Energy and Buildings 81 (2014) 391–403

Contents lists available at ScienceDirect

Energy and Buildings

journal homepage: www.elsevier.com/locate/enbuild

Building Information Modeling (BIM)-based daylighting simulation

and analysis
Sandeep Kota, Jeff S. Haberl, Mark J. Clayton, Wei Yan ∗
Department of Architecture, Texas A&M University, College Station, TX, United States

a r t i c l e i n f o a b s t r a c t

Article history: Daylighting is an important aspect in designing high performance buildings. Many simulation tools have
Received 30 April 2014 been developed to study the daylighting performance of buildings. These tools primarily use CAD envi-
Received in revised form 17 June 2014 ronments for creating architectural models, which are then converted into daylighting models to run on
Accepted 25 June 2014
the daylighting simulation engines. Once the architect defines the architectural model in CAD, a sim-
Available online 3 July 2014
ulation expert creates the simulation input file to perform daylighting analysis. Each tool has its own
rules that the architect and the engineer have to follow to prepare the simulation input files, and the
Keywords:
complexity depends on the tools. Currently, Building Information Modeling (BIM) is widely used in the
Building Information Modeling
BIM
AECO industries and BIM models are used as a means of exchanging data among different professionals
Daylighting simulation involved in the design and construction of buildings. The present paper discusses the use of BIM for build-
Radiance ing performance simulations and mainly focuses on how daylighting analysis can be incorporated into
DAYSIM a BIM environment, and what challenges and benefits exist in the process of integrating BIM with day-
lighting simulation tools. The paper presents the development and validation of a prototype to integrate
the BIM tool, Revit with the daylighting simulation tools, Radiance and DAYSIM.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction [2] by many design firms in their design process. Research in the
integration of BIM with building performance tools has been a main
Building performance analyses are important aspects of design- focus of both the developers of the BIM authoring tools as well
ing sustainable buildings. One of the performance analyses done as the Building Performance Simulation community. The following
by architects is to predict how buildings are performing in terms of sections present a methodology of integrating BIM authoring tools
their luminous environment as a result of daylighting. Daylight- (Autodesk Revit as a sample BIM tool) with daylighting analysis
ing analysis is done using either hand calculations or computer tools (Radiance and DAYSIM as sample simulation tools, which are
simulation tools. For simulation, once the architect defines the widely used to study the daylighting performance of buildings).
architectural model of a building, the simulation expert prepares a
simulation model for performing the analysis. The preparation of
simulation models (i.e. input files) can be a very lengthy, laborious, 2. Literature review
and resource consuming process as the work mostly consists of
manual or semi-manual translation from architectural model data 2.1. Daylighting simulation tools
to simulation data, which often results in numerous coding errors
[1]. To ease the process of creating the input files, graphical user Over the years many analysis tools have been developed to
interfaces for defining model geometry have been created for sim- study the daylighting performance of buildings. These tools extend
ulation tools. In addition, geometry modeling tools (CAD tools) have from simple charts to sophisticated computer tools. A brief his-
been linked with daylighting simulation tools. Currently Building torical development of these tools can be found in literature [3].
Information Modeling (BIM) is being used in building design in the Among these tools, Radiance is considered to be a state-of-the-art
Architecture/Engineering/Construction/Operation (AECO) industry backward ray tracer, which is based on a mixed stochastic and
deterministic ray tracing approach [4,5]. Simulation results with
Radiance have been physically validated for a range of building
geometries and shading devices. Radiance has been studied sys-
∗ Corresponding author. Tel.: +1 979 8450584. tematically over the last decade [6–8]. Studies have shown that
E-mail address: [email protected] (W. Yan). Radiance can be combined with a Daylight Coefficient (DC) [9]

http://dx.doi.org/10.1016/j.enbuild.2014.06.043
0378-7788/© 2014 Elsevier B.V. All rights reserved.
392 S. Kota et al. / Energy and Buildings 81 (2014) 391–403

