Integrating building information modeling and virtual reality development engines for building indoor lighting design

Using BIM-based simulations are increasingly being used to assess the success or failure of indoor lighting design. Lighting simulation plugins for BIM applications help users to analyze design options in order to provide sufficient lighting for occupants and improve building energy efficiency and overall building performance (Aksamija & Zaki, 2010). Two types of lighting simulation plugins are compatible with the BIM: external plugins and internal plugins (e.g., Lighting Analysis in Revit, Lighting Assistant in 3ds Max, Radiance, Daysim, Diva, and DesignBuilder). Internal plugins or internal extensions can be added to BIM applications, e.g., Revit, to enable the highest degree of interoperability (Nasyrov et al., 2014). External plugins are needed for data exchange across platforms to perform simulations and visualizations. Generating quantitative results is the main use of such plugins. However, qualitative results of lighting design, which refers to the illumination of an entire scene and has complex characteristics, such as aesthetics, are difficult to quantify (Sorger et al., 2016).

A number of previous studies proposed methods for using lighting analysis tools with the BIM model. For example, Özener et al. (Özener et al., 2010) developed BIM-based daylight-ing simulations and analyses for educational purposes. Such tools help students to understand the impact of design decisions on the performance of daylight and artificial lights in different design options. Daysim, Radiance, and Ecotect were used in their experimentation. In addition, Yan et al. (Yan et al., 2013) and Kota et al. (Kota et al., 2014) developed a system to connect the BIM database and building energy modeling (BEM) to support thermal simulations and daylighting simulations using an application programming interface (API) called Physical BIM (PBIM). Stavrakantonaki et al. (Stavrakantonaki, 2013) proposed a framework for selecting different BIM tools for daylighting assessment. DesignBuilder, Diva, and Ecotect were used in their experiments. However, Hu et al. (Hu & Olbina, 2011) argued that limitation of traditional lighting simulation tools, e.g., Radiance, is that the setup process is time-consuming and such tools are not easy to use in real-time simulations.

BIM has created an opportunity for stakeholders to collaborate and share multidisciplinary building information (Eastman et al., 2008) that can help designers to save time when creating new models (Ham & Golparvar-Fard, 2014). Integration with computer gaming environments is a new capability of BIM technology that allows users to better observe facilities and experience their surrounding environment before it is constructed (Figueres-Munoz & Merschbrock, 2015). Coupling of a game engine and the BIM can create a highly interactive environment with a digital geometry model derived from the BIM software. Figueres-Munoz et al. (Shiratuddin & Thabet, 2011) indicated that integration of BIM and gaming technology has the potential for combinatorial innovation. A number of previous studies have proposed methods for integrating game engine, BIM, and VR experience for various purposes, e.g., design review (e.g., (Shiratuddin & Thabet, 2011; Wu, 2015)), construction management (e.g., (Bille et al., 2014)), education (e.g., (Wu, 2015; Goedert et al., 2011)), energy conservation (e.g., (Niu et al., 2015; Jalaei & Jrade, 2014; Brewer et al., 2013)), real-time architectural visualization (e.g., (Yan et al., 2011)), and design feedback (e.g., (Hosokawa et al., 2016; Fukuda et al., 2015; Bahar et al., 2013; Motamedi et al., 2016)).

A few studies have focused on using game engines and VR for lighting design visualization. For example, Gröhn et al. (Gröhn et al., 2001) presented a method for visualizing indoor climate and visual comfort in a 3D VR environment. In their method, photo-realistic visualization was used to visualize the lighting distribution on the surfaces of the architectural model. However, they did not include real-time control of lighting conditions. In addition, Sik-lányi et al. (Sik-lányi, 2009) studied how using different types of lights influences the development of realistic VR scenes. Sampaio et al. (Sampaio et al., 2010) developed a virtual interactive model as a tool to support building management by focusing on the lighting system. Santos et al. (Santos & Derani, 2003) developed an immersive VR system for interior and lighting design using the cave automatic virtual environment (CAVE). This system provided users with an interface to change the design elements of the building, e.g., walls, floor, furniture, and decoration, in a virtual environment. However, their system focused only on the appearance of lighting design using ray tracing as the rendering algorithm, which did not provide a function to simultaneously generate quantitative outputs of lighting design, e.g., illuminance level and energy consumption feedback.

