LIST OF FIGURES
Fig. 1 UML diagram of the relationship between wall and its dimension in IFC ....................... 7
Fig. 2 Geometry and sensor representations in gbXML .............................................................. 8
Fig. 3 UML diagram of the relationship between sensor and sensed data in IFC ....................... 8
1. Executive Summary
A detailed identification and analysis is conducted on the interoperability gaps between Nbims/Open
standards and building performance simulation tools. As two prevalent Informational Infrastructures in
the architecture, engineering and construction (AEC) industry, IFC and gbXML are investigated and
compared with illustrated examples about geometry and sensor/sensed data. To evaluate the capabilities,
limitations and characteristics of various dominant building modeling and simulation tools,
interoperability issues between these tools are tabulated and analyzed in terms of the distortion of
building model information and the loss of geometric precision. Different types of elements in the
building model are used as illustrations in the comparison of the methods utilized in these tools in
handling the building models imported from other sources. The evaluations of the investigated tools are
summarized in detail for the reference of further study in solving these interoperability issues.
The challenges to design complex high performance buildings demand a paradigm shift from the
traditional sequential to an interactive concurrent design environment (Dong, Lam et al. 2007). This
implies the need to adopt new integrative design processes and the associated computer-based design
support tools need to be re-conceptualized and re-engineered within an integrated IT environment. Such a
system must not only be capable of supporting concurrent performance evaluation of building design and
engineering across domains, but be accessible across geographically distributed environments through
both synchronous and asynchronous modes of communication (Lam, Wong et al. 2004).
With the development of information technology, significant progress has been made in the area of
common data exchange in the building industry. The first generation of data exchange models such as
STEP (ISO 1992), COMBINE (Augenbroe 1992), ICADS (Pohi and Rep 1988), ARMILLA (Gauchel,
Hovestadt et al. 1992), and IBOM (Sanvido 1992) focus mainly on building geometry information. It is
necessary but evidently insufficient for any performance-based design support. The second generation
data exchange models expanded to include information that is required for particular domain-specific
performance modeling, such as energy/thermal comfort simulation, HVAC design, lighting design,
building life-cycle analysis and building plan checking. The hierarchical models (Bedell and Kohler 1993)
and KNODES (Rutherford 1993) for energy and life cycle cost modeling are examples. These object-
oriented data models begin to address the larger question of interoperability between various
performance domains in building design. Perhaps more importantly, it offers new opportunities for
integrative concurrent design support (Lam et al., 2004). The current third generation models are able to
potentially encapsulate all information related to a building in a comprehensive data schema and
facilitate the sharing of pertinent information throughout the entire building life-cycle, from design (for
evaluating total building performance) to construction (for evaluating cost and schedules) to operation
(occupancy and environmental sensing and system controls).
3. Prevalent Informational Infrastructures in the Industry: IFC and gbXML
One key factor in the success of any data schema is in the wide-spread adoption by all constituencies in
the architecture, engineering and construction (AEC) industry. Currently, the Industry Foundation Class
(IFC) and Green Building XML (gbXML) are two main and prevalent informational infrastructures in the
industry.
Industry Foundation Classes is a vendor neutral, open data exchange specification. It is an object oriented
file format developed for the building industry and is commonly used in Building Information Modeling
to facilitate interoperability between software platforms. IFC was originally developed in 1995 by a group
of American and European AEC firms and software vendors through the International Alliance for
Interoperability (IAI). Since 2005 it has been maintained by building SMART International (ASHRAE
2005). The goal of IFC is to provide a universal basis for process improvement and information sharing in
the construction and facilities management industries. EXPRESS-G is used to identify classes, the data
attributes of classes and the relationships that exist between classes in IFC (IAI 2006). It is a graphical
modeling notation developed within the ISO-STEP community for representing application interpreted
models. Some major implementations of this schema include CORENET (Singapore), BuildingSMART
(Australia), SARA (Finland), Building Lifecycle Interoperable Cost (BLIS), Simple Access to the
Building Lifecycle Exchange (SABLE) and BSPro (Bazjanac and Maile 2004).
