- Email: [email protected]
Building Information Modeling (BIM)-based daylighting simulation and analysis
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
Article history: Received 30 April 2014 Received in revised form 17 June 2014 Accepted 25 June 2014 Available online 3 July 2014 Keywords: Building Information Modeling BIM Daylighting simulation Radiance DAYSIM
a b s t r a c t Daylighting is an important aspect in designing high performance buildings. Many simulation tools have been developed to study the daylighting performance of buildings. These tools primarily use CAD environments for creating architectural models, which are then converted into daylighting models to run on the daylighting simulation engines. Once the architect defines the architectural model in CAD, a simulation 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 complexity depends on the tools. Currently, Building Information Modeling (BIM) is widely used in the AECO industries and BIM models are used as a means of exchanging data among different professionals involved in the design and construction of buildings. The present paper discusses the use of BIM for building performance simulations and mainly focuses on how daylighting analysis can be incorporated into a BIM environment, and what challenges and benefits exist in the process of integrating BIM with daylighting 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 Building performance analyses are important aspects of designing sustainable buildings. One of the performance analyses done by architects is to predict how buildings are performing in terms of their luminous environment as a result of daylighting. Daylighting analysis is done using either hand calculations or computer simulation tools. For simulation, once the architect defines the 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, and resource consuming process as the work mostly consists of manual or semi-manual translation from architectural model data to simulation data, which often results in numerous coding errors [1]. To ease the process of creating the input files, graphical user interfaces for defining model geometry have been created for simulation tools. In addition, geometry modeling tools (CAD tools) have been linked with daylighting simulation tools. Currently Building Information Modeling (BIM) is being used in building design in the Architecture/Engineering/Construction/Operation (AECO) industry
∗ Corresponding author. Tel.: +1 979 8450584. E-mail address: [email protected] (W. Yan). http://dx.doi.org/10.1016/j.enbuild.2014.06.043 0378-7788/© 2014 Elsevier B.V. All rights reserved.
[2] by many design firms in their design process. Research in the integration of BIM with building performance tools has been a main focus of both the developers of the BIM authoring tools as well as the Building Performance Simulation community. The following sections present a methodology of integrating BIM authoring tools (Autodesk Revit as a sample BIM tool) with daylighting analysis tools (Radiance and DAYSIM as sample simulation tools, which are widely used to study the daylighting performance of buildings).
2. Literature review 2.1. Daylighting simulation tools Over the years many analysis tools have been developed to study the daylighting performance of buildings. These tools extend from simple charts to sophisticated computer tools. A brief historical development of these tools can be found in literature [3]. Among these tools, Radiance is considered to be a state-of-the-art 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 systematically over the last decade [6–8]. Studies have shown that Radiance can be combined with a Daylight Coefficient (DC) [9]
392
S. Kota et al. / Energy and Buildings 81 (2014) 391–403
approach to reliably model the changing levels of daylight in a building over the course of a year. DAYSIM [8], which is an annual DC-based daylighting simulation tool with Radiance as its engine, has also been validated and is widely used in research and practice. Another daylighting simulation tool is Autodesk 3ds Max [10]. 3ds Max uses mental ray for doing daylighting analysis. It offers both a snapshot of daylight levels at a specified time point as well as annual illumination levels at the user specified locations in the building [11]. 3ds Max has been validated against Radiance and DAYSIM and proven to be accurate in predicting annual indoor illumination levels. Although there are several daylighting simulation tools available, the authors have chosen to use Radiance and DAYSIM for the present project as they are accurate, validated, and widely used simulation tools [12]. 2.2. CAD to daylighting Many of the daylighting simulation tools require defining the geometry of a building in a three dimensional (3D) coordinate system. Several of the sophisticated tools use text-based input files for modeling the geometry and the materials of the building. The syntax of a simulation input file is specific to the tool. For example, Radiance and DAYSIM input files are human readable text files that can also be interpreted by the computer programs. However, manually preparing an input file is a time-consuming process and often vulnerable to input mistakes. In order to make input file preparation more convenient and easy, several CAD modeling tools have been linked with daylighting simulation tools. Generally the CAD tools assist in the preparation of the geometry and material properties, and subsequently export the model into text-based, tool-specific input files to do simulation. Some of these tools are discussed below. 2.2.1. Advanced Daylight and Electric Lighting Integrated New Environment (ADELINE 2.0) ADELINE is an integrated lighting design computer tool for doing daylighting and electric lighting simulation [13]. ADELINE comes with SCRIBE-Modeler as a CAD interface and it connects to SUPERLITE [14] and Radiance to do daylighting and lighting simulation. It is also linked with energy simulation tools using a special program called SUPERLINK. Once the geometry is created with SCRIBEModeler, it is assigned materials using a program called PLINK, which also helps to define climate data for doing daylighting analysis. 2.2.2. Desktop Radiance Desktop Radiance is a CAD-based daylighting and electric lighting simulation tool with AutoCAD as a geometry modeler and Radiance as a simulation engine [15]. For linking AutoCAD with Radiance a special program called RADOUT was created [16]. It runs in the AutoCAD environment and facilitates the translation of CAD geometry into Radiance input files. Apart from having a CAD interface for creating geometry, Desktop Radiance also has a Radiance material library for assigning materials. RADOUT provides an interface to selectively export the geometry. Once the Radiance input file is created, different Radiance utilities are called using MS Windows BATCH scripting for doing a daylighting analysis [15]. In addition to AutoCAD, Radiance also has many utility programs that help in converting different geometry formats used by CAD-based tools into the Radiance format [16]. 2.2.3. DAYSIM/DIVA Another CAD-based modeling tool is Rhinoceros, a NURBS modeling program [17]. It has been linked to Radiance and DAYSIM through a plug-in called DIVA [18]. DIVA provides a graphical user interface to assign materials to different surfaces created in
Rhinoceros and export the models to Radiance and DAYSIM to perform the simulation. 2.2.4. Ecotect-Radiance-DAYSIM Ecotect is a widely used tool in the building simulation community for doing daylighting simulation [19]. Ecotect provides a CAD-based modeling environment for creating building geometry, assigning materials to geometry, and exporting the model to Radiance [20] and DAYSIM [21]. Fig. 1 shows linking of CAD-based tools with Radiance and DAYSIM. 2.2.5. Advantages and limitations of CAD-based daylighting simulation tools CAD-based daylighting simulation tools help in creating the geometry quickly. Without the aid of CAD interfaces it would be very time consuming to define geometry and material properties manually for daylighting simulations. However, some of the CAD-based tools have certain limitations in terms of creating geometry. These limitations are generally to require an idiosyncratic or specific method of defining geometry. For example in Desktop Radiance to create a window in a wall, the window glass has to be modeled as a single surface and the wall surface hosting the window has to be divided into four different surfaces. Similarly, in DIVA for Rhinoceros, windows have to be modeled as surfaces with zero thickness. The user has to adhere to these rules for creating daylighting simulation models. Also, if the building geometry has to be represented with thickness but the daylighting surfaces are modeled with zero thickness, then the complexity and time/effort for defining the model increase with the increase in the number of surfaces considerably. Furthermore, the materials have to be assigned by selecting individual surfaces, which is also a tedious task if the model is complex. 2.3. Building Information Modeling to daylighting Building Information Modeling (BIM) is an emerging technology in the AECO industry [22]. According to a public report there is a considerable increase in the use of BIM in AECO [2]. BIM tools not only help in designing a building using 3D graphics, but also are capable of holding non-graphical information such as material properties related to the building elements, which are not available in CAD tools. For example a wall can be modeled as an object with layers of materials using different properties. In addition, BIM tools help in modeling the geometry very quickly with a high degree of accuracy and detail compared to CAD tools by enforcing topology and connectivity among elements such as walls, windows, doors, roofs, and floors. Also, it is very easy to parametrically change the dimensions of the BIM objects and spaces with the geometry updating automatically based on any change. BIM is not only used to create building geometry, but also as a repository of the building information, which can be retrieved to perform different analyses on buildings, such as energy analysis, daylighting analysis, cost estimation, and structural analysis. 2.3.1. From BIM to CAD-based daylighting (Autodesk Revit to 3ds Max Design) 3ds Max Design is a product of Autodesk that is used mostly for animation and rendering. A module called Exposure in 3ds Max Design enables the user to do a physically based accurate daylighting simulation. The simulation engine is mental ray. Both time series rendering of High Dynamic Range (HDR) images and climate-based annual daylighting simulation can be conducted using Exposure. It has been validated comparing with measured data as well as with Radiance and DAYSIM daylighting simulating tools for a number of cases that extend from a simple window to complex fenestration systems consisting of shades, light shelves, and indoor venetian 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 tool, 3ds Max Design, is achieved through a file format called FBX [11]. The FBX file format contains the geometry, material information, and camera views necessary for the daylighting analysis in 3ds Max Design. However, in order to do a physically accurate daylighting simulation, a series of manipulations has to be performed initially such as replacing the lights in the Revit model with a daylighting system in 3ds MAX Design, and replacing the nonphysically based materials of Revit with physically based materials. Along with these changes, some of the changes have to be done to the Revit model, especially when modeling window glass panes. A detailed workflow of doing daylighting analysis using 3ds Max Design from Revit can be found in the literature [23]. Even though the process is fast compared with CAD-based daylighting simulation, it still requires manual checking and preparing the scene for accurate daylighting simulation. The process can be described as semi-automated. 2.3.2. ThermalOpt daylighting simulation ThermalOpt is a methodology for automated BIM-based simulation intended for use in multidisciplinary design optimization (MDO) environments [24]. It can do both thermal and daylighting simulations, supporting study of the effect of daylighting on the thermal performance. It uses EnergyPlus for thermal simulation and Radiance for daylighting simulation. ThermalOpt is implemented via four different plugins: (1) IFC2ThermalSim, (2) ThermalSim, (3) EnergyPlus Wrapper, and (4) Radiance Wrapper. In ThermalOpt, the building description is represented in IFC BIM interoperability format and translated into EnergyPlus input format. Geometry for daylighting simulation is taken from the EnergyPlus model. The EnergyPlus model does not represent the actual geometry of BIM. Taking geometry from EnergyPlus does not give accurate geometry for daylighting simulation. Even though the whole process is automated, the geometry for the daylighting simulation is not the same as the BIM model and this can alter the results. Learning from the shortcomings of the above studies, it was thought necessary to develop a process and methodology to accurately translate the BIM models’ geometry and material information into daylighting models (Radiance and DAYSIM). This was achieved by developing custom add-in programs that can take Revit BIM models and translate them into Radiance and DAYSIM input files, and then automatically perform the simulation without any
