Improving the Interoperability of gbXML Data Model through Redefining Data Mapping Rules of HVAC Systems

Abstract

Green Building XML (gbXML) is one of the most prevalent data models used for Building Information Modeling (BIM). It enables interoperability between BIM and Building Energy Modeling (BEM). BEM is often used in building performance analysis software tools. However, through interviewing engineers and building energy modeling professionals in the industry, it turned out that most gbXML files are only used for importing and exporting building geometry information. Information such as Heating, Ventilation, and Ait-conditioning (HVAC) systems and internal loads are rarely handled due to the lack of functionality in current BIM software. Taking account of more than 15% of the total energy consumption in the US is used by HVAC systems (DOE 2011), it is crucial to enable seamless HVAC data exchange between BIM software and BEM software. This paper researched the definition rules of HVAC systems in the current gbXML schema 6.01 version. Firstly, through a detailed data mapping of ASHRAE baseline HVAC system from IDF data model (EnergyPlus version 9.0) to current gbXML schema (version 6.01), interoperability issues were discovered and were concluded into three catecgories: missing components, overlapping components, and complex data mapping rules. Secondly, through redefining data mapping rules of current gbXML schema in terms of HVAC systems, the ASHRAE baseline HVAC system is able to be coded as a gbXML file. Finally, the gbXML file will be validated through a translation case study, focusing on primary HVAC system and boiler data. Revit 2020.1 Architecture is used in this study as the BIM tool. OpenStudio 2.8.1 and EnergyPlus 9.1.0 are used as the BEM tools. They are open-source and cross-platform, being adopted by lots of mainstream building performance analysis software. Based on the result of this study, current gbXML schema version 6.01 is capable of defining HVAC systems data, but the data mapping rules need to be documented and presented. Missing components and overlapping components issue could be solved by updating the schema. Interoperability improvement will eliminate the duplicate generation of HVAC data and allows a bidirectional information update between BIM and BEM software tools, supporting a more accurate and efficient building performance analysis process.

Full text

UC Berkeley
UC Berkeley Previously Published Works
Title
Improving the Interoperability of gbXML Data Model through Redefining Data Mapping Rules
of HVAC Systems
Permalink
https://escholarship.org/uc/item/1899g42q
Journal
ASHRAE Transactions, 126(2)
Author
Sun, Ruiji
Publication Date
2020-07-01
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California

Ruiji Sun and Zhizhang Hu are M.S. students in the Center for Building Performance and Diagnostics (CBPD) at Carnegie Mellon Universtiy, Pittsburgh, PA.
Weili Xu is the co-founder of BuildSimHub, Inc., Pittsburgh, PA. Krishnan Gowri is a senior consultant at Intertek Building Science Solutions, Inc., Seattle, WA.
Improving the Interoperability of
gbXML Data Model through
Redefining Data Mapping Rules of
HVAC Systems
Ruiji Sun Zhizhang Hu Krishnan Gowri Weili Xu, PhD
Student Member ASHRAE Fellow ASHRAE Associate Member ASHRAE
ABSTRACT
Green Building XML (gbXML) is one of the most prevalent data models used for Building Information Modeling (BIM). It enables interoperability
between BIM and Building Energy Modeling (BEM). BEM is often used in building performance analysis software tools. However, through
interviewing engineers and building energy modeling professionals in the industry, it turned out that most gbXML files are on ly used for importing and
exporting building geometry information. Information such as Heating, Ventilation, and Ait-conditioning (HVAC) systems and internal loads are
rarely handled due to the lack of functionality in current BIM software. Taking account of more than 15% of the total energy consumption in the US is
used by HVAC systems (DOE 2011), it is crucial to enable seamless HVAC data exchange between BIM software and BEM software.
This paper researched the definition rules of HVAC systems in the current gbXML schema 6.01 version. Firstly, through a detailed data mapping of
ASHRAE baseline HVAC system from IDF data model (EnergyPlus version 9.0) to current gbXML schema (version 6.01), interoperability issues
were discovered and were concluded into three catecgories: missing components, overlapping components, and complex data mapping rules. Secondly,
through redefining data mapping rules of current gbXML schema in terms of HVAC systems, the ASHRAE baseline HVAC system is able to be
coded as a gbXML file. Finally, the gbXML file will be validated through a translation case study, focusing on primary HVAC system and boiler data.
Revit 2020.1 Architecture is used in this study as the BIM tool. OpenStudio 2.8.1 and EnergyPlus 9.1.0 are used as the BEM tools. They are open-
source and cross-platform, being adopted by lots of mainstream building performance analysis software.
Based on the result of this study, current gbXML schema version 6.01 is capable of defining HVAC systems data, but the data m apping rules need to
be documented and presented. Missing components and overlapping components issue could be solved by updating the schema. Interoperability improvement
will eliminate the duplicate generation of HVAC data and allows a bidirectional information update between BIM and BEM software tools, supporting
a more accurate and efficient building performance analysis process.
INTRODUCTION
BIM-based building performance analysis
Building information model (BIM) is a digital representation of a building. It serves as a shared knowledge
resource for communication in architecture, engineering, construction (AEC) and facility management (FM) industries
(Eastman, C. et al. 2011). The first generations of data exchange models only include building geometry data,