approach to reliably model the changing levels of daylight in a Rhinoceros and export the models to Radiance and DAYSIM to per-
building over the course of a year. DAYSIM [8], which is an annual form the simulation.
DC-based daylighting simulation tool with Radiance as its engine,
has also been validated and is widely used in research and practice. 2.2.4. Ecotect-Radiance-DAYSIM
Another daylighting simulation tool is Autodesk 3ds Max [10]. 3ds Ecotect is a widely used tool in the building simulation com-
Max uses mental ray for doing daylighting analysis. It offers both a munity for doing daylighting simulation [19]. Ecotect provides a
snapshot of daylight levels at a specified time point as well as annual CAD-based modeling environment for creating building geometry,
illumination levels at the user specified locations in the building assigning materials to geometry, and exporting the model to Radi-
[11]. 3ds Max has been validated against Radiance and DAYSIM ance [20] and DAYSIM [21]. Fig. 1 shows linking of CAD-based tools
and proven to be accurate in predicting annual indoor illumina- with Radiance and DAYSIM.
tion levels. Although there are several daylighting simulation tools
2.2.5. Advantages and limitations of CAD-based daylighting
available, the authors have chosen to use Radiance and DAYSIM for
simulation tools
the present project as they are accurate, validated, and widely used
CAD-based daylighting simulation tools help in creating the
simulation tools [12].
geometry quickly. Without the aid of CAD interfaces it would be
very time consuming to define geometry and material proper-
2.2. CAD to daylighting
ties manually for daylighting simulations. However, some of the
CAD-based tools have certain limitations in terms of creating geom-
Many of the daylighting simulation tools require defining the
etry. These limitations are generally to require an idiosyncratic
geometry of a building in a three dimensional (3D) coordinate sys-
or specific method of defining geometry. For example in Desktop
tem. Several of the sophisticated tools use text-based input files
Radiance to create a window in a wall, the window glass has to
for modeling the geometry and the materials of the building. The
be modeled as a single surface and the wall surface hosting the
syntax of a simulation input file is specific to the tool. For example,
window has to be divided into four different surfaces. Similarly, in
Radiance and DAYSIM input files are human readable text files that
DIVA for Rhinoceros, windows have to be modeled as surfaces with
can also be interpreted by the computer programs. However, man-
zero thickness. The user has to adhere to these rules for creating
ually preparing an input file is a time-consuming process and often
daylighting simulation models. Also, if the building geometry has
vulnerable to input mistakes. In order to make input file preparation
to be represented with thickness but the daylighting surfaces are
more convenient and easy, several CAD modeling tools have been
modeled with zero thickness, then the complexity and time/effort
linked with daylighting simulation tools. Generally the CAD tools
for defining the model increase with the increase in the number
assist in the preparation of the geometry and material properties,
of surfaces considerably. Furthermore, the materials have to be
and subsequently export the model into text-based, tool-specific
assigned by selecting individual surfaces, which is also a tedious
input files to do simulation. Some of these tools are discussed
task if the model is complex.
below.
2.3. Building Information Modeling to daylighting
2.2.1. Advanced Daylight and Electric Lighting Integrated New
Environment (ADELINE 2.0) Building Information Modeling (BIM) is an emerging technol-
ADELINE is an integrated lighting design computer tool for doing ogy in the AECO industry [22]. According to a public report there
daylighting and electric lighting simulation [13]. ADELINE comes is a considerable increase in the use of BIM in AECO [2]. BIM tools
with SCRIBE-Modeler as a CAD interface and it connects to SUPER- not only help in designing a building using 3D graphics, but also
LITE [14] and Radiance to do daylighting and lighting simulation. It are capable of holding non-graphical information such as material
is also linked with energy simulation tools using a special program properties related to the building elements, which are not available
called SUPERLINK. Once the geometry is created with SCRIBE- in CAD tools. For example a wall can be modeled as an object with
Modeler, it is assigned materials using a program called PLINK, layers of materials using different properties. In addition, BIM tools
which also helps to define climate data for doing daylighting anal- help in modeling the geometry very quickly with a high degree of
ysis. accuracy and detail compared to CAD tools by enforcing topology
and connectivity among elements such as walls, windows, doors,
2.2.2. Desktop Radiance roofs, and floors. Also, it is very easy to parametrically change the
Desktop Radiance is a CAD-based daylighting and electric light- dimensions of the BIM objects and spaces with the geometry updat-
ing simulation tool with AutoCAD as a geometry modeler and ing automatically based on any change. BIM is not only used to
Radiance as a simulation engine [15]. For linking AutoCAD with create building geometry, but also as a repository of the building
Radiance a special program called RADOUT was created [16]. It runs information, which can be retrieved to perform different analy-
in the AutoCAD environment and facilitates the translation of CAD ses on buildings, such as energy analysis, daylighting analysis, cost
geometry into Radiance input files. Apart from having a CAD inter- estimation, and structural analysis.
face for creating geometry, Desktop Radiance also has a Radiance
material library for assigning materials. RADOUT provides an inter- 2.3.1. From BIM to CAD-based daylighting (Autodesk Revit to 3ds
face to selectively export the geometry. Once the Radiance input file Max Design)
is created, different Radiance utilities are called using MS Windows 3ds Max Design is a product of Autodesk that is used mostly
BATCH scripting for doing a daylighting analysis [15]. In addition to for animation and rendering. A module called Exposure in 3ds Max
AutoCAD, Radiance also has many utility programs that help in con- Design enables the user to do a physically based accurate daylight-
verting different geometry formats used by CAD-based tools into ing simulation. The simulation engine is mental ray. Both time series
the Radiance format [16]. rendering of High Dynamic Range (HDR) images and climate-based
annual daylighting simulation can be conducted using Exposure. It
2.2.3. DAYSIM/DIVA has been validated comparing with measured data as well as with
Another CAD-based modeling tool is Rhinoceros, a NURBS mod- Radiance and DAYSIM daylighting simulating tools for a number of
eling program [17]. It has been linked to Radiance and DAYSIM cases that extend from a simple window to complex fenestration
through a plug-in called DIVA [18]. DIVA provides a graphical systems consisting of shades, light shelves, and indoor venetian
user interface to assign materials to different surfaces created in blind systems [10].
 S. Kota et al. / Energy and Buildings 81 (2014) 391–403 393

Fig. 1. CAD to daylighting workflow.