manual intervention. The main objective of the research project is to make the process quick and friendly to architects and engineers. The following sections will describe the whole process in detail. 2.3.3. Revit to Radiance and DAYSIM The present research is a part of a project on Physical Building Information Modeling (PBIM). The aim of the project is to integrate BIM with building performance analysis tools in multiple domains, including thermal and daylighting [25]. Even though there are several BIM authoring tools available in the market today, Autodesk Revit has been adopted for the present project because of its parametric BIM authoring capability and its wide acceptance in the industry. The main aim of the project is to link BIM with Radiance and DAYSIM by translating the Revit models into Radiance/DAYSIM input files and automatically performing the simulation. In a prior study [26] a Revit model is translated into a Radiance input file through other file formats and tools. Some of the major findings reported by the study are (1) translating Revit data into Radiance requires several steps to convert Revit geometry into Radiance geometry, (2) manual intervention is required in the whole process, and (3) certain information is required to be added manually to create a complete Radiance input file. In our project, we aim to develop a direct translator from Revit to Radiance and DAYSIM in order to significantly reduce the model preparation time and, in the meantime, preserve the parametric modeling advantages of Revit. 2.4. Radiance utility programs Several methods were explored to convert a Revit model to a Radiance input file. Radiance has several utility programs for converting a wide variety of CAD geometry data formats into the Radiance format. Some of the widely recognized data formats, which are used by different CAD-based tools, are 3DS [27] and DXF [28] developed by Autodesk, OBJ [27] developed by Wavefront and the MGF developed particularly for Radiance [29]. Autodesk offers a comprehensive exporting utility called FBX Converter that can convert the Revit file into the FBX format [30] and further into different geometry formats such as DWG, DXF, OBJ and 3DS. For the present project, to explore existing methods for converting BIM to daylighting simulation models, Revit models 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 [32], 3DS2MGF [33], and MGF2RAD [34] utility programs are used to convert these formats into the Radiance format. Fig. 2 shows different steps involved in converting a Revit model into a Radiance input file using these different utilities. The steps involved in translation using different utilities are represented as paths (i.e. P1, P2, P3,. . .). P6 is the translation path describing our method presented in this paper. The method proposed is a more direct path involving only one step for creating the Radiance input file and also performing the simulation. The main advantage of this method is that it is easy and does not require in-depth knowledge of different
tools, which are needed in other paths. Table 1 shows the number of steps involved in each method for converting a Revit model into Radiance/DAYSIM input files. 2.5. Issues involved in Revit to daylighting using Radiance utilities In all the existing approaches, geometry information could be exported automatically using the utilities but not material information. Manual assignment is needed for materials. This is a major setback for creating Radiance input files. Another major setback 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 exporting/translating
Importing to other programs
Assigning materials or Exporting to other formats
Exporting/Converting
Exporting/Converting
No. of steps
P1
Exported into DWG format
DWG format imported into 3DS Max
Materials assigned in 3ds Max
Exported into Wavefront OBJ format
5
P2
Exported to FBX format
Imported into Autodesk FBX converter utility
Exported into 3DS format *.3ds
Converted into MGF format using 3DS2MGF
Exported into Radiance format using OBJ2RAD utility Converted into Radiance format using MGF2RAD utility
P3
Exported to FBX format
Imported into Autodesk FBX converter utility
Exported into Wavefront OBJ format *.OBJ
P4
Exported to FBX format
Imported into Autodesk FBX converter utility
Exported into DXF format *.DXF
P5
Exported to FBX format
Imported into Autodesk 3ds Max
Materials assigned in 3ds Max
Converted into Radiance format using OBJ2RAD utility Converted into Radiance format using DXF2RAD utility Simulated in 3ds Max
P6
Translated into Radiance input file using our Revit2Radiance Add-in Program
5
4
4
4 1
S. Kota et al. / Energy and Buildings 81 (2014) 391–403
are modeled with thickness whereas in Radiance window panes are modeled with zero thickness – this means manual adjustment is needed for modeling the windows with the existing methods. From our experiments, some errors were identified when converting OBJ data into Radiance using OBJ2RAD utility. Manual inspection of the OBJ geometry file created by FBX Converter revealed some strange characters in the OBJ input file that caused errors in translating using the OBJ2RAD utility. We also observed that converting OBJ and DXF to RAD was a one-step process, but converting 3DS to RAD was a two-step process. During the process of converting 3DS to MGF using the 3DS2MGF utility, an error occurred resulting in the MGF file not being created. From our experiments, the only path that has successfully translated geometry from a Revit model to Radiance (without material definition) is DXF2RAD. Each of these methods has its own limitations. There are many steps involved in the process of translating the Revit model into the Radiance input file. In addition, a thorough knowledge about each tool is required to carry out the process, which can be a significant challenge for even an experienced architect. As the architect changes the design iteratively the whole process becomes a tedious task, as 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. Owing to these reasons, we proposed to translate the Revit models into Radiance input files directly and automatically (i.e., integrating the daylighting analysis into the Revit modeling environment), which does not involve human intervention and results in a process that is quick and user friendly. 3. Methodology – prototyping Our research method includes prototyping and validating the prototype with building models. For facilitating the data transfer from Revit to Radiance and DAYSIM, a prototype Revit2Radiance was developed. The prototype takes the building information, both geometric and non-geometric such as material properties, from a BIM model and translates it into Radiance and DAYSIM input files. The prototype was developed using Revit’s Application Programing Interface (API) with the C# programming language. Based on BIM (geometry, materials, location, date/time, a camera view, and sensor points), the prototype extracts the building surfaces in triangle meshes and other data needed to generate the input files of Radiance and DAYSIM. The main objective for developing the prototype is to translate the Revit models into Radiance/DAYSIM input files and perform the daylighting simulation and visualization without manual intervention in the translation process (after the BIM model is created and data are entered into BIM). This not only helps avoid errors due to manual input of the data in preparing simulation input files, but also makes the process of performing the analysis more user-friendly for the designers who are not well acquainted with the simulation tools. For the present prototype a total of four Revit add-in programs were developed that can automatically translate BIM models into simulation models, simulate the building performance, and generate analysis results. Table 2 shows the add-in programs and respective tasks they accomplish. Fig. 3 shows the overall workflow of Revit2Radiance. Fig. 4 shows the workflow with the add-in programs. In the subsequent sections each of the add-in programs is explained in detail.
395
Table 2 Revit add-in programs (Revit2Radiance modules) for translating Revit models into Radiance input files. Name of the add-in program
Action accomplished by the add-in
1
Parameter File Creation
2
Adding Material Parameter
3
Reading Material Information Model Translator
Creates a default shared parameter file in the project directory for adding shared parameters that are needed for defining building materials in Add-in #2 below Adds user-defined shared parameters for building materials that are necessary for holding materials data for Radiance models Reads values of the created shared material parameters Translates the Revit model’s geometry and material data into Radiance/DAYSIM input files and automatically run the simulation programs to generate the results
4
multiple parameters and values assigned to it. For example, one of the material parameters is “color” which gives the color property of the material in terms of the Red, Green, and Blue (RGB) channels. In Radiance there are different material types [16]. One of the basic material types is ‘Plastic’ which has a color component, diffuse and specular components, and a roughness factor. The color is defined by Red, Green and Blue reflectance values [16]. Given the color values in RGB of a surface it is possible to compute the reflectance of the material through a brightness formula. The Average Reflectance value is computed from the following formula for the standard Radiance R, G, and B components of the reflectance [16]: Average Reflectance = 0.265 × R + 0.670 × G + 0.065 × B
(1)
where R, G, and B are normalized values. In our prototype, we used the default Revit material R, G, and B values (in the range of 0–255), normalized them, and calculated the Average Reflectance using the above equation. The same Average Reflectance is used as the Radiance RGB values for calculating the illumination levels and rendering grayscale images by Radiance. The rendered grayscale images can help visualize the illumination levels without the effects of chromaticity. For rendering color images, Eq. (1) does not need to be called by Revit2Radiance and the normalized Revit RGB, instead of the Average Reflectance value, will be used as Radiance RGB. In addition, if the users intend to use measured RGB of a material instead of the default values in Revit, they can modify the values in Revit. However, Revit materials do not have specularity and roughness values that are necessary for complete Radiance material description. It is necessary that these values be first added to each material in the Revit material database. Revit offers the functionality for adding user-defined parameters called custom parameters. In the present project this functionality was used through API programming to add specularity and roughness parameters to each Revit material. In order to add user-defined parameters, a shared parameter file must be created first for the parameters. The first add-in 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 associated with the geometry. Materials are assigned to geometry through the Revit material database. Each material in Revit has
Once the default shared parameter file is created, the second add-in program “Adding Material Parameter” adds the two 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 Revit. 3.3. Reading Material Information As described briefly in the earlier section “Parameter File Creation”, Radiance materials have reflectance, specularity and roughness values. Reflectance is defined with individual channel reflectance (i.e. red-reflectance, blue-reflectance and greenreflectance) along with specularity and roughness. Specularity and roughness values must be added to the parameters created for each Revit material (i.e. adding appropriate values to the shared parameters for each material). In order to ascertain specularity and roughness values of Revit materials, the Revit materials were compared with similar material types for which Radiance material data are available from different databases. The specularity and roughness values of these Radiance materials were then adopted for Revit materials. A list was created in an Excel spreadsheet
containing Revit material names with their corresponding specularity and roughness values taken from Radiance materials. The “Reading Material Information” add-in program reads the Excel database file containing the specularity and roughness values and assigns these values to the custom parameters created for each material in the earlier steps. At this point each material in the Revit material database has all the necessary information that must be translated into a Radiance material description.