Methods for fully integrating the BIM database, game engine, and VR for real-time lighting design and energy consumption feedback have not yet been proposed, and this topic remains an area of research. Therefore, the present paper investigates a new method that uses BIM, game engines, and VR to facilitate lighting design and lighting energy performance analysis via an immersive and interactive user experience.

Lighting plays an important role in indoor environmental conditions. Lighting performance depends on several factors, such as energy cost, energy savings, illuminance, and energy consumption (Deru et al., 2005). Designers want to achieve an optimal design with minimum energy consumption. The design of daylighting and artificial light rely on a combination of specific scientific principles, established standards and conventions, and aesthetic (Benya et al., 2001). The following five physical parameters that influence occupants’ visual comfort in architectural spaces should be considered: lighting illuminance levels, lighting distribution (diffusion), correlated color temperature (CCT), brightness ratio, and glare (Fielder, 2001), (Descottes & Ramos, 2011). Designers must ensure that the visual comfort parameters comply with the standard of lighting design. Regarding human visual perception, the most important variables are CCT and illuminance level (Shamsul et al., 2013). Visual perception is defined as an interpretation capability of visual sensation when people visually perceive their surrounding environment. As mentioned above, illumination level and CCT are two variables that have a significant influence on human perception. The illuminance value is an indicator for verifying the quantity of light in a given environment. International lighting standards and building regulations specify the level of illuminance required for providing visual comfort to facilitate human visual performance. Correlated color temperature describes the characteristics of light colors. Color temperature plays a particularly important role in both the physical properties of light and the physiological and psychological response of humans when light enters the eyes (Descottes & Ramos, 2011; Shamsul et al., 2013).

VR experience is based on users’ perception of the virtual world (Mihelj et al., 2013). VR technology enables the creation of realistic models and makes scenes look real (Stahre & Billger, 2006). Immersion (perception) and real-time interaction (action) are two properties of VR. VR allows users to experience an artificial environment through human senses (Souha et al., 2005). Head mounted displays (HMDs) have been used as 3D immersive tools to provide an experience that is close to the human perception of reality (Ciribini et al., 2014; Scarfe & Glennerster, 2015).

In the game environment, lighting plays a significant role in making scenes look realistic. Game engine technology has lighting features that resemble real-world lightings (Shiratuddin & Thabet, 2011). Lighting simulation for gaming and VR presentations relies on a mixture of rendering algorithms, including path tracing and photon mapping. It is possible to effectively and physically simulate indoor illuminated environments at a level of quality where concepts of photorealism and perceptual accuracy can be discussed and measured (Murdoch et al., 2015). Lighting simulation in game engines can produce realistic scenes based on the inverse square law, which states that illumination intensity varies directly with luminous intensity on a surface and is inversely proportional to the square of the distance between the light source and the lighted surface (IESNA, 2000; de Rousiers & Lagarde, 2014). Other lighting metrics, such as luminous flux and light color, are also provided in the game engine. Luminous flux (light output or intensity), measured in lumens (lm), determines the quantity of light emitted from the lamp to illuminate the entire scene. Light color is expressed as color temperature (kelvins or the SI abbreviation “K”). Another parameter for measuring light is luminance, measured in footlamberts (fL) or candelas/m² (cd/m²). Luminance refers to perceive brightness which describes the amount of light delivered into space and reflecting off of a surface in a space that affects human ability to see (The Energy and Resources Institute (TERI), 2004). Luminance measurement affects the measurement of exterior/interior design and finishes in reflecting light rather than measuring lighting quality. Luminance is a useful baseline metric to identify sources of glare within a person’s field of view.

The advanced dynamic lighting feature in game engines allows pre-visualization of various designs. Consequently, lighting features in game engines replace conventional methods and enable faster lighting simulation when changing parameters. Moreover, it is possible for players to perceive the characteristics of light in order to determine which design pattern is more appropriate for the users.

