[38] Integrated Environmental Solutions, Virtual Environment 2017, https://www.iesve. [61] R. Baetens, R. De Coninck, F. Jorissen, D. Picard, L. Helsen, D. Saelens, OpenIDEAS -
com/VE2017, (2017) , Accessed date: 31 January 2018. An Open Framework for Integrated District Energy Assessment, 14th International
[39] NREL (National Renewable Energy Laboratory), OpenStudio, https://www. Conference of the International Building Performance Simulation Association
openstudio.net/, (2015) , Accessed date: 31 January 2018. (IBPSA), BS2015, 2015 http://www.ibpsa.org/proceedings/BS2015/p2243.pdf no.
[40] A. Cemesova, C.J. Hopfe, R.S. Mcleod, PassivBIM: enhancing interoperability be- 2243 [Accessed on: 2018-01-31].
tween BIM and low energy design software, Autom. Constr. 57 (2015) 17–32, [62] T. Maile, J. O’Donnell, V. Bazjanac, C. Rose, BIM Geometry Modelling Guidelines
http://dx.doi.org/10.1016/j.autcon.2015.04.014. for energy performance simulation, 13th International Conference of the
[41] K.-u. Ahn, Y.-j. Kim, C.-s. Park, I. Kim, K. Lee, BIM interface for full vs. semi-au- International Building Performance Simulation Association (IBPSA), BS2013, 2013
tomated building energy simulation, Energy Build. 68 (2014) 671–678, http://dx. http://www.ibpsa.org/proceedings/BS2013/p_1510.pdf no. 1510 [Accessed on:
doi.org/10.1016/j.enbuild.2013.08.063. 2018-01-31].
[42] V. Bazjanac, Spaceboundary requirements for modeling of building geometry for [63] E. El Asmi, S. Robert, D. Mazza, K. Zreik, Multi-physical Model Simulation and
energy and other performance simulation, Proceedings of the CIB W78 2010: 27th Interoperability through BIM/ IFC using COMETH Simulation Engine, Architecture,
International Conference Cairo, 2010, pp. 16–18 Egypt http://www.cibw78.org/ City & Information Design, EuropIA.14: 14th International conference on the
conferences/ , Accessed date: 31 January 2018. Advances in Design Sciences and Technology, 2014 ISBN 979-10-90094-18-5.
[43] M. Weise, T. Liebich, R. See, V. Bazjanac, T. Laine, IFC implementation guide: space [64] R. Steinmann, K. Linhard, G. Dangl, G. Calleja-rodriguez, R. Guruz, Collaboration
boundaries for thermal analysis, Tech. rep. US Government Services Administration requirements and interoperability fundamentals in BIM based multi-disciplinary
(GSA), 2009, http://www.buildingsmart-_tech.org/downloads/accompanying-_ building design processes, eWork and eBusiness in Architecture, Engineering and
documents/agreements/IFC2x3-_space-_boundary-_implAgreement-_2009-_09-_17. Construction: ECPPM, 2016 ISBN 978-1-315-38690-4.
pdf [Accessed on: 2018-01-31]. [65] J. Wix, J. Karlshøj, Information Delivery Manual Guide to Components and
[44] D. Ladenhauf, K. Battisti, R. Berndt, E. Eggeling, D.W. Fellner, M. Gratzl-michlmair, Development Methods, Tech. rep. buildingSMART International, 2010, http://iug.
T. Ullrich, Computational geometry in the context of building information mod- buildingsmart.org/idms/development/IDMC_004_1_2.pdf.
eling, Energy Build. 115 (2016) 78–84, http://dx.doi.org/10.1016/j.enbuild.2015. [66] I.J. Ramaji, A.M. Memari, Extending the current model view definition standards to
02.056. support multi-storey modular building projects. Archit. Eng. Des. Manag. 14 (2017)
[45] G.N. Lilis, G.I. Giannakis, D.V. Rovas, Automatic generation of second-level space 158–176, http://dx.doi.org/10.1080/17452007.2017.1386083.
boundary topology from IFC geometry inputs, Autom. Constr. 76 (2017) 108–124, [67] T. Liebich, T. Chipman, IFC4 Add1 Specification, http://www.buildingsmart-tech.
http://dx.doi.org/10.1016/j.autcon.2016.08.044. org/ifc/IFC4/Add1/html/, (2015) , Accessed date: 31 January 2018.
[46] C.M. Rose, V. Bazjanac, An algorithm to generate space boundaries for building [68] CIC-RP (Computer Integrated Construction Research Program), BIM Project
energy simulation, Eng. Comput. 31 (2013) 271–280, http://dx.doi.org/10.1007/ Execution Planning Guide version 2.1, Tech. rep. Pennsylvania State Univ.,
s00366-013-0347-5. University Park, PA, 2011, http://bim.psu.edu/project/resources/ [Accessed on:
[47] J. O’Donnell, R. See, C. Rose, T. Maile, V. Bazjanac, P. Haves, SIMMODEL: a domain 2018-01-31].
data model for whole building energy simulation, 12th International Conference of [69] The British Standards Institution (bsi), Publically Available Specification (PAS)
the International Building Performance Simulation Association (IBPSA), BS2011, 1192-2: Specification for Information Management for the Capital/Delivery Phase
2011 http://www.ibpsa.org/proceedings/BS2011/P_1223.pdf no. 1223 [Accessed of Construction Projects using Building Information Modelling, (2013) ISBN 978 0
on 2018-01-31]. 580 82666 5.