Green Building XML (gbXML) is developed based on the XML format. It uses an XML-based schema to
represent the building information, such as geometry and material properties), for energy simulation
purpose. It was developed by Green Building Studio (formerly GeoPraxis) with the support of the
California Energy Commission Public Interest Energy Research (PIER) Program, and the California
Utilities (Pacific Gas and Electric Company, Southern California EDISON and Sempra Energy Utility), in
order to facilitate the transfer of building information stored in CAD building information models,
enabling integrated interoperability between building design models and a wide variety of engineering
analysis tools and models available today.
gbXML has the industry support and wide adoption by the leading CAD vendors and HVAC software
vendors. With the development of export and import capabilities in several major engineering modeling
tools, gbXML has become a defacto industry standard schema. Its use dramatically streamlines the
transfer of building information to and from engineering models, eliminating the need for time consuming
plan take-offs. This removes a significant cost barrier to designing resource efficient buildings and
specifying associated equipment. It enables building design teams to truly collaborate and realize the
potential benefits of Building Information Modeling (Faye C McQuiston 2005). Several popular CAD
software (e.g., Autodesk Revit, Architectural Desktop, and ArchiCAD) and energy analysis applications
(e.g., DOE-2. e-QUEST, HAP) can exchange data using this schema, through the Green Building Studio
web service (DOE-2, 2007).
4. Comparison of Informational Infrastructures for Data Exchange in Computational
Design Support Environments
In this section, a comparative study of the IFC and gbXML informational infrastructures for data
exchange in computational design support environments is conducted with illustrated examples about
geometry and sensor/sensed data.
IFC gbXML
Geometry information
Fig.1 shows how the IFC schema represents the In order to compare with IFC approach, the same
coordinates and dimensions of an IfcWall object. examples are given in gbXML approach.
Sensor information
Fig.3 shows the UML diagram of the IFC schema As shown in Fig.2, gbXML has an element named
representations of sensors and sensed data using Meter which is a description of a resource
temperature and CO2 parameters as examples. An measurement. Any measured data can be defined in
IfcSensorType defines a particular type of sensor which this element. It has children elements: Name,
is used for detection in a control system such as a Description, UtilityRate and others."Name" defines a
building automation control. Parameters of the sensors name of sensor, while"other" could contain the value
are specified through property sets that are enumerated and unit of this sensor and it has any occurrences.
in the IfcSensorTypeEnum data type. IfcSensorType has Using the same temperature and CO2 parameters, the
an attribute called PreDefinedType, which is an object generated XML file could be as follows:
of IfcSensorTypeEnum. IfcSensorTypeEnum defines the
range of different types of sensor that can be specified. There are only three levels from root element
A temperature sensor and CO2 sensor type are instances "gbXML" to data element"value". In addition, there is
of IfcSensorTypeEnum. The usage of IfcSensorType no need to contain"UtilityRate" element in "Meter"
defines the parameters for one or more occurrences of element if there was no UtilityRate information.
IfcDistributionControlElementType, which inherited Furthermore, a simple XML stylesheet can extract the
from the IfcDistributionElementType This is how information from the above XML file. Hence, to store
sensors are connected with the equipment. and obtain the sensor data using gbXML is relatively
Here is an example of how the sensor values are simple and straightforward (Mazzeo 2011).
represented. IfcPropertySingleValue is one of the
subtypes of IfcSimpleProperty. It has two attributes:
NorminalValue and Unit. The NorminalValue is an
object of IfcValue, while Unit is an IfcUnit object,
where "parts per million (ppm)" is stored. IfcValue is a
SELECT type for selecting between more specialized
types IfcSimpleValue, IfcMeasureValue and
IfcDerivedMeasureValue. Attributes of
IfcMeasureValue include
IfcThermodynamicTemperatureMeasure and
IfcRatioMeasure which are REAL types.
IfcThermodynamicTemperatureMeasure is where the
sensed temperature value is stored, while
IfcRatioMeasure is where the sensed CO2 value is
stored (IAI, 2006).