A number of previous research has studied visual perception dimensions of VR technology. Menshikova et al. (Menshikova et al., 2012) studied the Simultaneous Lightness Contrast (SLC) illusion and the results of their study showed that VR technologies may be effectively used in studies of lightness perception. Their study showed that the VR technology enables reproducing visual illusions in depth and to construct complex 3D scenes with controlled parameters to create articulated effects. Scarfe et al. (Scarfe & Glennerster, 2015) stated that immersive VR opens up new possibilities for studying visual perception in much more natural conditions.

The quality of an HMD display screen depends on its resolution as well as luminance and contrast ratio of the screen, field of view (FOV), exit pupil, eye relief, and overlap (Vince, 2004). The contrast is the ratio between higher luminance and the lower luminance that defines the feature to be detected, a value of 100:1 is typical. Luminance is the amount of light energy emitted or reflected from a screen in a specific direction (Fuchs, 2017) and a measure of a screen’s brightness, and ideally should exceed 1000 cd/m² (Vince, 2004). The luminance that can be observed in nature covers a range of 10⁻⁶ to 10⁶ cd/m² and the human visual system can accommodate a range of about 10⁴ cd/m² in a single glance (Fuchs, 2017). The FOV is a measure of the horizontal and vertical angle of view. The horizontal FOV of HMDs is restricted to about 100–120 degrees for each eye (Fuchs, 2017).

VR has a great potential to become a usable design tool for the planning of lighting design in buildings (Stahre & Billger, 2006). However, due to the technological limitations, there are still problems in representing realistic lighting condition. A virtual world may produce semi-realistic lighting phenomena, however, precisely perceiving illuminance and glare of scenes in HMD is impossible with the current technology. The FOV of a typical head based-display is limited and the resolution can vary from the relatively few pixels offered in the early system to very high resolution (Sherman & Craig, 2003). The screen of VR headset should display a huge number of pixels about 6000 pixels horizontally and 8400 pixels vertically, which are not compatible with the current technology (Fuchs, 2017). Although the visual perception is still different from the reality using the current technology, there are many visual details in the luminous environment that cannot be expressed through numerical information, such as misaligned luminaries and disturbing light-shade patterns (Inanici, 2007).

Glare refers to a physical discomfort and is caused by non-uniform luminance distribution within the visual field (Fasi & Budaiwi, 2015). Glare occurs in two ways; excessive brightness and the excessive range of luminance in a visual environment (IESNA, 2000). Glare is typically divided into three categories: 1) disability glare, which is caused by excessive brightness areas in the field of view of a person with greater luminance than that of which eyes can adapt to; 2) veiling glare, which is caused by reflections on specular or diffusive materials that might reduce the contrast and visibility of the task; 3) discomfort glare, a psychological sensation that does not necessarily impair vision in short term and can remain unnoticed to observers, it can cause annoyance, headaches or eyestrain after long exposure (Abraham, 2017). Discomfort glare is caused by luminance that are high relative to the average luminance in the field of view (IESNA, 2000). A glare index is a numerical evaluation of the acceptability of the presence of glare using a mathematical formula.

Several principal indices of discomfort glare have been proposed to quantify discomfort glare generated from small lighting sources (e.g., artificial lighting) and large luminous sources (e.g., windows) (Abraham, 2017). For discomfort generated by small sources, examples of discomfort indices include visual comfort probability (VCP) (Guth, 1963), unified glare rating (UGR) (Sorensen, 1987), and CIE glare index (CGI) (Einhorn, 1973). Example of discomfort indices generated by large luminous sources includes daylight glare index (DGI) (Chauvel et al., 1982) and daylight glare probability (DGP) (Wienold & Christoffersen, 2006). VCP is a rating that evaluates lighting systems in term of the percentages of people that will find a given discomfort glare (IESNA, 2000). VCP is defined as a number that corresponds to the percentage on a scale of 0–100. UGR is used to measure glare possibility in a given environment. UGR is the logarithm of glare from electric light sources (Boyce, 2003). UGR value ranging from 10 for low levels of discomfort glare to 28 for high levels of discomfort glare. DGI index was developed to predict glare from large area sources, such as a window. The DGI uses categorical rating to explain quantitative value between 16 for just noticeable to 28 for intolerable glare (Hirning et al., 2014). DGP is a metric for an estimate of the appearance of discomfort glare in daylit spaces. DGP considers the vertical illumination at the eye level as well as on the glare source luminance, its solid angle and its position index (Wienold & Christoffersen, 2006).