transferring traditional 2-D drawings to 3-D object-based models. However, with the development of sustainable
architecture, material properties, HVAC equipment, and electrical product data have been considered in current BIM
data models. Green Building XML (gbXML) and Industry Foundation Class (IFC) are the two prevalent ones. (Dong
B. et al. 2007). The new BIM-based performance analysis workflow makes the analysis process less expensive and
labor-intensive and increases the accuracy of the performance simulation results. Especially during design phases, the
value of BIM and performance analysis can be maximized (Moon, H. J. et al. 2011).
Interoperability of gbXML between BIM and BEM
Building energy modeling (BEM) is a subset of BIM. It allows more information inputs for building energy
simulation such as HVAC systems data, operation schedules, envelope materials. BEM is mainly used for sustainable
architecture design, HVAC systems design and operation, and building performance rating. The interoperability of
gbXML between BIM and BEM should be seamless model translations or data exchange among disparate building
design and performance analysis software tools. However, by interviewing more than twenty energy modelers,
engineers, software developers, and other stakeholders, the current gbXML workflow only enables geometry data
transformation. HVAC systems have to be manually created in BEM tools. In addition to the interviews, eleven
current mainstream BIM and BEM software and tools are investigated in the gbXML interoperability from different
aspects, shown in Table1.
Table 1. Data exchange as gbXML in BIM and BEM Software Tools
Software Tools
Geometry
Material Properties
HVAC Systems
Internal Loads
BuildSimHub
Yes
Yes
No
No
Design Builder 6.1
Yes
Yes
No
No
IES VE 2019
Yes
Yes
Yes
Yes
OpenStuido 2.9.0
Yes
Yes
Yes
Yes
HVAC Solution 9.5.1
No
No
Yes
Yes
TRACE 700 6.3.4
Yes
Yes
No
Yes
Revit 2020.1 Architecture
Yes
Yes
No
No
Revit 2020.1 MEP
Yes
Yes
No
No
Revit 2020.1 System Analysis
Yes
Yes
Yes
Yes
Spider gbXML Viewer 0.17
Yes
Yes
No
No
Talece BIMPort
Yes
No
No
No
Same as the interview results, most BEM software tools in the industry facilitate the gbXML as a data exchange
model. However, compared with geometry, material properties, and internal loads information, HVAC systems data
was turned out to be the least handled one. Therefore, to improve the interoperability of gbXML, HVAC systems
data exchange is supposed to be enabled. A survey has reported that 78% of contractors ranked interoperability as
the top way to increase the business value of BIM (McGraw Hill, 2009). Currently, some researchers are working on
developing translator tools from gbXML directly to EnergyPlus, which is critical to the gbXML interoperability
improvement (Xu, W. et al. 2019).
gbXML as an XML-based green building schema
gbXML is an XML-based data schema representing building information. XML is an extensible markup language.
It is already a mature and powerful data model. Elements are the key objects of this language. They start with the
opening tag "<tag>" and end with the closing tag "</ tag>". The attributes included in open tags are used to
distinguish elements "<tag attribute = 'something' />". The content between the opening and closing tags may be
text, other elements, or a mixture of them (W3Schools, 2017). Due to the extensibility of the language, XML has
many derived forms. For example, based on XML data structures, Hypertext Markup Language (HTML) has