Fig. 1 UML diagram of the relationship between wall and its dimension in IFC
Fig. 2 Geometry and sensor representations in gbXML
Fig. 3 UML diagram of the relationship between sensor and sensed data in IFC
5. Pros and Cons of the Technical and Semantic Implementations in IFC and gbXML
The pons and cons of the technical and semantic implementations in IFC and gbXML can be analyzed
from the following two aspects:
Firstly, IFC adopts a comprehensive and generic approach to represent an entire building project. From
IAI online information website, it shows a holistic approach of IFC representation, covering domains
from building construction to building operation. IFC representation was also extended in the building
commission domain and implemented in several cases studies. The application of gbXML deployed by
Green Building Studio Inc. is currently only on the energy simulation domain. Users can upload a well-
formatted gbXML file to get a quick summary of simulation results from DOE-2 based on contextually
related assumptions. It is also possible to get input files for EnergyPlus and eQuest from GBS after the
simulation runs. However, gbXML has the ability to carry building environmental sensing information.
gbXML schema can be extended for certain purposes such as conducting lighting simulation. In terms of
geometry, the generic approach of IFC has the ability to represent any shape of building geometry, while
gbXML only accepts rectangular shape, which is enough for energy simulation.
Secondly, IFC uses a "top-down" and relational approach, which yields in a relative complex data
representation schema and a large data file size. gbXML adopts a "bottom-up" approach, which is flexible,
open source, and a relatively straight-forward data schema. The "top-down" approach can trace back all
the semantic changes when one value of the element in the schema changed. Ideally, it has the ability to
maintain semantic integrity automatically. However, it is very complex to program and be implemented in
software. The "bottom-up" approach has less layer of complexity. This approach has proven successful in
offering web-based simulation service (notably Green Building Studio) for the industry.
In this section, the interoperability issues between various dominant building modeling and simulation
tools are analyzed in terms of the distortion of building model information (e.g., definition of thermal
zones, building materials and HVAC system) and the loss of geometric precision.
All the tools investigated in the study are current dominant tools in the building modeling and simulation
field, including DesignBuilder, Energyplus, eQUEST, Autodesk Revit, Autodesk CAD, Google Sketchup
and Autodesk Ecotect. Different types of elements in the building model, such as building openings,
adiabatic component blocks, external louvres and air walls, are used as illustrations to compare the
methods utilized in different tools in handling the building models imported from other sources, which is
useful to analyze the capabilities, limitations and characteristics of these tools.
The evaluations of the investigated tools are summarized and tabulated in detail for the reference of
further study in solving these interoperability issues.
GOOGLE SKETCH UP (OPEN STUDIO
DESIGN BUILDER (DB) V2.3 ENERGYPLUS V6.0 [DXF Output] Problem EXPLANATION
PLUG IN)
Generation of sub-surfaces and
Design Builder 3D GUI EnergyPlus model exported from Design
Google Sketch up model exported from associated model elements for
View from Building Level Builder
EnergyPlus building openings of non-
(Openings on the building surfaces) (3D DXF output after a simulation run is
(3D view with X-RAY surfaces) rectangular geometry developed in
(Holes, doors, windows, vents) shown)
DB and exported to EnergyPlus
Room 7 is successfully
differentiated from the Room objects identified by
surrounding room objects room separation lines (and
Room separation surface is surfaces) which are also used
missing. Thermal zone defined by to separate two adjacent zones
Room 7 is ignored and merged by air surfaces (without a
with the adjacent meeting room thermal mass) in Design
Builder (virtual partitions) and
(Room 6)
EnergyPlus (air walls)
environments cannot be
successfully exported to
Design Builder v2.3 through
the use of REVIT generated
Plan View gbXML models.
3D View Rooms Separation surface is also
in place as seen from Although analysis of gbXML
Room Separation line between analytical surfaces building model shows
office space and the meeting analysis existence of such room
room surfaces, it is seen that Design
Builder gbXML import
functions cannot differentiate
separation surfaces causing
merging/combination of the
3D Block View from DB two thermal zones that are
supposed to be divided by a
separation surface. Loss of
Section View model detail is observed by
missing thermal zones.