Recently, most architect and lighting designers prefer not to be guided by glare indices (Hirning et al., 2014). However, preventing glare is important for design consideration (Reinhart et al., 2012) such as work safety. Various studies have investigated that there are other potential subfactors which influence the large variations in individual glare sensitivity, such as age of the person (Reinhart & Wienold, 2011; Van den Berg et al., 2009), gender (Wolska & Sawicki, 2014), culture (Amirkhani et al., 2016), and physical differences (Pulpitlova & Detkova, 1993). Although various metrics have been developed attempting to quantify glare, a major obstacle in quantifying discomfort glare is the difficulty in analyzing complex lighting distributions due to the dynamic nature of daylight (Hirning et al., 2014). Every equation evaluate glare differently, and the evaluation of discomfort glare indices in practice is complicated (Stone & Harker, 1973; Clear, 2012). However, in the advanced lighting research, the assessment of discomfort glare in spaces increasingly involves the use of simulated high dynamic range (HDR) images (Boyce, 2003; Giraldo Vásquez et al., 2016).

A number of studies have focused on using HDR images for discomfort glare analysis from daylight, e.g., (Clear, 2012; Suk et al., 2017; Mcneil & Burrell, 2016). Existing glare assessment studies were typically based on real-world building. A few studies have proposed methods to simulate light illuminance and glare in a virtual environment, e.g., (Mangkuto et al., 2014; Inanici & Mehlika, 2003). HDR images store luminance data on a pixel scale, which enable the possibility of detecting glare sources to compute discomfort glare metrics. With the digital evaluation of HDR, it is possible to integrate the HDR rendering scenes generated by a game engine with glare evaluation tools, such as Evalglare, Photolux, Findglare, to improve the lighting design and discomfort glare measurement during the design process.

The prototype system supports interactivity with which users can experiment and adjust lighting design parameters in real-time. The following interactions are supported: (a) First-person movement: to control the movement of users’ avatars as they walk in the space; (b) User interface control: to customize parameters such as time (to observe the dynamic of sunlight), lighting fixtures, types, orientation, intensity, and color temperature of the bulbs, as well as the location of furniture. Unreal Motion Graphics (UMG), a visual user interface in Unreal Engine, is used to create various GUI elements in the BLDF system. The GUI interface of the proposed system is shown in Fig. 2. The menus of the game environment enable players to change parameters and freely navigate among game objects using input controllers. The following widgets are created, as shown in Fig. 2: (a) Plan view: to show the layout of furniture and the position of the user. Users can switch between false-color view and realistic views; (b) Shading devices menu: to change the type of shading for windows; (c) Material types menu: to change the materials of indoor surfaces; (d) Time slider: to set time for adjusting the daylight; (e, f) Light switches: to turn on/off individual artificial lights; (g) Lighting intensity menu: to change the intensity of light sources; (h) Color temperature control: to change the color of lights; (i) Lighting fixture types menu: to choose the type of lighting fixtures; (j) Moving and rotation tools: to move and rotate the light source, lighting fixtures, and furniture; (k) Lighting illuminance legend: to help measure the illuminance level (for false-color views); (l) Total energy usage: to display the total lighting energy consumption of the room; (m) Compass: to show the orientation of users when moving; (n) Lighting occupancy sensor: to switch lights on/off when the space is occupied and unoccupied; (o) Task light: to add individual light controls to each workstation.