3.4. Radiance sky model and view description Radiance requires a sky, a view and a scene description. There are different ways a sky description can be modeled in Radiance. By providing the latitude, longitude, and time of the day, the Radiance utility Gensky [16] can create a sky description. In Revit, each view has sun setting parameters, which also provide access 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
the day (month, day, and hour) for shadow calculating and image rendering purposes. For the present prototype, this information is accessed, formatted, and provided to Gensky program to create the sky description by our prototype. Similarly, the camera view information is translated from BIM into a Radiance view description that is necessary to produce Radiance images that can be post-processed for different analyses. 3.5. Sensor point description In Radiance the simulation results can be displayed as images, numerical values, and illumination contour plots. Apart from producing a picture that has the photometric information for a given camera view and scene description, Radiance can also compute lighting levels at a specific location inside a given space by specifying a sensor point. The sensor point is specified by giving the X, Y and Z coordinates and view directions at which Radiance calculates the illumination levels. For the prototype, a sensor object was created as a custom Revit family that the user can place inside or outside of the building to compute the illumination levels. The prototype automatically translates the coordinates of the sensor locations into the sensor points coordinates for both Radiance and DAYSIM. 3.6. Weather file for DAYSIM DAYSIM for daylighting simulation uses a different sky description program called Gendaylit, which uses a weather file to generate a sky description for annual daylighting analysis based on the Perez all-weather sky model [8]. It uses the EnergyPlus weather data format, which is readily available for many different locations throughout the world. In the present project, the EnergyPlus weather file is formatted automatically into the DAYSIM weather format by the epw2wea program [35], which comes with DAYSIM installation. 3.7. Model Translator The main add-in program of the prototype is the Model Translator. It translates the project information, including geometry, materials, weather, location, date and time, and sensor point data into Radiance/DAYSIM input files. The following are the set of input files created and organized into a project folder structure created by Model Translator. • A Radiance material file, which has Radiance material description translated from the Revit materials. • Radiance geometry file, which has Radiance geometry description translated from the Revit model geometry. • A Radiance camera view file for generating Radiance images. • Radiance sky description files. • A weather file necessary for DAYSIM. • Two batch files, which have Radiance and DAYSIM commands to sequentially execute the Radiance and DAYSIM routines to produce the results. Model Translator also runs the batch files that automatically launch both the programs. Once the simulations are complete all the results are automatically written to a folder in the project directory. 3.8. Window glass pane translation One of the challenges encountered in translating the geometry of BIM models into Radiance is the translation of window glass panes. BIM represents the building geometry as surfaces or solids,
397
Table 3 Project and material description. Building parameters
Values
Location Length of the room Width of the room No. of windows Dimensions of windows Wall reflectance Floor reflectance Roof/ceiling reflectance Window transmittance
Denver, CO 6M 8M 2 3M×2M 0.5 0.2 0.8 0.75
which can have multiple surfaces depending on the shape of the geometry (Fig. 5A shows a solid of a window pane). The Model Translator add-in converts each surface of a solid into a Radiance geometry surface with material information attached to it. In Revit, window panes are mostly cuboids of 6 surfaces with two of the surfaces representing outside and inside surfaces of a single pane and other four representing the sides of the pane (Fig. 5B). Translating the window glass pane solid into 6 individual glass panes will lead to a wrong model in Radiance because Radiance would create internal reflection of light among the 6 surfaces of the glass pane. In Radiance the window should be represented as a single surface without any thickness (Fig. 5C). To represent the Revit window glass pane correctly with a Radiance window model (i.e., from Fig. 5A to C) only one surface of the Revit glass pane has to be translated. An algorithm was developed to translate only one surface (i.e. front surface of the glass pane) into a Radiance glass surface. 3.9. Running the prototype (Revit Add-in programs) In order to perform the daylighting simulation the Add-in programs need to be loaded into Revit and accessed under the “External 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 programs and the sequence (1–4, displayed alphabetically based on the program names in Add-In Manager) shows the order of execution of the programs to perform BIM-based daylighting analysis. Once the main program (4) is executed, the Revit model is translated into Radiance and DAYSIM input files, and Radiance and DAYSIM programs are launched automatically to perform the daylighting simulation (Fig. 7). 4. Test cases and validation We used two test cases to demonstrate the BIM to daylighting simulation process, the tool Revit2Radiance, the results, and validation. 4.1. Test case 1 and validation For testing the prototype, the BESTEST Case 600 building’s BIM model was created in Revit (Fig. 8). The reason for using this building for testing is that the same BIM model of the building can be used for testing the translations from BIM to both thermal and daylighting simulations, while the thermal simulation results can be validated with benchmark results [25]. Fig. 8 shows the 3D BIM model, the floor plan, the section, the windows’ locations, and the placement of the daylight sensors. Table 3 provides the project and material description, which are used for Radiance and DAYSIM simulations. The weather location used for the simulation is Denver, Colorado. The sky description is for summer solstice (June 21st) with the sun option turned on in the Gensky program. The time for 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 Radiance and DAYSIM programs. The results from the Radiance program are primarily images, which have photometric data. The results of the DAYSIM program are annual illumination data. The simulation results are then written into a result folder in the project directory. The Radiance image generating utility program takes geometry and materials files, the sky description file, and the view description to generate an image (Fig. 9A) that shows the illumination levels. Once the image is generated, subsequently three images are generated from the first image by post-processing: (1) the human sensitive image (Fig. 9B), (2) the iso lux contour plot (Fig. 9C), and (3) the false color image (Fig. 9D). Similar to Radiance, by taking the geometry, material, weather, and sensor point files, DAYSIM performs an annual illumination simulation at the specified sensor points (Fig. 10). An annual illumination profile file (*.ill) is then written into the project folder. For the present prototype Revit2Radiance, a validation study was conducted. The validation study was not aimed to find out
the accuracy of the simulation results since the simulation engines Radiance and DAYSIM are already validated in other research. Instead, the aim of the validation is to ascertain that the Revit model is accurately translated into the Radiance/DAYSIM model. A visual comparison was done to check the accuracy of the geometry translation. The Radiance model, which is generated by our Revit2Radiance, is visualized using the RVIEW program. The view generated by the RVIEW program exactly matches the Revit model showing the accuracy of the geometry translation (Fig. 11). Fig. 12 shows the Revit material properties for a material “Sash”. Revit2Radiance uses Revit RGB and computes the Average Reflectance value using Eq. (1). As described in Section 3.1, the reason for using the same Average Reflectance for all Radiance RGB components is to render a grayscale image in order to visualize only the illumination levels. (For rendering a color image, Eq. (1) will not be used by Revit2Radiance). Revit2Radiance also takes the custom parameter values “Specularity” and “Roughness” to write the Radiance 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 Fig. 13 shows the prototype output of Radiance material description. Both the prototype-generated Average Reflectance (0.73 for RGB components) and the manually calculated Average Reflectance
(0.73) match exactly. Also the Specularity” and “Roughness” values are passed from the Revit material to the Radiance material correctly. This is a sample comparison that validates the accuracy of 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 Manually calculated Average Reflectance for the material “Sash”. Color channels Value Normalized value (value/255) Color coefficients (from Eq. (1)) Normalized value × coefficients
Red 224 0.878 0.265 0.233
Green 178 0.698 0.670 0.468
Blue 126 0.494 0.065 0.032
Average Reflectance
0.73
4.2. Test case 2 and validation
to
The prototype was also tested on a complex case study check the accuracy in translating the geometry for
performing the daylighting simulation. The test case is Stanford University Solar Decathlon 2013 house project (http://solardecathlon.stanford.edu), the Revit model of which was provided to us by the project team. Fig. 14A shows the floor plan and 3D views of the house. The main hall of the house has windows 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 transmittance of windows. Both renderings and annual illumination profiles were produced using Revit2Radiance. The accuracy of translating the model was tested. Once the Radiance file is Table 5 Surface reflectance’s of different building element.