AutoDesk REVIT gbXML Export Analysis: Visual Parts of the floor slab projecting out from Geometry of the EnergyPlus model after being
analysis of the analytical surfaces exported from the internal geometry is identified as exported to Google Sketch Up. Same external
REVIT Model as gbXML file indicates that internal external, detached shading surfaces (and shading surface attached to the ground is
surface (inward-facing) of the floor slab (tagged as attached to the ground, z=0 elevation) by exported to EnergyPlus model. The unused
an external element) projects to the outer surface of Design Builder. A continuous strip of a surface here is seen as ring encircling the
the ext. walls. However, room object geometry is shading element is seen at the ground level. building perimeter. The distance between the
bounded by the inside surface of the ext. walls. This Such an element has no functionality to the strip of unused shading element and external
situation denotes a possible discrepancy in the building model analyzed with Design wall surface of the thermal zone is equal to
exported floor slab geometry. Builder. ext. wall thickness (if room objects are defined
from interior face) of half of it (if room objects
are defined from wall centroid).
AUTODESK REVIT Architecture 2011
AUTODESK REVIT Architecture 2011 DESIGN BUILDER V2.3 Problem EXPLANATION
(gbXML)
Room objects enclosing other
AutoDesk REVIT gbXML Export View Views from Building Level
AutoDesk REVIT 3D Rendered + Plan View room objects are missing in after
(gbXML Rooms 3D & Analytical Surfaces (Axonometric + Rendered Perspective)
Design Builder export
roof plan)
Room objects
enclosing/encircling other room
objects cannot be successfully
exported from REVIT medium to
Design Builder through the use of
gbXML export schema. Although
enclosed room objects are
exported without lack of
geometric detail, geometry of
enclosing room objects cannot be
identified. This situation results
3D View Rooms in entire omission of enclosing
room objects.
Axonometric Model View
Encircling Room 10 is present in gbXML
3D Rendered View Room object encircled by another model. Room 10 floor & roof surfaces are Furthermore, with the omission
Missing thermal zone in Design Builder.
fragmented into multiple triangular and of such room objects/thermal
room object in REVIT model Remaining zones are assumed to be zones, remaining thermal zones
trapezoidal sub-surfaces exposed to outside conditions are assumed to have all surfaces
adjacent to outside/exterior. The
structure of the entire simulation
model is altered drastically.
Room objects are not transferred
as thermal zones of a single
building block conversely they
are considered as distinct building
blocks.
When, Ecotect model is imported from Design Builder v2.0 in the form of a When Ecotect model is imported from Design Builder v2.0 in the form
simple DXF format, there exist many redundant tangling surfaces. Many surfaces of detailed RADIANCE format (*.RAD files), it is seen that fewer
are triangulated increasing the modeling complexity. Some surfaces for detailed redundant surfaces exists. With RAD import, window surfaces are
fenestration geometry are missing from the model. With DXF import, all surfaces assigned with glazing materials (but still not the same specification
are assigned with a brick material (whether transparent or opaque). Thermal zone originally given in Design Builder program). Thermal zone distinctions
distinctions and related information are missed in the imported model. are also missed in the imported model.
Design Builder v3.0.0.091 Design Builder v3.0.0.091 Design Builder v3.0.0.091 Problem EXPLANATION
Customized glass pane definitions
for spectral data type are not
Glazing Data Input View - Main Glazing Data Input View - Main Glazing Data Input View Sub-Category
allowed for user defined material
libraries
Glass Pane Modeling Dialog Box Type 1 Glass Pane Modeling Dialog Box Type 2 Glass Pane Modeling Spectral Input Dialog
Box Design Builder v3.0.0.091 program
doesnt allow users to create
customized glass material libraries
in which glass panes are defined
with spectral entries instead of
broadband (averaged) type.