developed a set of predefined tags and attributes to help display content on a website. Similarly, gbXML has the same
XML data structure. It also has predefined tags and attributes that are specifically designed for green building
information modeling (gbXML, 2017). Therefore, only XML files that follow the gbXML schema can be called
gbXML. However, confusion occurs when the BIM model follows part of the gbXML schema and customizes some
other elements and attributes. For example, in the Revit 2020.1 system analysis, the exported file follows the gbXML
schema in the geometric data, but for HVAC system data, it adds two custom elements. For clarity, the customized
gbXML is referred to as an XML file in this article.
METHODOLOGY
The objective of this research is improving the interoperability of gbXML. The study scope is limited to the
ASHRAE baseline HVAC system seven, which is a Variable Air Volume (VAV) and reheat fan system. The validation
method is a translation case study. Firstly, a research of current gbXML data schema was conducted through mapping
the baseline HVAC system data one by one from the IDF data model to the gbXML schema. Three issues were
found and analyzed. Secondly, based on the research result, the gbXML data mapping rules in terms of HVAC
systems were redefined. A gbXML example file was created and was parsed to OpenStudio and EnergyPlus
environment. Boiler data was selected as a typical HVAC equipment data. Finally, the translation result suggested the
effectiveness of the redefined data mapping rules. The workflow of this study is shown in Figure 1.
Figure 1. gbXML interoperability improvement workflow
DATA MAPPING
HVAC systems data mapping from IDF to gbXML
Compared with the XML, the IDF data structure is flattened. There are only three levels in the IDF model:
“class”, “field”, and “objective”. The "field" is used to define the characteristics of a "class". The "objective" is for
different objects, belonging to the same class. In the IDF data model of the ASHRAE baseline system type seven,
HVAC related classes are “PlantLoop”, “CondenserLoop”, “ZoneHVAC: AirDistributionUnit”, “Boiler: HotWater”,
etc. These classes have predefined IDF fields and each respective object has its customized value, following
corresponding fields. Take the "Boiler: HotWater" class as an example to explain the data mapping process from IDF
to gbXML in detail. IDF representation of the “Boiler: HotWater” is shown in Table 2.

Table 2. IDF Representation of a Boiler
Class
Field
Units
Object1
Boiler: HotWater
Name
-
Boiler1
Boiler: HotWater
Fuel Type
-
NaturalGas
Boiler: HotWater
Nominal Capacity
W
autosize
Boiler: HotWater
Nominal Thermal Efficiency
-
0.8
Boiler: HotWater
Efficiency Curve Temperature Evaluation Variable
-
LeavingBoiler
Boiler: HotWater
Normalized Boiler Efficiency Curve Name
-
Boiler Efficiency Curve
Boiler: HotWater
Design Water Flow Rate
m3/s
autosize
Boiler: HotWater
Minimum Part Load Ratio
-
0
Boiler: HotWater
Maximum Part Load Ratio
-
1.1
Boiler: HotWater
Optimum Part Load Ratio
-
1
Boiler: HotWater
Boiler Water Inlet Node Name
-
Boiler1 HW Inlet
Boiler: HotWater
Boiler Water Outlet Node Name
-
Boiler1 HW Outlet
Boiler: HotWater
Water Outlet Upper Temperature Limit
C
100
Boiler: HotWater
Boiler Flow Mode
-
LeavingSetpointModulated
Boiler: HotWater
Parasitic Electric Load
W
0
Boiler: HotWater
Sizing Factor
-
1.25
Boiler: HotWater
End-Use Subcategory
-
-
All fields of the “Boiler: HotWater” class in IDF are mapped to elements and attributes in gbXML data schema.
To define the boiler as a “HydronicloopEquipemt”, its parent element “HydronicLoop” needs to be defined. In Table
3, gbXML defines HVAC systems in a hierarchy way. “AirLoop” and “HydronicLoop” are two main elements for
primary system data. “AirLoopEquipemt” and “HydronicloopEquipemt” are used to define the secondary system
data or equipment data. “Zone” and “Space” are used to assign HVAC systems to specific areas.
Table 3. gbXML Representation of a Boiler
1st Element
2nd Element
3rd Element
4th Element
Attributes
Object1
gbXML
HydronicLoop
-
-
loopType
HotWater
gbXML
HydronicLoop
-
-
fluidType
Water
gbXML
HydronicLoop
-
-
id
DHW1
gbXML
HydronicLoop
HydronicLoopEquipment
-
id
DHW1-B1
gbXML
HydronicLoop
HydronicLoopEquipment
-
equipmentType
Boiler
gbXML
HydronicLoop
HydronicLoopEquipment
Name
-
Boiler1
gbXML
HydronicLoop
HydronicLoopEquipment
Temp
unit
C
gbXML
HydronicLoop
HydronicLoopEquipment
Temp
tempType
Max
gbXML
HydronicLoop
HydronicLoopEquipment
Power
unit
Watt
gbXML
HydronicLoop
HydronicLoopEquipment
Power
useType
Heating
gbXML
HydronicLoop
HydronicLoopEquipment
Power
powerType
NaturalGas
gbXML
HydronicLoop
HydronicLoopEquipment
Efficiency
efficiencyType
BoilerEff
gbXML
HydronicLoop
HydronicLoopEquipment
Control
controlType
Boiler
gbXML
HydronicLoop
HydronicLoopEquipment
Control
minPowerRatio
0
gbXML
HydronicLoop
FlowControl
DesignFlow
unit
LPerSec
Issue one: missing components
During the data mapping process from the IDF data model to the gbXML schema. It turned out that some IDF
fields do not have corresponding elements or attributes in gbXML. Take the “Boiler: HotWater” IDF class as an
example. In the IDF data model, the “Maximum Part Load Ratio”, “Optimum Part Load Ratio”, “Boiler Flow Mode”,
and “Sizing Factor” are clearly defined and they critical to boiler performance as well as building energy simulation.
But in current gbXML schema, they can not be defined using existing elements or attributes, unless custom elements
or attributes are created. Some other missing components are unnecessary for gbXML data exchange due to the