Fig. 13. Revit2Radiance prototype-generated Radiance material description for Revit material “Sash”.
Building element
Reflectance/Transmittance
Walls Floor Roof Clearstory window transmittance Curtain panel window transmittance Southern side window transmittance
0.5 0.2 0.8 0.8 0.7 0.7
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 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 and D) show the RVIEW views of the Radiance input file generated by Revit2Radiance. Upon visual inspection it can be concluded that the prototype is accurately translating the BIM model into the Radiance input file. Fig. 16(A–C) show the Radiance simulation results of the Solar Decathlon house.
5. Conclusion and future work The research paper presents a method to enable direct integration of BIM with daylighting simulation tools. The prototype directly creates Radiance and DAYSIM input files from the Revit models through automated steps with high efficiency and accuracy. In addition, the prototype is an easy-to-use tool for architects and designers without requiring in-depth knowledge of the simulation
S. Kota et al. / Energy and Buildings 81 (2014) 391–403
tools. The research has identified that BIM does not have all the information that is necessary for creating the simulation input files for Radiance and DAYSIM; however it provides options to incorporate the required information. The research has also identified that the representations of the same building elements are not the same between Revit and Radiance/DAYSIM. For example a glass pane, which is represented with a thickness in Revit, has to be represented as a surface without a thickness in Radiance and DAYSIM. The prototype has been validated for the accuracy in the geometry translation and proved to be accurate and reliable. With the parametric change capability of BIM, it will be easy for the architects and designers to use the tool to quickly study different options since it provides a seamless link between Revit and Radiance/DAYSIM. The integration of daylighting analysis into the BIM environment helps in making informed design decisions. Our method depends upon Revit and thus lacks the format neutrality of using IFC. Even though IFC data schema is being promoted and developed to facilitate the interoperability among different BIM-based programs, it is thought to be complex to program and implement in software [36]. Nonetheless, IFC is considered to play an important role in the translation from BIM to energy simulation [1]. The IFC approach has benefits compared with the approach adopted in this paper as it can facilitate the data transfer between different IFC compatible tools. The major limitation of the present method is that it works with Revit but not other BIM authoring tools. However, the concept and methodology can be applied to any BIM authoring tools that support API programming. An extension of the present work could be to study the effect of daylight on the electric lighting integrated into the BIM environment. Yan et al. [25] demonstrate the integration of thermal performance analysis into Revit using a Modelica-based thermal simulation engine. The next step could be integrating the results of these two performance simulations (thermal and daylighting) for multi-objective design optimization. As of now the prototype can only handle simple fenestration systems; future improvement could be to include the simulation of Complex Fenestration Systems (CFS) such as windows with blinds, fritted glass, and laser cut panels. Acknowledgements This material is based upon work supported by the National Science Foundation under Grant No. 0967446. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. We are grateful to the Stanford Solar Decathlon 2013 project team and their Construction Manager Rob Best and Project Manager Derek Ouyang for providing us with the BIM model of the project. We thank Jong Bum Kim and WoonSeong Jeong for providing us with the Revit model of the BESTEST Case 600 and their valuable input for the research project. References [1] V. Bazjanac, IFC BIM-based methodology for semi-automated building energy performance simulation, in: Presented at the CIB-W78 25th International Conference of Information Technology in Construction, Santiago, Chili, 2008. [2] N. Young, S. Jones, H.M. Bernstein, Building Information Modeling (BIM)Transforming Design and Construction to Achieve Greater Industry Productivity. SmartMarket Report, 48, 2008. [3] S. Kota, J.S. Haberl, Historical survey of daylighting calculations methods and their use in energy performance simulations, in: Paper presented at 9th International Conference for Enhanced Building Operations, Austin, Texas, November 17–19, 2009. [4] G.J. Ward, The RADIANCE lighting simulation and rendering system, in: Proceedings of the 21st Annual Conference on Computer Graphics and Interactive Techniques, ACM, 1994 July, pp. 459–472.
403
[5] G.J. Ward, F.M. Rubinstein, A new technique for computer simulation of illuminated spaces, Journal of the Illuminating Engineering Society 17 (1) (1988) 80–91. [6] J. Mardaljevic, Validation of a lighting simulation program under real sky conditions, Lighting Research and Technology 27 (4) (1995) 181–188. [7] C.F. Reinhart, M. Andersen, Development and validation of a Radiance model for a translucent panel, Energy and Buildings 38 (7) (2006) 890– 904. [8] C.F. Reinhart, O. Walkenhorst, Validation of dynamic RADIANCE-based daylight simulations for a test office with external blinds, Energy and Buildings 33 (7) (2001) 683–697. [9] J. Mardaljevic, Simulation of annual daylighting profiles for internal illuminance, Lighting Research and Technology 32 (3) (2000) 111–118. [10] C. Reinhart, P.F. Breton, Experimental validation of Autodesk® 3ds Max® Design 2009 and Daysim 3.0, Leukos 6 (1) (2009) 7–35. [11] C. Reinhart, M. Landry, P.F. Breton, Daylight simulation in 3ds Max Design 2009 – Getting Started. Autodesk White Paper, 2009. [12] C. Reinhart, A. Fitz, Findings from a survey on the current use of daylight simulations in building design, Energy and Buildings 38 (7) (2006) 824– 835. [13] H. Erhorn, J. De Boer, M. Dirksmöller, ADELINE – an integrated approach to lighting simulation, Proceedings of IBPSA Building Simulation’97 1 (1997) 79– 85. [14] LBL, SUPERLITE 2.0 User Manual. Building Technology Program, Lawrence Berkeley Laboratory, Berkeley, CA, 1989. [15] LBNL, Desktop Radiance 2.0 BETA User Manual, Environmental Energy Technologies Division, Building Technologies Department, Lawrence Berkeley National Laboratory, Berkeley, CA, 2000. [16] G.W. Larson, R. Shakespeare, C. Ehrlich, J. Mardaljevic, E. Phillips, P. ApianBennewitz, Rendering With Radiance: The Art and Science of Lighting Visualization, Morgan Kaufmann, San Francisco, CA, 1998. [17] R. McNeel, Rhinoceros Version 5.0, Service Release 8 [Software], 2014, Available from http://www.rhino3d.com [18] K. Lagios, A. Jakubiec, J. Niemasz, C. Reinhart, Quick Start Guide for the DIVAFOR-RHINO Toolbar, 2010, Retrieved from http://www.diva-for-rhino.com/ DIVA-for-Rhino QuickStartGuide.pdf [19] A.J. Marsh, Performance Analysis and Conceptual Design (Ph.D. dissertation), The University of Western Australia, Perth, Australia, 1997. [20] A. Marsh, The Interaction of Radiance and ECOTECT, in: Presented at 5th International Radiance Scientific Workshop, De Montfort University, Leicester, UK, 2006. [21] C. Reinhart, Tutorial on the Use of Daysim Simulations for Sustainable Design, 2010, Retrieved from http://www.daysim.com/pub/Daysim3.0.Tutorial.pdf [22] F. Farias, Contemporary Strategies for Sustainable Design. Doctoral dissertation (Ph.D. dissertation), Texas A&M University, 2013, Retrieved from http://hdl.handle.net/1969.1/149379 [23] M. Landry, P.F. Breton, Daylight simulation in 3ds Max Design 2009 – Advanced Concepts. Autodesk White Paper, 2009. [24] B. Welle, J. Haymaker, Z. Rogers, ThermalOpt: a methodology for automated BIM-based multidisciplinary thermal simulation for use in optimization environments, Building Simulation 4 (4) (2011) 293–313. [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 Performance Simulation Association (IBPSA), August 25–28, Chambery, France, 2013. [26] J. McGrew, Revit to Radiance (via an expensive translator. . .), 2006, Retrieved from http://becausewecan.org/node/190 [27] K. McHenry, P. Bajcsy, An Overview of 3d Data Content, File Formats and Viewers, National Center for Supercomputing Applications, 2008, pp. 1205. [28] S. Kumar, Interoperability between building information models (BIM) and energy analysis programs, 2008 (Master’s thesis): Available through ProQuest Dissertations and Theses database. [29] G. Ward, The Material and Geometry Format, 1996, Retrieved from http://radsite.lbl.gov/mgf/mgfhtml/intro.html [30] B.S. Russell, Autodesk FBX-Based Revit Architecture 2009 to Autodesk 3ds Max Design 2009 Workflow: Autodesk White Paper, 2009. [31] G. Mischler, dxf2rad 0.9 Convert Geometry from DXF to Radiance, 2000, 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/ man html/obj2rad.1.html (accessed 12.06.14). [33] S. Anger, J. Bowermaster, G. Ward, 3ds2mgf – convert 3D Studio binary file to Materials and Geometry Format, 2014, Retrieved from http:// radsite.lbl.gov/mgf/mgfhtml/3ds2mgf.html (accessed 12.06.14). [34] G. Ward, mgf2rad – convert Materials and Geometry Format file to RADIANCE description, 2014, Retrieved from http://radsite.lbl.gov/radiance/man html/mgf2rad.1.html (accessed 12.06.14). [35] C. Reinhart, DAYSIM Subprogram: epw2wea, 2014, Retrieved from http://daysim.ning.com/subprogram/epw2wea (accessed 12.06.14). [36] B. Dong, K.P. Lam, Y.C. Huang, G.M. Dobbs, A comparative study of the IFC and gbXML informational infrastructures for data exchange in computational design support environments, Building Simulation (2007) 1530–1537.