[48] S. Pinheiro, J. O’Donnell, R. Wimmer, V. Bazjanac, S. Muhic, T. Maile, J. Frisch, [70] M. Oh, J. Lee, S. Wan, Y. Jeong, Integrated system for BIM-based collaborative
C. Van Treeck, Model view definition for advanced building energy performance design, Autom. Constr. 58 (2015) 196–206, http://dx.doi.org/10.1016/j.autcon.
simulation, Proceedings of the CESBP Central European Symposium on Building 2015.07.015.
Physics / BauSIM 2016, 2016 http://www.ibpsa.org/proceedings/bausimPapers/ [71] K. Stapleton, B. Gledson, Z. Alwan, Understanding technological interoperability
2016/D-_02-_1.pdf [Accessed on: 2018-01-31]. through observations of data leakage in Building Information Modelling (BIM)
[49] W. Reinhard, S. Pinheiro, T. Maile, J. O’Donnell, J. Cao, V. Bazjanac, J. Frisch, based transactions, Proceedings of the 32nd eCAADe Conference, 10–12 September
C.V. Treeck, Improving information exchange between Building Information 2014, Northumbria University, Newcastle Upon Tyne. 2014 http://nrl.
Modelling (BIM) and Advanced Building Energy Performance Simulation (BEPS), northumbria.ac.uk/17696/ no. 052 [Accessed on: 2018-01-31].
Gebaudetechnik Wiss. Prax. 4 (2017) 276–291 ISBN 2195-643X. [72] T. Krijnen, A. Kämäräinen, IfcOpenShell, http://ifcopenshell.org/python.html,
[50] S. Imam, D.A. Coley, I. Walker, The building performance gap: are modellers lit- (2011) , Accessed date: 31 January 2018.
erate? Build. Serv. Eng. Res. Technol. 38 (3) (2017) 351–375, http://dx.doi.org/10. [73] R. Santos, A.A. Costa, A. Grilo, Bibliometric analysis and review of Building
1177/0143624416684641. Information Modelling literature published between 2005 and 2015, Autom.
[51] G. Benndorf, N. Rehault, M. Clairembault, T. Rist, Describing HVAC controls in IFC Constr. 80 (2017) 118–136, http://dx.doi.org/10.1016/j.autcon.2017.03.005.
method and application, Energy Procedia 122 (2017) 319–324, http://dx.doi.org/ [74] F. Jorissen, R. Baetens, D. Picard, IDEAS v1.0.0, https://github.com/open-ideas/
10.1016/j.egypro.2017.07.330. IDEAS, (2017) , Accessed date: 31 January 2018.
[52] M. Thorade, J. Rädler, P. Remmen, T. Maile, R. Wimmer, J. Cao, M. Lauster, [75] G. Mignone, M.R. Hosseini, N. Chileshe, M. Arashpour, Enhancing collaboration in
D. Müller, C. Nytsch-geusen, C. Van Treeck, An open toolchain for generating BIM-based construction networks through organisational discontinuity theory : a
Modelica code from Building Information Models, Proceedings of the 11th case study of the new Royal Adelaide Hospital, Archit. Eng. Des. Manag. 12 (5)
International Modelica Conference September 21.23, 2015, Versailles, France, (2016) 333–352, http://dx.doi.org/10.1080/17452007.2016.1169987.
2015, http://dx.doi.org/10.3384/ecp15118383. [76] Verein Deutscher Ingenieure (VDI), VDI 3805-1: Product data exchange in the
[53] R. Wimmer, J. Cao, P. Remmen, T. Maile, J.O. Donnell, J. Frisch, R. Streblow, Building Services - Fundamentals, Tech. rep. Beuth Verlag GmbH, 2011, http://
D. Müller, C.V. Treeck, Implementation of advanced BIM-based mapping rules for www.vdi3805.eu/en/vdi-3805.html , Accessed date: 13 March 2018ICS 35.240.99,
automated converions to Modelica, 14th International Conference of the 97.120..
International Building Performance Simulation Association (IBPSA), BS2015, 2015 [77] International Organization for Standardization (ISO), ISO 16757-2015: Data
http://www.ibpsa.org/proceedings/BS2015/p2424.pdf no. 2424 [Accessed on: structures for electronic product catalogues for building services, Tech. rep. 2015
2018-0131]. https://www.iso.org/standard/57613.html , Accessed date: 13 March 2018ICS
[54] A. Constantin, R. Streblow, D. Muller, The Modelica HouseModels Library: pre- 91.010.01..
sentation and evaluation of a room model with the ASHRAE Standard 140, [78] E. Atam, Current software barriers to advanced model-based control design for
Proceedings of the 10 th International Modelica Conference; March 10–12, 2014, energy-efficient buildings, Renew. Sustain. Energy Rev. 73 (2017) 1031–1040,
Lund, Sweden, no. 96, 2014, pp. 293–299, , http://dx.doi.org/10.3384/ http://dx.doi.org/10.1016/j.rser.2017.02.015.
ecp14096293. [79] S. Pinheiro, R. Wimmer, T. Maile, J. O’Donnell, J. Cao, V. Bazjanac, J. Frisch, C. van
[55] University of Indiana, USA, BIM Guidelines & Standards for Architects, Engineers, Treeck, MVD based Information Exchange between BIM and Advanced Building
and Contractors, Tech. rep. University of Indiana, USA, 2015, http://www.iu.edu/ Energy Performance Simulation, Autom. Constr. 90 (2018) 91–103, http://dx.doi.
~vpcpf/consultant-_contractor/standards/bim-_standards.shtml [Accessed on: org/10.1016/j.autcon.2018.02.009 (issn: 0926-5805).
2018-01-31].

181