The interoperability between the BIM tool, Revit, and the CAD manual intervention. The main objective of the research project is
tool, 3ds Max Design, is achieved through a file format called FBX to make the process quick and friendly to architects and engineers.
[11]. The FBX file format contains the geometry, material infor- The following sections will describe the whole process in detail.
mation, and camera views necessary for the daylighting analysis
in 3ds Max Design. However, in order to do a physically accurate 2.3.3. Revit to Radiance and DAYSIM
daylighting simulation, a series of manipulations has to be per- The present research is a part of a project on Physical Building
formed initially such as replacing the lights in the Revit model with Information Modeling (PBIM). The aim of the project is to integrate
a daylighting system in 3ds MAX Design, and replacing the non- BIM with building performance analysis tools in multiple domains,
physically based materials of Revit with physically based materials. including thermal and daylighting [25]. Even though there are sev-
Along with these changes, some of the changes have to be done to eral BIM authoring tools available in the market today, Autodesk
the Revit model, especially when modeling window glass panes. Revit has been adopted for the present project because of its para-
A detailed workflow of doing daylighting analysis using 3ds Max metric BIM authoring capability and its wide acceptance in the
Design from Revit can be found in the literature [23]. Even though industry. The main aim of the project is to link BIM with Radiance
the process is fast compared with CAD-based daylighting simula- and DAYSIM by translating the Revit models into Radiance/DAYSIM
tion, it still requires manual checking and preparing the scene for input files and automatically performing the simulation. In a prior
accurate daylighting simulation. The process can be described as study [26] a Revit model is translated into a Radiance input file
semi-automated. through other file formats and tools. Some of the major findings
reported by the study are (1) translating Revit data into Radiance
2.3.2. ThermalOpt daylighting simulation requires several steps to convert Revit geometry into Radiance
ThermalOpt is a methodology for automated BIM-based sim- geometry, (2) manual intervention is required in the whole pro-
ulation intended for use in multidisciplinary design optimization cess, and (3) certain information is required to be added manually
(MDO) environments [24]. It can do both thermal and daylight- to create a complete Radiance input file. In our project, we aim
ing simulations, supporting study of the effect of daylighting on to develop a direct translator from Revit to Radiance and DAYSIM
the thermal performance. It uses EnergyPlus for thermal sim- in order to significantly reduce the model preparation time and,
ulation and Radiance for daylighting simulation. ThermalOpt is in the meantime, preserve the parametric modeling advantages of
implemented via four different plugins: (1) IFC2ThermalSim, (2) Revit.
ThermalSim, (3) EnergyPlus Wrapper, and (4) Radiance Wrap-
per. In ThermalOpt, the building description is represented in IFC 2.4. Radiance utility programs
BIM interoperability format and translated into EnergyPlus input
format. Geometry for daylighting simulation is taken from the Ener- Several methods were explored to convert a Revit model to
gyPlus model. The EnergyPlus model does not represent the actual a Radiance input file. Radiance has several utility programs for
geometry of BIM. Taking geometry from EnergyPlus does not give converting a wide variety of CAD geometry data formats into the
accurate geometry for daylighting simulation. Even though the Radiance format. Some of the widely recognized data formats,
whole process is automated, the geometry for the daylighting simu- which are used by different CAD-based tools, are 3DS [27] and DXF
lation is not the same as the BIM model and this can alter the results. [28] developed by Autodesk, OBJ [27] developed by Wavefront and
Learning from the shortcomings of the above studies, it was the MGF developed particularly for Radiance [29].
thought necessary to develop a process and methodology to Autodesk offers a comprehensive exporting utility called FBX
accurately translate the BIM models’ geometry and material infor- Converter that can convert the Revit file into the FBX format [30]
mation into daylighting models (Radiance and DAYSIM). This was and further into different geometry formats such as DWG, DXF,
achieved by developing custom add-in programs that can take Revit OBJ and 3DS. For the present project, to explore existing methods
BIM models and translate them into Radiance and DAYSIM input for converting BIM to daylighting simulation models, Revit mod-
files, and then automatically perform the simulation without any els were exported into the FBX format and then subsequently into
394 S. Kota et al. / Energy and Buildings 81 (2014) 391–403

Fig. 2. Different translation paths from Revit to Radiance: P1, P2, P3, P4 and P5 are the paths of existing methods, and P6 is the path of our proposed new method.

DXF, OBJ and 3DS using the FBX Converter. DXF2RAD [31], OBJ2RAD tools, which are needed in other paths. Table 1 shows the number
[32], 3DS2MGF [33], and MGF2RAD [34] utility programs are used of steps involved in each method for converting a Revit model into
to convert these formats into the Radiance format. Fig. 2 shows Radiance/DAYSIM input files.
different steps involved in converting a Revit model into a Radi-
ance input file using these different utilities. The steps involved in 2.5. Issues involved in Revit to daylighting using Radiance utilities
translation using different utilities are represented as paths (i.e. P1,
P2, P3,. . .). P6 is the translation path describing our method pre- In all the existing approaches, geometry information could be
sented in this paper. The method proposed is a more direct path exported automatically using the utilities but not material infor-
involving only one step for creating the Radiance input file and also mation. Manual assignment is needed for materials. This is a major
performing the simulation. The main advantage of this method is setback for creating Radiance input files. Another major setback
that it is easy and does not require in-depth knowledge of different found in the existing methods is that in Revit the window panes

Table 1
Steps involved in each path for converting a Revit model into Radiance/DAYSIM input files.

Path Name Revit file Importing to other Assigning materials Exporting/Converting Exporting/Converting No. of steps
exporting/translating programs or Exporting to
other formats

P1 Exported into DWG DWG format Materials assigned Exported into Exported into 5
format imported into 3DS in 3ds Max Wavefront OBJ Radiance format
Max format using OBJ2RAD
utility
P2 Exported to FBX format Imported into Exported into 3DS Converted into Converted into 5
Autodesk FBX format *.3ds MGF format using Radiance format
converter utility 3DS2MGF using MGF2RAD
utility
P3 Exported to FBX format Imported into Exported into Converted into 4
Autodesk FBX Wavefront OBJ Radiance format
converter utility format *.OBJ using OBJ2RAD
utility
P4 Exported to FBX format Imported into Exported into DXF Converted into 4
Autodesk FBX format *.DXF Radiance format
converter utility using DXF2RAD
utility
P5 Exported to FBX format Imported into Materials assigned Simulated in 3ds 4
Autodesk 3ds Max in 3ds Max Max
P6 Translated into 1
Radiance input file
using our
Revit2Radiance Add-in
Program
 S. Kota et al. / Energy and Buildings 81 (2014) 391–403 395