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
Article history: Received 30 April 2014 Received in revised form 17 June 2014 Accepted 25 June 2014 Available online 3 July 2014 Keywords: Building Information Modeling BIM Daylighting simulation Radiance DAYSIM
a b s t r a c t Daylighting is an important aspect in designing high performance buildings. Many simulation tools have been developed to study the daylighting performance of buildings. These tools primarily use CAD environments for creating architectural models, which are then converted into daylighting models to run on the daylighting simulation engines. Once the architect defines the architectural model in CAD, a simulation 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 complexity depends on the tools. Currently, Building Information Modeling (BIM) is widely used in the AECO industries and BIM models are used as a means of exchanging data among different professionals involved in the design and construction of buildings. The present paper discusses the use of BIM for building performance simulations and mainly focuses on how daylighting analysis can be incorporated into a BIM environment, and what challenges and benefits exist in the process of integrating BIM with daylighting 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 Building performance analyses are important aspects of designing sustainable buildings. One of the performance analyses done by architects is to predict how buildings are performing in terms of their luminous environment as a result of daylighting. Daylighting analysis is done using either hand calculations or computer simulation tools. For simulation, once the architect defines the 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, and resource consuming process as the work mostly consists of manual or semi-manual translation from architectural model data to simulation data, which often results in numerous coding errors [1]. To ease the process of creating the input files, graphical user interfaces for defining model geometry have been created for simulation tools. In addition, geometry modeling tools (CAD tools) have been linked with daylighting simulation tools. Currently Building Information Modeling (BIM) is being used in building design in the Architecture/Engineering/Construction/Operation (AECO) industry
∗ Corresponding author. Tel.: +1 979 8450584. E-mail address: [email protected] (W. Yan). http://dx.doi.org/10.1016/j.enbuild.2014.06.043 0378-7788/© 2014 Elsevier B.V. All rights reserved.
[2] by many design firms in their design process. Research in the integration of BIM with building performance tools has been a main focus of both the developers of the BIM authoring tools as well as the Building Performance Simulation community. The following sections present a methodology of integrating BIM authoring tools (Autodesk Revit as a sample BIM tool) with daylighting analysis tools (Radiance and DAYSIM as sample simulation tools, which are widely used to study the daylighting performance of buildings).
2. Literature review 2.1. Daylighting simulation tools Over the years many analysis tools have been developed to study the daylighting performance of buildings. These tools extend from simple charts to sophisticated computer tools. A brief historical development of these tools can be found in literature [3]. Among these tools, Radiance is considered to be a state-of-the-art 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 systematically over the last decade [6–8]. Studies have shown that Radiance can be combined with a Daylight Coefficient (DC) [9]
392
S. Kota et al. / Energy and Buildings 81 (2014) 391–403
approach to reliably model the changing levels of daylight in a building over the course of a year. DAYSIM [8], which is an annual DC-based daylighting simulation tool with Radiance as its engine, has also been validated and is widely used in research and practice. Another daylighting simulation tool is Autodesk 3ds Max [10]. 3ds Max uses mental ray for doing daylighting analysis. It offers both a snapshot of daylight levels at a specified time point as well as annual illumination levels at the user specified locations in the building [11]. 3ds Max has been validated against Radiance and DAYSIM and proven to be accurate in predicting annual indoor illumination levels. Although there are several daylighting simulation tools available, the authors have chosen to use Radiance and DAYSIM for the present project as they are accurate, validated, and widely used simulation tools [12]. 2.2. CAD to daylighting Many of the daylighting simulation tools require defining the geometry of a building in a three dimensional (3D) coordinate system. Several of the sophisticated tools use text-based input files for modeling the geometry and the materials of the building. The syntax of a simulation input file is specific to the tool. For example, Radiance and DAYSIM input files are human readable text files that can also be interpreted by the computer programs. However, manually preparing an input file is a time-consuming process and often vulnerable to input mistakes. In order to make input file preparation more convenient and easy, several CAD modeling tools have been linked with daylighting simulation tools. Generally the CAD tools assist in the preparation of the geometry and material properties, and subsequently export the model into text-based, tool-specific input files to do simulation. Some of these tools are discussed below. 2.2.1. Advanced Daylight and Electric Lighting Integrated New Environment (ADELINE 2.0) ADELINE is an integrated lighting design computer tool for doing daylighting and electric lighting simulation [13]. ADELINE comes with SCRIBE-Modeler as a CAD interface and it connects to SUPERLITE [14] and Radiance to do daylighting and lighting simulation. It is also linked with energy simulation tools using a special program called SUPERLINK. Once the geometry is created with SCRIBEModeler, it is assigned materials using a program called PLINK, which also helps to define climate data for doing daylighting analysis. 2.2.2. Desktop Radiance Desktop Radiance is a CAD-based daylighting and electric lighting simulation tool with AutoCAD as a geometry modeler and Radiance as a simulation engine [15]. For linking AutoCAD with Radiance a special program called RADOUT was created [16]. It runs in the AutoCAD environment and facilitates the translation of CAD geometry into Radiance input files. Apart from having a CAD interface for creating geometry, Desktop Radiance also has a Radiance material library for assigning materials. RADOUT provides an interface to selectively export the geometry. Once the Radiance input file is created, different Radiance utilities are called using MS Windows BATCH scripting for doing a daylighting analysis [15]. In addition to AutoCAD, Radiance also has many utility programs that help in converting different geometry formats used by CAD-based tools into the Radiance format [16]. 2.2.3. DAYSIM/DIVA Another CAD-based modeling tool is Rhinoceros, a NURBS modeling program [17]. It has been linked to Radiance and DAYSIM through a plug-in called DIVA [18]. DIVA provides a graphical user interface to assign materials to different surfaces created in
Rhinoceros and export the models to Radiance and DAYSIM to perform the simulation. 2.2.4. Ecotect-Radiance-DAYSIM Ecotect is a widely used tool in the building simulation community for doing daylighting simulation [19]. Ecotect provides a CAD-based modeling environment for creating building geometry, assigning materials to geometry, and exporting the model to Radiance [20] and DAYSIM [21]. Fig. 1 shows linking of CAD-based tools with Radiance and DAYSIM. 2.2.5. Advantages and limitations of CAD-based daylighting simulation tools CAD-based daylighting simulation tools help in creating the geometry quickly. Without the aid of CAD interfaces it would be very time consuming to define geometry and material properties manually for daylighting simulations. However, some of the CAD-based tools have certain limitations in terms of creating geometry. These limitations are generally to require an idiosyncratic or specific method of defining geometry. For example in Desktop Radiance to create a window in a wall, the window glass has to be modeled as a single surface and the wall surface hosting the window has to be divided into four different surfaces. Similarly, in DIVA for Rhinoceros, windows have to be modeled as surfaces with zero thickness. The user has to adhere to these rules for creating daylighting simulation models. Also, if the building geometry has to be represented with thickness but the daylighting surfaces are modeled with zero thickness, then the complexity and time/effort for defining the model increase with the increase in the number of surfaces considerably. Furthermore, the materials have to be assigned by selecting individual surfaces, which is also a tedious task if the model is complex. 2.3. Building Information Modeling to daylighting Building Information Modeling (BIM) is an emerging technology in the AECO industry [22]. According to a public report there is a considerable increase in the use of BIM in AECO [2]. BIM tools not only help in designing a building using 3D graphics, but also are capable of holding non-graphical information such as material properties related to the building elements, which are not available in CAD tools. For example a wall can be modeled as an object with layers of materials using different properties. In addition, BIM tools help in modeling the geometry very quickly with a high degree of accuracy and detail compared to CAD tools by enforcing topology and connectivity among elements such as walls, windows, doors, roofs, and floors. Also, it is very easy to parametrically change the dimensions of the BIM objects and spaces with the geometry updating automatically based on any change. BIM is not only used to create building geometry, but also as a repository of the building information, which can be retrieved to perform different analyses on buildings, such as energy analysis, daylighting analysis, cost estimation, and structural analysis. 2.3.1. From BIM to CAD-based daylighting (Autodesk Revit to 3ds Max Design) 3ds Max Design is a product of Autodesk that is used mostly for animation and rendering. A module called Exposure in 3ds Max Design enables the user to do a physically based accurate daylighting simulation. The simulation engine is mental ray. Both time series rendering of High Dynamic Range (HDR) images and climate-based annual daylighting simulation can be conducted using Exposure. It has been validated comparing with measured data as well as with Radiance and DAYSIM daylighting simulating tools for a number of cases that extend from a simple window to complex fenestration systems consisting of shades, light shelves, and indoor venetian 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 tool, 3ds Max Design, is achieved through a file format called FBX [11]. The FBX file format contains the geometry, material information, and camera views necessary for the daylighting analysis in 3ds Max Design. However, in order to do a physically accurate daylighting simulation, a series of manipulations has to be performed initially such as replacing the lights in the Revit model with a daylighting system in 3ds MAX Design, and replacing the nonphysically based materials of Revit with physically based materials. Along with these changes, some of the changes have to be done to the Revit model, especially when modeling window glass panes. A detailed workflow of doing daylighting analysis using 3ds Max Design from Revit can be found in the literature [23]. Even though the process is fast compared with CAD-based daylighting simulation, it still requires manual checking and preparing the scene for accurate daylighting simulation. The process can be described as semi-automated. 2.3.2. ThermalOpt daylighting simulation ThermalOpt is a methodology for automated BIM-based simulation intended for use in multidisciplinary design optimization (MDO) environments [24]. It can do both thermal and daylighting simulations, supporting study of the effect of daylighting on the thermal performance. It uses EnergyPlus for thermal simulation and Radiance for daylighting simulation. ThermalOpt is implemented via four different plugins: (1) IFC2ThermalSim, (2) ThermalSim, (3) EnergyPlus Wrapper, and (4) Radiance Wrapper. In ThermalOpt, the building description is represented in IFC BIM interoperability format and translated into EnergyPlus input format. Geometry for daylighting simulation is taken from the EnergyPlus model. The EnergyPlus model does not represent the actual geometry of BIM. Taking geometry from EnergyPlus does not give accurate geometry for daylighting simulation. Even though the whole process is automated, the geometry for the daylighting simulation is not the same as the BIM model and this can alter the results. Learning from the shortcomings of the above studies, it was thought necessary to develop a process and methodology to accurately translate the BIM models’ geometry and material information into daylighting models (Radiance and DAYSIM). This was achieved by developing custom add-in programs that can take Revit BIM models and translate them into Radiance and DAYSIM input files, and then automatically perform the simulation without any