hierarchy structure, such as “Boiler Water Inlet Node Name”, “Boiler Water Outlet Node Name” and “End-Use
Subcategory”. They can be defined by order of “HydronicLoopEquipment” elements.
Issue two: overlapping components at loop intersections
HVAC engineering consists of various equipment and system loops. The equipment at the intersections between
two different loops could raise inconsistent issues. For example, “Coil” can be defined as an “equipmentType”
attribute of the “AirLoopEquipment” element, but it can also be defined as an “equipmentType” attribute of the
“HydronicLoopEquipment” element. Therefore, if two software map the “Coil” differently, then the data will not be
decoded or encoded coincidently, thus incurring interoperability issues. The same situations also occur at intersections
between the supply air loop and return air loop, primary and secondary chilling water loops, etc. Figure 2 shows some
typical interactions in an HVAC system.
Figure 2. Hydronic loops and air loops
Issue three: complex data mapping rules
In IDF, HVAC systems data are categorized into different classes. But in the gbXML schema, data is not
independent of each other. Most elements have complicated reference relationships, and the gbXML documentation
doesn’t explain this well. For example, a zone is made up of one space or multiple spaces and is served by one HVAC
system (McDowall, R. 2007). In the IDF, the “ZoneHVAC: AirDistributionUnit” class is named by zones. But in
gbXML schema, HVAC systems and equipment are identified by “id” attributes. A primary system connects with a
zone, the “AirLoop” and “HydronicLoop” elements have a “controlZoneIdRef” attribute. At the same time, the
“Zone” element also has “AirLoopId” and “HydronicLoopId” children elements. However, for HVAC equipment,
the “AirLoopEquipement” and “HydronicLoopEquipment” elements don’t have a zone or space reference. Instead,
“Space” elements in gbXML are connected with HVAC equipment and they have “AirLoopEquipementId” and
“HydronicLoopEquipmentId” children elements. The complex relationships are shown in Figure 3. They would cause
missed connections between zones or spaces with HVAC equipment, raising interoperability issues.
REDEFINING DATA MAPPING RULES
Compared with missing components, decoding performance curves, the problem of complex data mapping rules
is the most critical issue and needs to be redefined. The rule of thumb for defining HVAC systems in gbXML schema
is bottom-up. First of all, defining “HydronicLoopEquipment”, “AirLoopEquipment”, “AirLoop” and
“HydronicLoop” in order. Secondly, connecting spaces and HVAC equipment by “HydronicLoopEquipmentId” and
“AirLoopEquipmentId” in “Space” element; connecting zones and primary system by “HydronicLoopid” and

“AirLoopId” in “Zone” element. Additionally, connecting spaces and zones by “zoneIdRef” in the “Space” element.
In this way, zones are connected with their spaces’ equipment indirectly. For example, there are three spaces in a zone,
but only one of them has a VAV box. Then this zone is connected with the VAV box. Figure 4 shows redefined data
mapping rules, which are more effective and efficient than the relationships shown in Figure 3.
Figure 3. Current gbXML data mapping rules of HVAC systems
Figure 4. Redefined gbXML data mapping rules of HVAC systems