are modeled with thickness whereas in Radiance window panes Table 2

Revit add-in programs (Revit2Radiance modules) for translating Revit models into
are modeled with zero thickness – this means manual adjustment is
Radiance input files.
needed for modeling the windows with the existing methods. From
our experiments, some errors were identified when converting OBJ Name of the add-in Action accomplished by the
program add-in
data into Radiance using OBJ2RAD utility. Manual inspection of the
OBJ geometry file created by FBX Converter revealed some strange 1 Parameter File Creates a default shared
characters in the OBJ input file that caused errors in translating Creation parameter file in the project
directory for adding shared
using the OBJ2RAD utility. We also observed that converting OBJ
parameters that are needed for
and DXF to RAD was a one-step process, but converting 3DS to RAD defining building materials in
was a two-step process. During the process of converting 3DS to Add-in #2 below
MGF using the 3DS2MGF utility, an error occurred resulting in the 2 Adding Material Adds user-defined shared
Parameter parameters for building
MGF file not being created. From our experiments, the only path
materials that are necessary for
that has successfully translated geometry from a Revit model to holding materials data for
Radiance (without material definition) is DXF2RAD. Each of these Radiance models
methods has its own limitations. There are many steps involved 3 Reading Material Reads values of the created
in the process of translating the Revit model into the Radiance Information shared material parameters
4 Model Translator Translates the Revit model’s
input file. In addition, a thorough knowledge about each tool is
geometry and material data
required to carry out the process, which can be a significant chal- into Radiance/DAYSIM input
lenge for even an experienced architect. As the architect changes files and automatically run the
the design iteratively the whole process becomes a tedious task, as simulation programs to
generate the results
each translation involves working with several tools and multiple
steps. Furthermore, the architect cannot study design alternatives
quickly due to the laborious process involved in each method.
multiple parameters and values assigned to it. For example, one of
Owing to these reasons, we proposed to translate the Revit models
the material parameters is “color” which gives the color property
into Radiance input files directly and automatically (i.e., integrat-
of the material in terms of the Red, Green, and Blue (RGB) chan-
ing the daylighting analysis into the Revit modeling environment),
nels. In Radiance there are different material types [16]. One of the
which does not involve human intervention and results in a process
basic material types is ‘Plastic’ which has a color component, dif-
that is quick and user friendly.
fuse and specular components, and a roughness factor. The color
is defined by Red, Green and Blue reflectance values [16]. Given
3. Methodology – prototyping the color values in RGB of a surface it is possible to compute the
reflectance of the material through a brightness formula. The Aver-
Our research method includes prototyping and validating the age Reflectance value is computed from the following formula for
prototype with building models. For facilitating the data transfer the standard Radiance R, G, and B components of the reflectance
from Revit to Radiance and DAYSIM, a prototype Revit2Radiance [16]:
was developed. The prototype takes the building information, both
geometric and non-geometric such as material properties, from a Average Reflectance = 0.265 × R + 0.670 × G + 0.065 × B (1)
BIM model and translates it into Radiance and DAYSIM input files. where R, G, and B are normalized values. In our prototype, we
The prototype was developed using Revit’s Application Program- used the default Revit material R, G, and B values (in the range of
ing Interface (API) with the C# programming language. Based on 0–255), normalized them, and calculated the Average Reflectance
BIM (geometry, materials, location, date/time, a camera view, and using the above equation. The same Average Reflectance is used as
sensor points), the prototype extracts the building surfaces in tri- the Radiance RGB values for calculating the illumination levels and
angle meshes and other data needed to generate the input files rendering grayscale images by Radiance. The rendered grayscale
of Radiance and DAYSIM. The main objective for developing the images can help visualize the illumination levels without the effects
prototype is to translate the Revit models into Radiance/DAYSIM of chromaticity. For rendering color images, Eq. (1) does not need to
input files and perform the daylighting simulation and visualiza- be called by Revit2Radiance and the normalized Revit RGB, instead
tion without manual intervention in the translation process (after of the Average Reflectance value, will be used as Radiance RGB.
the BIM model is created and data are entered into BIM). This In addition, if the users intend to use measured RGB of a material
not only helps avoid errors due to manual input of the data in instead of the default values in Revit, they can modify the values in
preparing simulation input files, but also makes the process of per- Revit.
forming the analysis more user-friendly for the designers who are However, Revit materials do not have specularity and roughness
not well acquainted with the simulation tools. For the present pro- values that are necessary for complete Radiance material descrip-
totype a total of four Revit add-in programs were developed that tion. It is necessary that these values be first added to each material
can automatically translate BIM models into simulation models, in the Revit material database. Revit offers the functionality for
simulate the building performance, and generate analysis results. adding user-defined parameters called custom parameters. In the
Table 2 shows the add-in programs and respective tasks they present project this functionality was used through API program-
accomplish. Fig. 3 shows the overall workflow of Revit2Radiance. ming to add specularity and roughness parameters to each Revit
Fig. 4 shows the workflow with the add-in programs. In the sub- material. In order to add user-defined parameters, a shared param-
sequent sections each of the add-in programs is explained in eter file must be created first for the parameters. The first add-in
detail. program creates this shared parameter file.