manual intervention. The main objective of the research project is to make the process quick and friendly to architects and engineers. The following sections will describe the whole process in detail. 2.3.3. Revit to Radiance and DAYSIM The present research is a part of a project on Physical Building Information Modeling (PBIM). The aim of the project is to integrate BIM with building performance analysis tools in multiple domains, including thermal and daylighting [25]. Even though there are several BIM authoring tools available in the market today, Autodesk Revit has been adopted for the present project because of its parametric BIM authoring capability and its wide acceptance in the industry. The main aim of the project is to link BIM with Radiance and DAYSIM by translating the Revit models into Radiance/DAYSIM input files and automatically performing the simulation. In a prior study [26] a Revit model is translated into a Radiance input file through other file formats and tools. Some of the major findings reported by the study are (1) translating Revit data into Radiance requires several steps to convert Revit geometry into Radiance geometry, (2) manual intervention is required in the whole process, and (3) certain information is required to be added manually to create a complete Radiance input file. In our project, we aim to develop a direct translator from Revit to Radiance and DAYSIM in order to significantly reduce the model preparation time and, in the meantime, preserve the parametric modeling advantages of Revit. 2.4. Radiance utility programs Several methods were explored to convert a Revit model to a Radiance input file. Radiance has several utility programs for converting a wide variety of CAD geometry data formats into the Radiance format. Some of the widely recognized data formats, which are used by different CAD-based tools, are 3DS [27] and DXF [28] developed by Autodesk, OBJ [27] developed by Wavefront and the MGF developed particularly for Radiance [29]. Autodesk offers a comprehensive exporting utility called FBX Converter that can convert the Revit file into the FBX format [30] and further into different geometry formats such as DWG, DXF, OBJ and 3DS. For the present project, to explore existing methods for converting BIM to daylighting simulation models, Revit models 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 [32], 3DS2MGF [33], and MGF2RAD [34] utility programs are used to convert these formats into the Radiance format. Fig. 2 shows different steps involved in converting a Revit model into a Radiance input file using these different utilities. The steps involved in translation using different utilities are represented as paths (i.e. P1, P2, P3,. . .). P6 is the translation path describing our method presented in this paper. The method proposed is a more direct path involving only one step for creating the Radiance input file and also performing the simulation. The main advantage of this method is that it is easy and does not require in-depth knowledge of different
tools, which are needed in other paths. Table 1 shows the number of steps involved in each method for converting a Revit model into Radiance/DAYSIM input files. 2.5. Issues involved in Revit to daylighting using Radiance utilities In all the existing approaches, geometry information could be exported automatically using the utilities but not material information. Manual assignment is needed for materials. This is a major setback for creating Radiance input files. Another major setback 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 exporting/translating
Importing to other programs
Assigning materials or Exporting to other formats
Exporting/Converting
Exporting/Converting
No. of steps
P1
Exported into DWG format
DWG format imported into 3DS Max
Materials assigned in 3ds Max
Exported into Wavefront OBJ format
5
P2
Exported to FBX format
Imported into Autodesk FBX converter utility
Exported into 3DS format *.3ds
Converted into MGF format using 3DS2MGF
Exported into Radiance format using OBJ2RAD utility Converted into Radiance format using MGF2RAD utility
P3
Exported to FBX format
Imported into Autodesk FBX converter utility
Exported into Wavefront OBJ format *.OBJ
P4
Exported to FBX format
Imported into Autodesk FBX converter utility
Exported into DXF format *.DXF
P5
Exported to FBX format
Imported into Autodesk 3ds Max
Materials assigned in 3ds Max
Converted into Radiance format using OBJ2RAD utility Converted into Radiance format using DXF2RAD utility Simulated in 3ds Max
P6
Translated into Radiance input file using our Revit2Radiance Add-in Program
5
4
4
4 1
S. Kota et al. / Energy and Buildings 81 (2014) 391–403
are modeled with thickness whereas in Radiance window panes are modeled with zero thickness – this means manual adjustment is needed for modeling the windows with the existing methods. From our experiments, some errors were identified when converting OBJ data into Radiance using OBJ2RAD utility. Manual inspection of the OBJ geometry file created by FBX Converter revealed some strange characters in the OBJ input file that caused errors in translating using the OBJ2RAD utility. We also observed that converting OBJ and DXF to RAD was a one-step process, but converting 3DS to RAD was a two-step process. During the process of converting 3DS to MGF using the 3DS2MGF utility, an error occurred resulting in the MGF file not being created. From our experiments, the only path that has successfully translated geometry from a Revit model to Radiance (without material definition) is DXF2RAD. Each of these methods has its own limitations. There are many steps involved in the process of translating the Revit model into the Radiance input file. In addition, a thorough knowledge about each tool is required to carry out the process, which can be a significant challenge for even an experienced architect. As the architect changes the design iteratively the whole process becomes a tedious task, as 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. Owing to these reasons, we proposed to translate the Revit models into Radiance input files directly and automatically (i.e., integrating the daylighting analysis into the Revit modeling environment), which does not involve human intervention and results in a process that is quick and user friendly. 3. Methodology – prototyping Our research method includes prototyping and validating the prototype with building models. For facilitating the data transfer from Revit to Radiance and DAYSIM, a prototype Revit2Radiance was developed. The prototype takes the building information, both geometric and non-geometric such as material properties, from a BIM model and translates it into Radiance and DAYSIM input files. The prototype was developed using Revit’s Application Programing Interface (API) with the C# programming language. Based on BIM (geometry, materials, location, date/time, a camera view, and sensor points), the prototype extracts the building surfaces in triangle meshes and other data needed to generate the input files of Radiance and DAYSIM. The main objective for developing the prototype is to translate the Revit models into Radiance/DAYSIM input files and perform the daylighting simulation and visualization without manual intervention in the translation process (after the BIM model is created and data are entered into BIM). This not only helps avoid errors due to manual input of the data in preparing simulation input files, but also makes the process of performing the analysis more user-friendly for the designers who are not well acquainted with the simulation tools. For the present prototype a total of four Revit add-in programs were developed that can automatically translate BIM models into simulation models, simulate the building performance, and generate analysis results. Table 2 shows the add-in programs and respective tasks they accomplish. Fig. 3 shows the overall workflow of Revit2Radiance. Fig. 4 shows the workflow with the add-in programs. In the subsequent sections each of the add-in programs is explained in detail.
395
Table 2 Revit add-in programs (Revit2Radiance modules) for translating Revit models into Radiance input files. Name of the add-in program
Action accomplished by the add-in
1
Parameter File Creation
2
Adding Material Parameter
3
Reading Material Information Model Translator
Creates a default shared parameter file in the project directory for adding shared parameters that are needed for defining building materials in Add-in #2 below Adds user-defined shared parameters for building materials that are necessary for holding materials data for Radiance models Reads values of the created shared material parameters Translates the Revit model’s geometry and material data into Radiance/DAYSIM input files and automatically run the simulation programs to generate the results
4
multiple parameters and values assigned to it. For example, one of the material parameters is “color” which gives the color property of the material in terms of the Red, Green, and Blue (RGB) channels. In Radiance there are different material types [16]. One of the basic material types is ‘Plastic’ which has a color component, diffuse and specular components, and a roughness factor. The color is defined by Red, Green and Blue reflectance values [16]. Given the color values in RGB of a surface it is possible to compute the reflectance of the material through a brightness formula. The Average Reflectance value is computed from the following formula for the standard Radiance R, G, and B components of the reflectance [16]: Average Reflectance = 0.265 × R + 0.670 × G + 0.065 × B
(1)
where R, G, and B are normalized values. In our prototype, we used the default Revit material R, G, and B values (in the range of 0–255), normalized them, and calculated the Average Reflectance using the above equation. The same Average Reflectance is used as the Radiance RGB values for calculating the illumination levels and rendering grayscale images by Radiance. The rendered grayscale images can help visualize the illumination levels without the effects of chromaticity. For rendering color images, Eq. (1) does not need to be called by Revit2Radiance and the normalized Revit RGB, instead of the Average Reflectance value, will be used as Radiance RGB. In addition, if the users intend to use measured RGB of a material instead of the default values in Revit, they can modify the values in Revit. However, Revit materials do not have specularity and roughness values that are necessary for complete Radiance material description. It is necessary that these values be first added to each material in the Revit material database. Revit offers the functionality for adding user-defined parameters called custom parameters. In the present project this functionality was used through API programming to add specularity and roughness parameters to each Revit material. In order to add user-defined parameters, a shared parameter file must be created first for the parameters. The first add-in 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 associated with the geometry. Materials are assigned to geometry through the Revit material database. Each material in Revit has
Once the default shared parameter file is created, the second add-in program “Adding Material Parameter” adds the two 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 Revit. 3.3. Reading Material Information As described briefly in the earlier section “Parameter File Creation”, Radiance materials have reflectance, specularity and roughness values. Reflectance is defined with individual channel reflectance (i.e. red-reflectance, blue-reflectance and greenreflectance) along with specularity and roughness. Specularity and roughness values must be added to the parameters created for each Revit material (i.e. adding appropriate values to the shared parameters for each material). In order to ascertain specularity and roughness values of Revit materials, the Revit materials were compared with similar material types for which Radiance material data are available from different databases. The specularity and roughness values of these Radiance materials were then adopted for Revit materials. A list was created in an Excel spreadsheet
containing Revit material names with their corresponding specularity and roughness values taken from Radiance materials. The “Reading Material Information” add-in program reads the Excel database file containing the specularity and roughness values and assigns these values to the custom parameters created for each material in the earlier steps. At this point each material in the Revit material database has all the necessary information that must be translated into a Radiance material description.