Customization of default program
library is already prohibited,
however, when user creates a new
object from scratch or duplicates an
existing object then any
modification is allowed. Due to
current structure of design Builder
program such functionality is not
working properly for glass panes to
be defined with spectral data
entries.
During glazing data input, Design Builder Model libraries can be updated by the users for Duplicating existing object and converting
normally does not allow users to modify default glazing data type of broadband (as explained library type from locked to model have no effect
libraries (locked library data). However, users above). However, with spectral type data on the restrictions. Any modification is restored
can create their own libraries and new objects by modifications, updates, and adding new data to to the original after refreshing the current screen.
starting from scratch or duplicating an existing the library are still prohibited even though data is
object (model library data) in model library (which should be customizable).
Design Builder v3.0.0.091 Problem EXPLANATION
Limitations of assignment
different thermal zones into
HVAC template library and modeling input and assignment views
separate air loop of the same
HVAC type
Two HVAC templates for the same type When users select, Compact
HVAC design detail within
(Constant Volume DX Unitary Multi-zone) are
Design Builder (which is the
created (by duplication method) and current most common and easy to use
HVAC template library is updated. To create two type where template HVAC
different air-loop in the same building model, systems can be assigned to
different thermal zones are assigned to different thermal zones and later
HVAC templates (of the same type). However, exported in full detail to
analysis of simulation models to EnergyPlus EnergyPlus program),
developing multiple air-loops
revealed that only one of the assigned systems is
of the same type of system is
assumed for the whole building and all thermal not possible. Such as, users
zones are grouped in the same air-loop of this cannot create two different air-
single HVAC system. loops of constant volume DX
system, by grouping different
Design Builder HVAC Template Library thermal zones and assigning
them to different HVAC
objects of the same type.
3D Model View with roof assembly composed of multiple surfaces (for eave 3D Model View with roof assembly exported with correct geometry for
sections) eave sections (defined as shading type surfaces)
In Revit modeling environment, building roof assembly can be defined as a Different roof assembly definition methods have different results when
continuous surface covering thermal zones underneath and extending outwards to exported to eQUEST environment (through gbXML format).
form eaves on the sides. Alternatively, same roof assembly can be defined as Continuous roof assembly is broken into multiple surfaces and the ones
separate/distinct multiple surfaces covering thermal zones, and forming eave assigned as shading type (for eave sections) are distorted and have
sections on different sides of the building. irregular geometry (irregular azimuth angles).
7. References
Augenbroe, G. (1992). "Integrated building performance evaluation in early design stages." Building and
Environment 27(2).
Bazjanac, V. and T. Maile (2004). "IFC HVAC Interface to EnergyPlus - A Case of Expanded
Interoperability for Energy Simulation." Proceedings of the SimBuild 2004 Conference in Boulder, CO.
Dong, B., K. P. Lam, et al. (2007). "A comparative study of the IFC and gbXML informational
infrastructures for data exchange in computational design support environments." Proceedings: Building
Simulation 2007.
Faye C McQuiston, Jerald D Parker, Jeffery D Spitler (2005). "Heating, Ventilating and Air Conditioning
Analysis and Design." John Wiley & Sons, Inc.
Gauchel, J., S. Hovestadt, et al. (1992). "Building modeling based on concepts of autonomy." Proceedings
of AID 1992, Pittsburgh, PA.
ISO (1992). Product data representation and exchange part 1: Overview and Fundamental Principles
Lam, K., N. Wong, et al. (2004). "SEMPER-II: an internet-based multi-domain building performance
simulation environment for early design support." Automation in Construction 13(15).
Mazzeo, N. (2011). Chemistry, Emission Control, Radioactive Pollution and Indoor Air Quality InTech
Pohi, J. and I. Rep (1988). "An integrated intelligent CAD environment. Proceedings." 4th International
Conference on Systems Research, Informatics and Cybernetics, Baden-Baden, Germany.
Rutherford, J. (1993). " KNODES: knowledged-based design decision support. ." CADD Future.
Sanvido, V. (1992). "Linking levels of abstraction of a building design." Building and Environment 27(2).