3.1. Parameter File Creation 3.2. Adding Material Parameters

A Revit model has building geometry and material information Once the default shared parameter file is created, the sec-
associated with the geometry. Materials are assigned to geome- ond add-in program “Adding Material Parameter” adds the two
try through the Revit material database. Each material in Revit has custom parameters “Specularity” and “Roughness” to the model,
396 S. Kota et al. / Energy and Buildings 81 (2014) 391–403

Fig. 3. Revit2Radiance workflow.

which then appear in the custom parameter list of each material in containing Revit material names with their corresponding spec-
Revit. ularity and roughness values taken from Radiance materials. The
“Reading Material Information” add-in program reads the Excel
3.3. Reading Material Information database file containing the specularity and roughness values and
assigns these values to the custom parameters created for each
As described briefly in the earlier section “Parameter File material in the earlier steps. At this point each material in the Revit
Creation”, Radiance materials have reflectance, specularity and material database has all the necessary information that must be
roughness values. Reflectance is defined with individual channel translated into a Radiance material description.
reflectance (i.e. red-reflectance, blue-reflectance and green-
reflectance) along with specularity and roughness. Specularity and
roughness values must be added to the parameters created for 3.4. Radiance sky model and view description
each Revit material (i.e. adding appropriate values to the shared
parameters for each material). In order to ascertain specularity Radiance requires a sky, a view and a scene description. There
and roughness values of Revit materials, the Revit materials were are different ways a sky description can be modeled in Radiance.
compared with similar material types for which Radiance mate- By providing the latitude, longitude, and time of the day, the Radi-
rial data are available from different databases. The specularity and ance utility Gensky [16] can create a sky description. In Revit,
roughness values of these Radiance materials were then adopted each view has sun setting parameters, which also provide access
for Revit materials. A list was created in an Excel spreadsheet to location (latitude and longitude) information and the time of

Fig. 4. The overall process of translating a Revit model into Radiance/DAYSIM input files.
 S. Kota et al. / Energy and Buildings 81 (2014) 391–403 397

the day (month, day, and hour) for shadow calculating and image Table 3
Project and material description.
rendering purposes. For the present prototype, this information is
accessed, formatted, and provided to Gensky program to create the Building parameters Values
sky description by our prototype. Similarly, the camera view infor- Location Denver, CO
mation is translated from BIM into a Radiance view description that Length of the room 6M
is necessary to produce Radiance images that can be post-processed Width of the room 8M
for different analyses. No. of windows 2
Dimensions of windows 3M×2M
Wall reflectance 0.5
3.5. Sensor point description Floor reflectance 0.2
Roof/ceiling reflectance 0.8
In Radiance the simulation results can be displayed as images, Window transmittance 0.75
numerical values, and illumination contour plots. Apart from pro-
ducing a picture that has the photometric information for a given
camera view and scene description, Radiance can also compute which can have multiple surfaces depending on the shape of the
lighting levels at a specific location inside a given space by spec- geometry (Fig. 5A shows a solid of a window pane). The Model
ifying a sensor point. The sensor point is specified by giving the Translator add-in converts each surface of a solid into a Radiance
X, Y and Z coordinates and view directions at which Radiance cal- geometry surface with material information attached to it. In Revit,
culates the illumination levels. For the prototype, a sensor object window panes are mostly cuboids of 6 surfaces with two of the sur-
was created as a custom Revit family that the user can place inside faces representing outside and inside surfaces of a single pane and
or outside of the building to compute the illumination levels. The other four representing the sides of the pane (Fig. 5B). Translat-
prototype automatically translates the coordinates of the sensor ing the window glass pane solid into 6 individual glass panes will
locations into the sensor points coordinates for both Radiance and lead to a wrong model in Radiance because Radiance would create
DAYSIM. internal reflection of light among the 6 surfaces of the glass pane.
In Radiance the window should be represented as a single surface
3.6. Weather file for DAYSIM without any thickness (Fig. 5C). To represent the Revit window glass
pane correctly with a Radiance window model (i.e., from Fig. 5A to
DAYSIM for daylighting simulation uses a different sky descrip- C) only one surface of the Revit glass pane has to be translated. An
tion program called Gendaylit, which uses a weather file to generate algorithm was developed to translate only one surface (i.e. front
a sky description for annual daylighting analysis based on the surface of the glass pane) into a Radiance glass surface.
Perez all-weather sky model [8]. It uses the EnergyPlus weather
data format, which is readily available for many different loca- 3.9. Running the prototype (Revit Add-in programs)
tions throughout the world. In the present project, the EnergyPlus
weather file is formatted automatically into the DAYSIM weather In order to perform the daylighting simulation the Add-in pro-
format by the epw2wea program [35], which comes with DAYSIM grams need to be loaded into Revit and accessed under the “External
installation. Tools” menu. Fig. 6 shows the Add-In Manager with all the Add-in
programs loaded. The sequence (a–e) shows the steps to load the
3.7. Model Translator programs and the sequence (1–4, displayed alphabetically based
on the program names in Add-In Manager) shows the order of
The main add-in program of the prototype is the Model Trans- execution of the programs to perform BIM-based daylighting anal-
lator. It translates the project information, including geometry, ysis. Once the main program (4) is executed, the Revit model is
materials, weather, location, date and time, and sensor point data translated into Radiance and DAYSIM input files, and Radiance and
into Radiance/DAYSIM input files. The following are the set of input DAYSIM programs are launched automatically to perform the day-
files created and organized into a project folder structure created lighting simulation (Fig. 7).
by Model Translator.