VALIDATION
gbXML file preparation
A corresponding building model was created in Revit 2020.1, and zones were assigned by the core-perimeter
method. Therefore, building geometry data can be exported as a gbXML file from Revit. Then, HVAC systems data is
manually added to this gbXML file by the redefined data mapping rules, following current gbXML schema. Though
not all HVAC equipment is covered, a primary HVAC system and boiler data is sufficient enough to prove the
effectiveness of the redefined rules. Figure 5 shows boiler data in the gbXML file. “Sizing Factor” and other critical
parameters were not handled due to the missing components issue of the current schema. Instead of adding a new
customized XML tag, “Sizing Factor” value will be added in BEM tools to be translated to IDF.
Figure 5. Boiler information added in gbXML
OpenStudio manipulation
The gbXML file can be imported into OpenStudio by applying customized measures. It becomes an OSM file.
Figure 6 shows an example Ruby code that converts the boiler data in gbXML to OpenStudio. OpenStudio uses a
built-in “ForwardTranslator” to translate the OSM model to an IDF model. Information such as the climate,
schedules, internal loads, etc. is added in the IDF model by a python script and is kept the same as the original IDF
model of the ASHRAE baseline HVAC system.
Figure 6. Ruby code for converting boiler data in gbXML
Translation results
Successful import of the primary HVAC system and the boiler data suggested the effectiveness of the redefined
data mapping rules of gbXML for HVAC systems. Figure 7 shows a graphical layout of the converted primary HVAC
system in OpenStudio, and Figure 8 shows the converted boiler data in EnergyPlus IDF editor. Both HVAC system-
level information and equipment-level information are the same as the information in the original IDF model of the
ASHRAE baseline HVAC system type seven. Due to time constraints, the energy simulation of the converted IDF
model cannot be performed until the HVAC system is fully mapped to the current gbXML schema. It should be the
same as the original IDF model of the ASHRAE HVAC system.

Figure 7. Converted primary HVAC system in OpenStudio 2.8.1
Figure 8. Converted boiler data in EnergyPlus 9.1.0 IDF editor
CONCLUSION
The current gbXML schema (Version 6.01) is rarely used for HVAC systems representation due to complex data
mapping rules. It has been one of the most critical obstacles to gbXML interoperability improvement. Through
detailed data mapping of the ASHRAE standard system type seven from the IDF model (Version 9.0) to current
gbXML schema, gbXML should be able to define HVAC systems though three interoperability issues were figured
out. The complex data mapping rules have been redefined, using four elements: “Space”, “Zone”, “AirLoop” and
“HydronicLoop”. The “AirLoop” and “HydronicLoop” elements are used for primary HVAC system data. Their
children elements “AirLoopEquipment” and “HydronicLoopEquipment” are used for HVAC equipment data.
“Space” can be connected to the HVAC equipment by its children elements “HydronicLoopEquipmentId” and
“AirLoopEquipmentId”. “Zone” can be connected to the HVAC primary system by its children elements
“HydronicLoopId” and “AirLoopId”. In terms of the other two interoperability problems: missing components and
overlapping components. They could be handled by updating the gbXML schema.
ACKNOWLEDGEMENT
This study was funded through the research project 1810 of the American Society of Heating, Refrigerating, and
Air-Conditioning Engineers (ASHRAE). The authors acknowledge the support of employees at BuildSimHub. Inc.
and faculties at Carnegie Mellon University for this research. Finally, Sun R. wants to thank, in particular, the
invaluable support received from Zhang C. over the months.

REFERENCES
DOE, U. (2011). 2011 Building energy data book. US Department of Energy.
Dong, B., Lam, K. P., Huang, Y. C., & Dobbs, G. M. 2007. A comparative study of the IFC and gbXML informational
infrastructures for data exchange in computational design support environments. In Tenth International IBPSA
Conference (pp. 1530-1537).
Eastman, C., Teicholz, P., Sacks, R., & Liston, K. 2011. BIM handbook: A guide to building information modeling for owners,
managers, designers, engineers and contractors. Chapter. John Wiley & Sons.
gbXML Current Schema 2017. http://www.gbxml.org/Schema_Current_GreenBuildingXML_gbXML
McDowall, R. 2007. Fundamentals of HVAC systems, SI edition. Chapter 5, Zones. Academic Press.
Seyam, Shaimaa. 2018. Types of HVAC Systems. In HVAC System. IntechOpen.
McGraw Hill. 2009. The business value of BIM: Getting building information modeling to the bottom line. Smart Market
Report (2009): 1-50.
Moon, H. J., Choi, M. S., Kim, S. K., & Ryu, S. H. 2011. Case studies for the evaluation of interoperability between a BIM
based architectural model and building performance analysis programs. In Proceedings of 12th conference of international
building performance simulation association (Vol. 2011).
W3Schools Online Web Tutorials. 2017. https://www.w3schools.com/
Xu, W., Chong, A. & Lam, K. P. 2019. A new BIM to BEM framework: the development and verification of an open-
Source gbXML to EnergyPlus translator for supporting building life cycle performance analysis. In building simulation
2019 conference.
Zhang, Z., Chong, A., Pan, Y., Zhang, C., Lu, S., & Lam, K. P. (2018, September). A deep reinforcement learning approach
to using whole building energy model for hvac optimal control. In 2018 Building Performance Analysis Conference and
SimBuild.
View publication stats