3.4. Radiance sky model and view description Radiance requires a sky, a view and a scene description. There are different ways a sky description can be modeled in Radiance. By providing the latitude, longitude, and time of the day, the Radiance utility Gensky [16] can create a sky description. In Revit, each view has sun setting parameters, which also provide access 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
the day (month, day, and hour) for shadow calculating and image rendering purposes. For the present prototype, this information is accessed, formatted, and provided to Gensky program to create the sky description by our prototype. Similarly, the camera view information is translated from BIM into a Radiance view description that is necessary to produce Radiance images that can be post-processed for different analyses. 3.5. Sensor point description In Radiance the simulation results can be displayed as images, numerical values, and illumination contour plots. Apart from producing a picture that has the photometric information for a given camera view and scene description, Radiance can also compute lighting levels at a specific location inside a given space by specifying a sensor point. The sensor point is specified by giving the X, Y and Z coordinates and view directions at which Radiance calculates the illumination levels. For the prototype, a sensor object was created as a custom Revit family that the user can place inside or outside of the building to compute the illumination levels. The prototype automatically translates the coordinates of the sensor locations into the sensor points coordinates for both Radiance and DAYSIM. 3.6. Weather file for DAYSIM DAYSIM for daylighting simulation uses a different sky description program called Gendaylit, which uses a weather file to generate a sky description for annual daylighting analysis based on the Perez all-weather sky model [8]. It uses the EnergyPlus weather data format, which is readily available for many different locations throughout the world. In the present project, the EnergyPlus weather file is formatted automatically into the DAYSIM weather format by the epw2wea program [35], which comes with DAYSIM installation. 3.7. Model Translator The main add-in program of the prototype is the Model Translator. It translates the project information, including geometry, materials, weather, location, date and time, and sensor point data into Radiance/DAYSIM input files. The following are the set of input files created and organized into a project folder structure created by Model Translator. • A Radiance material file, which has Radiance material description translated from the Revit materials. • Radiance geometry file, which has Radiance geometry description translated from the Revit model geometry. • A Radiance camera view file for generating Radiance images. • Radiance sky description files. • A weather file necessary for DAYSIM. • Two batch files, which have Radiance and DAYSIM commands to sequentially execute the Radiance and DAYSIM routines to produce the results. Model Translator also runs the batch files that automatically launch both the programs. Once the simulations are complete all the results are automatically written to a folder in the project directory. 3.8. Window glass pane translation One of the challenges encountered in translating the geometry of BIM models into Radiance is the translation of window glass panes. BIM represents the building geometry as surfaces or solids,
397
Table 3 Project and material description. Building parameters
Values
Location Length of the room Width of the room No. of windows Dimensions of windows Wall reflectance Floor reflectance Roof/ceiling reflectance Window transmittance
Denver, CO 6M 8M 2 3M×2M 0.5 0.2 0.8 0.75
which can have multiple surfaces depending on the shape of the geometry (Fig. 5A shows a solid of a window pane). The Model Translator add-in converts each surface of a solid into a Radiance geometry surface with material information attached to it. In Revit, window panes are mostly cuboids of 6 surfaces with two of the surfaces representing outside and inside surfaces of a single pane and other four representing the sides of the pane (Fig. 5B). Translating the window glass pane solid into 6 individual glass panes will lead to a wrong model in Radiance because Radiance would create internal reflection of light among the 6 surfaces of the glass pane. In Radiance the window should be represented as a single surface without any thickness (Fig. 5C). To represent the Revit window glass pane correctly with a Radiance window model (i.e., from Fig. 5A to C) only one surface of the Revit glass pane has to be translated. An algorithm was developed to translate only one surface (i.e. front surface of the glass pane) into a Radiance glass surface. 3.9. Running the prototype (Revit Add-in programs) In order to perform the daylighting simulation the Add-in programs need to be loaded into Revit and accessed under the “External 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 programs and the sequence (1–4, displayed alphabetically based on the program names in Add-In Manager) shows the order of execution of the programs to perform BIM-based daylighting analysis. Once the main program (4) is executed, the Revit model is translated into Radiance and DAYSIM input files, and Radiance and DAYSIM programs are launched automatically to perform the daylighting simulation (Fig. 7). 4. Test cases and validation We used two test cases to demonstrate the BIM to daylighting simulation process, the tool Revit2Radiance, the results, and validation. 4.1. Test case 1 and validation For testing the prototype, the BESTEST Case 600 building’s BIM model was created in Revit (Fig. 8). The reason for using this building for testing is that the same BIM model of the building can be used for testing the translations from BIM to both thermal and daylighting simulations, while the thermal simulation results can be validated with benchmark results [25]. Fig. 8 shows the 3D BIM model, the floor plan, the section, the windows’ locations, and the placement of the daylight sensors. Table 3 provides the project and material description, which are used for Radiance and DAYSIM simulations. The weather location used for the simulation is Denver, Colorado. The sky description is for summer solstice (June 21st) with the sun option turned on in the Gensky program. The time for 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 Radiance and DAYSIM programs. The results from the Radiance program are primarily images, which have photometric data. The results of the DAYSIM program are annual illumination data. The simulation results are then written into a result folder in the project directory. The Radiance image generating utility program takes geometry and materials files, the sky description file, and the view description to generate an image (Fig. 9A) that shows the illumination levels. Once the image is generated, subsequently three images are generated from the first image by post-processing: (1) the human sensitive image (Fig. 9B), (2) the iso lux contour plot (Fig. 9C), and (3) the false color image (Fig. 9D). Similar to Radiance, by taking the geometry, material, weather, and sensor point files, DAYSIM performs an annual illumination simulation at the specified sensor points (Fig. 10). An annual illumination profile file (*.ill) is then written into the project folder. For the present prototype Revit2Radiance, a validation study was conducted. The validation study was not aimed to find out
the accuracy of the simulation results since the simulation engines Radiance and DAYSIM are already validated in other research. Instead, the aim of the validation is to ascertain that the Revit model is accurately translated into the Radiance/DAYSIM model. A visual comparison was done to check the accuracy of the geometry translation. The Radiance model, which is generated by our Revit2Radiance, is visualized using the RVIEW program. The view generated by the RVIEW program exactly matches the Revit model showing the accuracy of the geometry translation (Fig. 11). Fig. 12 shows the Revit material properties for a material “Sash”. Revit2Radiance uses Revit RGB and computes the Average Reflectance value using Eq. (1). As described in Section 3.1, the reason for using the same Average Reflectance for all Radiance RGB components is to render a grayscale image in order to visualize only the illumination levels. (For rendering a color image, Eq. (1) will not be used by Revit2Radiance). Revit2Radiance also takes the custom parameter values “Specularity” and “Roughness” to write the Radiance 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 Fig. 13 shows the prototype output of Radiance material description. Both the prototype-generated Average Reflectance (0.73 for RGB components) and the manually calculated Average Reflectance
(0.73) match exactly. Also the Specularity” and “Roughness” values are passed from the Revit material to the Radiance material correctly. This is a sample comparison that validates the accuracy of 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 Manually calculated Average Reflectance for the material “Sash”. Color channels Value Normalized value (value/255) Color coefficients (from Eq. (1)) Normalized value × coefficients
Red 224 0.878 0.265 0.233
Green 178 0.698 0.670 0.468
Blue 126 0.494 0.065 0.032
Average Reflectance
0.73
4.2. Test case 2 and validation
to
The prototype was also tested on a complex case study check the accuracy in translating the geometry for
performing the daylighting simulation. The test case is Stanford University Solar Decathlon 2013 house project (http://solardecathlon.stanford.edu), the Revit model of which was provided to us by the project team. Fig. 14A shows the floor plan and 3D views of the house. The main hall of the house has windows 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 transmittance of windows. Both renderings and annual illumination profiles were produced using Revit2Radiance. The accuracy of translating the model was tested. Once the Radiance file is Table 5 Surface reflectance’s of different building element.
Fig. 13. Revit2Radiance prototype-generated Radiance material description for Revit material “Sash”.