• A Radiance material file, which has Radiance material description 4. Test cases and validation
translated from the Revit materials.
• Radiance geometry file, which has Radiance geometry description We used two test cases to demonstrate the BIM to daylight-
translated from the Revit model geometry. ing simulation process, the tool Revit2Radiance, the results, and
• A Radiance camera view file for generating Radiance images. validation.
• Radiance sky description files.
• A weather file necessary for DAYSIM. 4.1. Test case 1 and validation
• Two batch files, which have Radiance and DAYSIM commands
to sequentially execute the Radiance and DAYSIM routines to For testing the prototype, the BESTEST Case 600 building’s BIM
produce the results. model was created in Revit (Fig. 8). The reason for using this build-
ing for testing is that the same BIM model of the building can be
Model Translator also runs the batch files that automatically used for testing the translations from BIM to both thermal and day-
launch both the programs. Once the simulations are complete all lighting simulations, while the thermal simulation results can be
the results are automatically written to a folder in the project direc- validated with benchmark results [25]. Fig. 8 shows the 3D BIM
tory. model, the floor plan, the section, the windows’ locations, and the
placement of the daylight sensors. Table 3 provides the project and
3.8. Window glass pane translation material description, which are used for Radiance and DAYSIM sim-
ulations. The weather location used for the simulation is Denver,
One of the challenges encountered in translating the geome- Colorado. The sky description is for summer solstice (June 21st)
try of BIM models into Radiance is the translation of window glass with the sun option turned on in the Gensky program. The time for
panes. BIM represents the building geometry as surfaces or solids, the simulation in Radiance is set to 12:00 PM.
398 S. Kota et al. / Energy and Buildings 81 (2014) 391–403

Fig. 5. Window pane representation: (A) A solid representing a window pane in a Revit model; (B) six individual surfaces of the solid in Revit; and (C) radiance representation
of the window pane, which is a surface without thickness.

The results of the experiment include the results of the Radi- the accuracy of the simulation results since the simulation engines
ance and DAYSIM programs. The results from the Radiance program Radiance and DAYSIM are already validated in other research.
are primarily images, which have photometric data. The results of Instead, the aim of the validation is to ascertain that the Revit
the DAYSIM program are annual illumination data. The simulation model is accurately translated into the Radiance/DAYSIM model.
results are then written into a result folder in the project directory. A visual comparison was done to check the accuracy of the
The Radiance image generating utility program takes geometry and geometry translation. The Radiance model, which is generated
materials files, the sky description file, and the view description to by our Revit2Radiance, is visualized using the RVIEW program.
generate an image (Fig. 9A) that shows the illumination levels. Once The view generated by the RVIEW program exactly matches the
the image is generated, subsequently three images are generated Revit model showing the accuracy of the geometry translation
from the first image by post-processing: (1) the human sensitive (Fig. 11). Fig. 12 shows the Revit material properties for a material
image (Fig. 9B), (2) the iso lux contour plot (Fig. 9C), and (3) the “Sash”. Revit2Radiance uses Revit RGB and computes the Average
false color image (Fig. 9D). Similar to Radiance, by taking the geom- Reflectance value using Eq. (1). As described in Section 3.1, the rea-
etry, material, weather, and sensor point files, DAYSIM performs son for using the same Average Reflectance for all Radiance RGB
an annual illumination simulation at the specified sensor points components is to render a grayscale image in order to visualize only
(Fig. 10). An annual illumination profile file (*.ill) is then written the illumination levels. (For rendering a color image, Eq. (1) will not
into the project folder. be used by Revit2Radiance). Revit2Radiance also takes the custom
For the present prototype Revit2Radiance, a validation study parameter values “Specularity” and “Roughness” to write the Radi-
was conducted. The validation study was not aimed to find out ance material description. Table 4 shows the manually calculated

Fig. 6. Launching the Revit2Radiance prototype.

 S. Kota et al. / Energy and Buildings 81 (2014) 391–403 399

Fig. 7. Daylighting simulation in Radiance and DAYSIM based on a BIM model. The Radiance commands are executed automatically by the Revit2Radiance Add-ins.

values for the Average Reflectance for Revit material “Sash” and (0.73) match exactly. Also the Specularity” and “Roughness” values
Fig. 13 shows the prototype output of Radiance material descrip- are passed from the Revit material to the Radiance material cor-
tion. Both the prototype-generated Average Reflectance (0.73 for rectly. This is a sample comparison that validates the accuracy of
RGB components) and the manually calculated Average Reflectance our material translation.

Fig. 8. BESTEST Case 600 BIM: 3D, floor plan, and section views of the model showing the dimensions of the room, the locations of the windows, and the sensors.
400 S. Kota et al. / Energy and Buildings 81 (2014) 391–403

Fig. 9. Radiance simulation results: (A) rendering, (B) human sensitivity image, (C) Iso-contour plots, and (D) false color image showing illumination levels.

Fig. 10. Annual DAYSIM illumination profile at a sensor point.

 S. Kota et al. / Energy and Buildings 81 (2014) 391–403 401

Fig. 11. (A) The Revit model and (B) RVIEW visualization of the Radiance model translated from the Revit model.

Fig. 12. Revit material information for “Sash” with custom parameters “Specularity” and “Roughness”.

Table 4 performing the daylighting simulation. The test case is

Manually calculated Average Reflectance for the material “Sash”.
Stanford University Solar Decathlon 2013 house project
Color channels Red Green Blue Average (http://solardecathlon.stanford.edu), the Revit model of which was
Value 224 178 126 Reflectance provided to us by the project team. Fig. 14A shows the floor plan
Normalized value (value/255) 0.878 0.698 0.494
and 3D views of the house. The main hall of the house has windows
Color coefficients (from Eq. (1)) 0.265 0.670 0.065
Normalized value × coefficients 0.233 0.468 0.032 0.73
on the north wall, which are clearstory windows (Fig. 14B) and
the curtain panel windows on the south wall (Fig. 14C). Table 5
gives the reflectance of different building elements and visible
4.2. Test case 2 and validation transmittance of windows. Both renderings and annual illumina-
tion profiles were produced using Revit2Radiance. The accuracy
The prototype was also tested on a complex case study of translating the model was tested. Once the Radiance file is
to check the accuracy in translating the geometry for
Table 5
Surface reflectance’s of different building element.

Building element Reflectance/Transmittance

Walls 0.5
Floor 0.2
Roof 0.8
Clearstory window transmittance 0.8
Curtain panel window transmittance 0.7
Fig. 13. Revit2Radiance prototype-generated Radiance material description for Southern side window transmittance 0.7
Revit material “Sash”.
402 S. Kota et al. / Energy and Buildings 81 (2014) 391–403

Fig. 14. (A) Floor plan of the Stanford Solar-decathlon 2013 house showing the camera location and direction and the sensor point location; (B) North-west isometric 3D
view showing the clear story windows on the north wall; and (C) South-west isometric 3D view showing the curtain panel windows on the south wall. (Source of images:
Stanford University Solar Decathlon 2013 project team. http://solardecathlon.stanford.edu).

Fig. 15. (A) The southwest view of the BIM model; (B) the southwest view of RVIEW rendering of the Radiance input file generated by Revit2Radiance; (C) the northwest
view of the BIM Model; and (D) the northwest view of RVIEW rendering of the Radiance input file generated by Revit2Radiance.

Fig. 16. Radiance simulation results of the Solar Decathlon house: (A) human sensitivity image; (B) false Color image; and (C) Iso-lux contour plot.

generated by Revit2Radiance, the Radiance model is viewed in the 5. Conclusion and future work
RVIEW program and is visually compared with the view of the BIM
model. Fig. 15(A and C) show the BIM model views and Fig. 15(B The research paper presents a method to enable direct inte-
and D) show the RVIEW views of the Radiance input file generated gration of BIM with daylighting simulation tools. The prototype
by Revit2Radiance. Upon visual inspection it can be concluded directly creates Radiance and DAYSIM input files from the Revit
that the prototype is accurately translating the BIM model into models through automated steps with high efficiency and accuracy.
the Radiance input file. Fig. 16(A–C) show the Radiance simulation In addition, the prototype is an easy-to-use tool for architects and
results of the Solar Decathlon house. designers without requiring in-depth knowledge of the simulation
 S. Kota et al. / Energy and Buildings 81 (2014) 391–403 403