Building element
Reflectance/Transmittance
Walls Floor Roof Clearstory window transmittance Curtain panel window transmittance Southern side window transmittance
0.5 0.2 0.8 0.8 0.7 0.7
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 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 and D) show the RVIEW views of the Radiance input file generated by Revit2Radiance. Upon visual inspection it can be concluded that the prototype is accurately translating the BIM model into the Radiance input file. Fig. 16(A–C) show the Radiance simulation results of the Solar Decathlon house.
5. Conclusion and future work The research paper presents a method to enable direct integration of BIM with daylighting simulation tools. The prototype directly creates Radiance and DAYSIM input files from the Revit models through automated steps with high efficiency and accuracy. In addition, the prototype is an easy-to-use tool for architects and designers without requiring in-depth knowledge of the simulation
S. Kota et al. / Energy and Buildings 81 (2014) 391–403
tools. The research has identified that BIM does not have all the information that is necessary for creating the simulation input files for Radiance and DAYSIM; however it provides options to incorporate the required information. The research has also identified that the representations of the same building elements are not the same between Revit and Radiance/DAYSIM. For example a glass pane, which is represented with a thickness in Revit, has to be represented as a surface without a thickness in Radiance and DAYSIM. The prototype has been validated for the accuracy in the geometry translation and proved to be accurate and reliable. With the parametric change capability of BIM, it will be easy for the architects and designers to use the tool to quickly study different options since it provides a seamless link between Revit and Radiance/DAYSIM. The integration of daylighting analysis into the BIM environment helps in making informed design decisions. Our method depends upon Revit and thus lacks the format neutrality of using IFC. Even though IFC data schema is being promoted and developed to facilitate the interoperability among different BIM-based programs, it is thought to be complex to program and implement in software [36]. Nonetheless, IFC is considered to play an important role in the translation from BIM to energy simulation [1]. The IFC approach has benefits compared with the approach adopted in this paper as it can facilitate the data transfer between different IFC compatible tools. The major limitation of the present method is that it works with Revit but not other BIM authoring tools. However, the concept and methodology can be applied to any BIM authoring tools that support API programming. An extension of the present work could be to study the effect of daylight on the electric lighting integrated into the BIM environment. Yan et al. [25] demonstrate the integration of thermal performance analysis into Revit using a Modelica-based thermal simulation engine. The next step could be integrating the results of these two performance simulations (thermal and daylighting) for multi-objective design optimization. As of now the prototype can only handle simple fenestration systems; future improvement could be to include the simulation of Complex Fenestration Systems (CFS) such as windows with blinds, fritted glass, and laser cut panels. Acknowledgements This material is based upon work supported by the National Science Foundation under Grant No. 0967446. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. We are grateful to the Stanford Solar Decathlon 2013 project team and their Construction Manager Rob Best and Project Manager Derek Ouyang for providing us with the BIM model of the project. We thank Jong Bum Kim and WoonSeong Jeong for providing us with the Revit model of the BESTEST Case 600 and their valuable input for the research project. References [1] V. Bazjanac, IFC BIM-based methodology for semi-automated building energy performance simulation, in: Presented at the CIB-W78 25th International Conference of Information Technology in Construction, Santiago, Chili, 2008. [2] N. Young, S. Jones, H.M. Bernstein, Building Information Modeling (BIM)Transforming Design and Construction to Achieve Greater Industry Productivity. SmartMarket Report, 48, 2008. [3] S. Kota, J.S. Haberl, Historical survey of daylighting calculations methods and their use in energy performance simulations, in: Paper presented at 9th International Conference for Enhanced Building Operations, Austin, Texas, November 17–19, 2009. [4] G.J. Ward, The RADIANCE lighting simulation and rendering system, in: Proceedings of the 21st Annual Conference on Computer Graphics and Interactive Techniques, ACM, 1994 July, pp. 459–472.
403
[5] G.J. Ward, F.M. Rubinstein, A new technique for computer simulation of illuminated spaces, Journal of the Illuminating Engineering Society 17 (1) (1988) 80–91. [6] J. Mardaljevic, Validation of a lighting simulation program under real sky conditions, Lighting Research and Technology 27 (4) (1995) 181–188. [7] C.F. Reinhart, M. Andersen, Development and validation of a Radiance model for a translucent panel, Energy and Buildings 38 (7) (2006) 890– 904. [8] C.F. Reinhart, O. Walkenhorst, Validation of dynamic RADIANCE-based daylight simulations for a test office with external blinds, Energy and Buildings 33 (7) (2001) 683–697. [9] J. Mardaljevic, Simulation of annual daylighting profiles for internal illuminance, Lighting Research and Technology 32 (3) (2000) 111–118. [10] C. Reinhart, P.F. Breton, Experimental validation of Autodesk® 3ds Max® Design 2009 and Daysim 3.0, Leukos 6 (1) (2009) 7–35. [11] C. Reinhart, M. Landry, P.F. Breton, Daylight simulation in 3ds Max Design 2009 – Getting Started. Autodesk White Paper, 2009. [12] C. Reinhart, A. Fitz, Findings from a survey on the current use of daylight simulations in building design, Energy and Buildings 38 (7) (2006) 824– 835. [13] H. Erhorn, J. De Boer, M. Dirksmöller, ADELINE – an integrated approach to lighting simulation, Proceedings of IBPSA Building Simulation’97 1 (1997) 79– 85. [14] LBL, SUPERLITE 2.0 User Manual. Building Technology Program, Lawrence Berkeley Laboratory, Berkeley, CA, 1989. [15] LBNL, Desktop Radiance 2.0 BETA User Manual, Environmental Energy Technologies Division, Building Technologies Department, Lawrence Berkeley National Laboratory, Berkeley, CA, 2000. [16] G.W. Larson, R. Shakespeare, C. Ehrlich, J. Mardaljevic, E. Phillips, P. ApianBennewitz, Rendering With Radiance: The Art and Science of Lighting Visualization, Morgan Kaufmann, San Francisco, CA, 1998. [17] R. McNeel, Rhinoceros Version 5.0, Service Release 8 [Software], 2014, Available from http://www.rhino3d.com [18] K. Lagios, A. Jakubiec, J. Niemasz, C. Reinhart, Quick Start Guide for the DIVAFOR-RHINO Toolbar, 2010, Retrieved from http://www.diva-for-rhino.com/ DIVA-for-Rhino QuickStartGuide.pdf [19] A.J. Marsh, Performance Analysis and Conceptual Design (Ph.D. dissertation), The University of Western Australia, Perth, Australia, 1997. [20] A. Marsh, The Interaction of Radiance and ECOTECT, in: Presented at 5th International Radiance Scientific Workshop, De Montfort University, Leicester, UK, 2006. [21] C. Reinhart, Tutorial on the Use of Daysim Simulations for Sustainable Design, 2010, Retrieved from http://www.daysim.com/pub/Daysim3.0.Tutorial.pdf [22] F. Farias, Contemporary Strategies for Sustainable Design. Doctoral dissertation (Ph.D. dissertation), Texas A&M University, 2013, Retrieved from http://hdl.handle.net/1969.1/149379 [23] M. Landry, P.F. Breton, Daylight simulation in 3ds Max Design 2009 – Advanced Concepts. Autodesk White Paper, 2009. [24] B. Welle, J. Haymaker, Z. Rogers, ThermalOpt: a methodology for automated BIM-based multidisciplinary thermal simulation for use in optimization environments, Building Simulation 4 (4) (2011) 293–313. [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 Performance Simulation Association (IBPSA), August 25–28, Chambery, France, 2013. [26] J. McGrew, Revit to Radiance (via an expensive translator. . .), 2006, Retrieved from http://becausewecan.org/node/190 [27] K. McHenry, P. Bajcsy, An Overview of 3d Data Content, File Formats and Viewers, National Center for Supercomputing Applications, 2008, pp. 1205. [28] S. Kumar, Interoperability between building information models (BIM) and energy analysis programs, 2008 (Master’s thesis): Available through ProQuest Dissertations and Theses database. [29] G. Ward, The Material and Geometry Format, 1996, Retrieved from http://radsite.lbl.gov/mgf/mgfhtml/intro.html [30] B.S. Russell, Autodesk FBX-Based Revit Architecture 2009 to Autodesk 3ds Max Design 2009 Workflow: Autodesk White Paper, 2009. [31] G. Mischler, dxf2rad 0.9 Convert Geometry from DXF to Radiance, 2000, 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/ man html/obj2rad.1.html (accessed 12.06.14). [33] S. Anger, J. Bowermaster, G. Ward, 3ds2mgf – convert 3D Studio binary file to Materials and Geometry Format, 2014, Retrieved from http:// radsite.lbl.gov/mgf/mgfhtml/3ds2mgf.html (accessed 12.06.14). [34] G. Ward, mgf2rad – convert Materials and Geometry Format file to RADIANCE description, 2014, Retrieved from http://radsite.lbl.gov/radiance/man html/mgf2rad.1.html (accessed 12.06.14). [35] C. Reinhart, DAYSIM Subprogram: epw2wea, 2014, Retrieved from http://daysim.ning.com/subprogram/epw2wea (accessed 12.06.14). [36] B. Dong, K.P. Lam, Y.C. Huang, G.M. Dobbs, A comparative study of the IFC and gbXML informational infrastructures for data exchange in computational design support environments, Building Simulation (2007) 1530–1537.