tools. The research has identified that BIM does not have all the [5] G.J. Ward, F.M. Rubinstein, A new technique for computer simulation of illu-
information that is necessary for creating the simulation input files minated spaces, Journal of the Illuminating Engineering Society 17 (1) (1988)
80–91.
for Radiance and DAYSIM; however it provides options to incor- [6] J. Mardaljevic, Validation of a lighting simulation program under real sky con-
porate the required information. The research has also identified ditions, Lighting Research and Technology 27 (4) (1995) 181–188.
that the representations of the same building elements are not the [7] C.F. Reinhart, M. Andersen, Development and validation of a Radiance
model for a translucent panel, Energy and Buildings 38 (7) (2006) 890–
same between Revit and Radiance/DAYSIM. For example a glass 904.
pane, which is represented with a thickness in Revit, has to be rep- [8] C.F. Reinhart, O. Walkenhorst, Validation of dynamic RADIANCE-based daylight
resented as a surface without a thickness in Radiance and DAYSIM. simulations for a test office with external blinds, Energy and Buildings 33 (7)
(2001) 683–697.
The prototype has been validated for the accuracy in the geometry
[9] J. Mardaljevic, Simulation of annual daylighting profiles for internal illumi-
translation and proved to be accurate and reliable. With the para- nance, Lighting Research and Technology 32 (3) (2000) 111–118.
metric change capability of BIM, it will be easy for the architects and [10] C. Reinhart, P.F. Breton, Experimental validation of Autodesk® 3ds Max® Design
2009 and Daysim 3.0, Leukos 6 (1) (2009) 7–35.
designers to use the tool to quickly study different options since it
[11] C. Reinhart, M. Landry, P.F. Breton, Daylight simulation in 3ds Max Design 2009
provides a seamless link between Revit and Radiance/DAYSIM. The – Getting Started. Autodesk White Paper, 2009.
integration of daylighting analysis into the BIM environment helps [12] C. Reinhart, A. Fitz, Findings from a survey on the current use of daylight
in making informed design decisions. simulations in building design, Energy and Buildings 38 (7) (2006) 824–
835.
Our method depends upon Revit and thus lacks the format neu- [13] H. Erhorn, J. De Boer, M. Dirksmöller, ADELINE – an integrated approach to
trality of using IFC. Even though IFC data schema is being promoted lighting simulation, Proceedings of IBPSA Building Simulation’97 1 (1997) 79–
and developed to facilitate the interoperability among different 85.
[14] LBL, SUPERLITE 2.0 User Manual. Building Technology Program, Lawrence
BIM-based programs, it is thought to be complex to program and Berkeley Laboratory, Berkeley, CA, 1989.
implement in software [36]. Nonetheless, IFC is considered to play [15] LBNL, Desktop Radiance 2.0 BETA User Manual, Environmental Energy Tech-
an important role in the translation from BIM to energy simulation nologies Division, Building Technologies Department, Lawrence Berkeley
National Laboratory, Berkeley, CA, 2000.
[1]. The IFC approach has benefits compared with the approach [16] G.W. Larson, R. Shakespeare, C. Ehrlich, J. Mardaljevic, E. Phillips, P. Apian-
adopted in this paper as it can facilitate the data transfer between Bennewitz, Rendering With Radiance: The Art and Science of Lighting
different IFC compatible tools. The major limitation of the present Visualization, Morgan Kaufmann, San Francisco, CA, 1998.
[17] R. McNeel, Rhinoceros Version 5.0, Service Release 8 [Software], 2014, Available
method is that it works with Revit but not other BIM authoring
from http://www.rhino3d.com
tools. However, the concept and methodology can be applied to [18] K. Lagios, A. Jakubiec, J. Niemasz, C. Reinhart, Quick Start Guide for the DIVA-
any BIM authoring tools that support API programming. FOR-RHINO Toolbar, 2010, Retrieved from http://www.diva-for-rhino.com/
DIVA-for-Rhino QuickStartGuide.pdf
An extension of the present work could be to study the effect
[19] A.J. Marsh, Performance Analysis and Conceptual Design (Ph.D. dissertation),
of daylight on the electric lighting integrated into the BIM envi- The University of Western Australia, Perth, Australia, 1997.
ronment. Yan et al. [25] demonstrate the integration of thermal [20] A. Marsh, The Interaction of Radiance and ECOTECT, in: Presented at 5th Inter-
performance analysis into Revit using a Modelica-based thermal national Radiance Scientific Workshop, De Montfort University, Leicester, UK,
2006.
simulation engine. The next step could be integrating the results [21] C. Reinhart, Tutorial on the Use of Daysim Simulations for Sustainable Design,
of these two performance simulations (thermal and daylighting) 2010, Retrieved from http://www.daysim.com/pub/Daysim3.0.Tutorial.pdf
for multi-objective design optimization. As of now the prototype [22] F. Farias, Contemporary Strategies for Sustainable Design. Doctoral disser-
tation (Ph.D. dissertation), Texas A&M University, 2013, Retrieved from
can only handle simple fenestration systems; future improvement http://hdl.handle.net/1969.1/149379
could be to include the simulation of Complex Fenestration Sys- [23] M. Landry, P.F. Breton, Daylight simulation in 3ds Max Design 2009 – Advanced
tems (CFS) such as windows with blinds, fritted glass, and laser cut Concepts. Autodesk White Paper, 2009.
[24] B. Welle, J. Haymaker, Z. Rogers, ThermalOpt: a methodology for automated
panels. BIM-based multidisciplinary thermal simulation for use in optimization envi-
ronments, Building Simulation 4 (4) (2011) 293–313.
Acknowledgements [25] W. Yan, M. Clayton, J. Haberl, W. Jeong, J. Kim, S. Kota, J. Bermudez Alcocer,
M. Dixit, Interfacing BIM with Building Thermal and Daylighting Modeling, in:
Paper presented at 13th International Conference of the International Building
This material is based upon work supported by the National Sci- Performance Simulation Association (IBPSA), August 25–28, Chambery, France,
ence Foundation under Grant No. 0967446. Any opinions, findings, 2013.
[26] J. McGrew, Revit to Radiance (via an expensive translator. . .), 2006, Retrieved
and conclusions or recommendations expressed in this material are
from http://becausewecan.org/node/190
those of the authors and do not necessarily reflect the views of the [27] K. McHenry, P. Bajcsy, An Overview of 3d Data Content, File Formats and View-
National Science Foundation. We are grateful to the Stanford Solar ers, National Center for Supercomputing Applications, 2008, pp. 1205.
[28] S. Kumar, Interoperability between building information models (BIM) and
Decathlon 2013 project team and their Construction Manager Rob
energy analysis programs, 2008 (Master’s thesis): Available through ProQuest
Best and Project Manager Derek Ouyang for providing us with the Dissertations and Theses database.
BIM model of the project. We thank Jong Bum Kim and WoonSeong [29] G. Ward, The Material and Geometry Format, 1996, Retrieved from
Jeong for providing us with the Revit model of the BESTEST Case 600 http://radsite.lbl.gov/mgf/mgfhtml/intro.html
[30] B.S. Russell, Autodesk FBX-Based Revit Architecture 2009 to Autodesk 3ds Max
and their valuable input for the research project. Design 2009 Workflow: Autodesk White Paper, 2009.
[31] G. Mischler, dxf2rad 0.9 Convert Geometry from DXF to Radiance, 2000,
References Retrieved from http://www.schorsch.com/en/download/dxf2rad/ (accessed
12.06.14).
[32] G. Ward, OBJ2RAD, 2014, Retrieved from http://radsite.lbl.gov/radiance/
[1] V. Bazjanac, IFC BIM-based methodology for semi-automated building energy man html/obj2rad.1.html (accessed 12.06.14).
performance simulation, in: Presented at the CIB-W78 25th International Con- [33] S. Anger, J. Bowermaster, G. Ward, 3ds2mgf – convert 3D Studio binary
ference of Information Technology in Construction, Santiago, Chili, 2008. file to Materials and Geometry Format, 2014, Retrieved from http://
[2] N. Young, S. Jones, H.M. Bernstein, Building Information Modeling (BIM)- radsite.lbl.gov/mgf/mgfhtml/3ds2mgf.html (accessed 12.06.14).
Transforming Design and Construction to Achieve Greater Industry Produc- [34] G. Ward, mgf2rad – convert Materials and Geometry Format file to RADI-
tivity. SmartMarket Report, 48, 2008. ANCE description, 2014, Retrieved from http://radsite.lbl.gov/radiance/man
[3] S. Kota, J.S. Haberl, Historical survey of daylighting calculations methods and html/mgf2rad.1.html (accessed 12.06.14).
their use in energy performance simulations, in: Paper presented at 9th Interna- [35] C. Reinhart, DAYSIM Subprogram: epw2wea, 2014, Retrieved from
tional Conference for Enhanced Building Operations, Austin, Texas, November http://daysim.ning.com/subprogram/epw2wea (accessed 12.06.14).
17–19, 2009. [36] B. Dong, K.P. Lam, Y.C. Huang, G.M. Dobbs, A comparative study of the IFC
[4] G.J. Ward, The RADIANCE lighting simulation and rendering system, in: and gbXML informational infrastructures for data exchange in computational
Proceedings of the 21st Annual Conference on Computer Graphics and Inter- design support environments, Building Simulation (2007) 1530–1537.
active Techniques, ACM, 1994 July, pp. 459–472.