The global illumination algorithm in the Unreal Engine is used to automatically produce lighting illumination scenes and calculate detailed shadows at runtime. The scenes can also be visualized in false-color, which shows the illuminances at different levels in the scene. Red indicates an illuminance level that exceeds than 1000 lx, and dark blue indicates an illuminance level of less than 100 lx. The user can hide/unhide each widget on the viewport. A keyboard and mouse are used as input devices to change design parameters in the proposed system. The designer can change the light intensity (measured in lumens) and the color temperature of a light (measured in Kelvins) using associated widgets on the interface (Figs. 2g and h). Additional lighting equipment items are modeled in the BIM application and are added to the game environment as an alternative equipment repository. However, it is possible to import libraries of lighting elements from the BIM application into the game engine. After adding items to the equipment repository, new types of lighting equipment are shown on the interface (Fig. 2i). The material and texture of indoor envelopes, e.g., the floor, walls, and ceiling, can be modified using the developed widget (Fig. 2c).

Ambient lights can be switched on/off using the light switch panel (Fig. 2e). The system provides the ability to set up lighting zones when lighting control per zone is required. An additional option for lighting control is the lighting occupancy sensor, which automatically controls ambient lights and task lights (Figs. 2n and o) in the BLDF system. In order to implement occupancy sensing, the Box Trigger tool and visual scripting of Unreal Engine are used. The lights will automatically turn on when the user walks into a specific zone and automatically turn off when the specific zone is vacated.

The total energy consumption of lighting equipment is shown in the energy consumption bar. A simple lighting energy calculation formula is used to calculate lighting energy in watts (number of fixtures × light power (W) = total consumption). The system also provides an interface to enter the time and date for adjusting the daylight duration (Fig. 2d). In order to simulate sunlight, directional light is added as an actor in the game environment, and visual scripting is used to automatically control the yaw and pitch of the sun and to control the position of the sun based on the time and the season.

Alternatives for shading devices, e.g., horizontal panel, horizontal multiple blades, vertical fin, slanted vertical fin, and egg crate (Fig. 2b), are provided in the system. Shading device models are created in BIM and are imported into the game environment as an alternative equipment repository. Users can compare the influence of shading devices on indoor lighting. The illuminance range legend (Fig. 2k) is used to identify the illuminance values, and the compass (Fig. 2m) helps to identify the orientation. Fig. 2a shows a false-color visualization, which helps designers to preview the amount of illuminance. The sample visualizations in Fig. 3 show the results of various simulations when parameters are changed using the developed widgets.

The usage process flow of the BLDF system (as shown in Fig. 4) is as follows: (a) The BIM model is created in Autodesk Revit, in which the geometry and material properties of building elements are modeled; (b) A static mesh is created after exporting the FBX file from Revit to 3ds Max; (c) A 3D geometry in the game environment is automatically created after importing the FBX file of the BIM model into Unreal Engine. Due to interoperability issues between the current versions of applications, although the geometry information of the building elements and lighting equipment are successfully imported into Unreal Engine, some of their properties, such as color/texture and lighting properties, are not transferred; (d) Light bulbs and their properties (e.g., intensity and color temperature) are manually added to the game engine. In addition, the orientation, the scale of the building, and the color and textures of interior envelopes are manually configured in the game environment; (e) Lighting parameters, such as intensity, color temperature, position of lighting equipment, and interior materials, are configured using the GUI (as explained in Subsection 4.1); (f) The simulation is executed, which enables users to immediately view quantitative results (e.g., energy consumption) and visualizations; (g) The user analyzes the lighting design outputs; (h) If the results are satisfactory, the game parameters are saved and a text file is generated; (i) The final design information parameters, such as lighting intensities, color temperatures, and positions of fixtures and bulbs are updated in the BIM using visual programming in Dynamo (explained in Subsection 4.3); (j) The BIM model is saved with the updated parameters. The user can set new parameters and run simulations until a satisfactory design is achieved. Realistic scenes with walk-through capability support qualitative assessment and the false-color scenes support quantitative assessment. The system is used in design meetings to facilitate discussions between designers and clients.

Dynamo (Dynamo, 2016) visual programming is used to update new design parameters in the BIM application. Figure 5 shows the process flow of updating the the design parameters in BIM. After finalizing parameters, such as intensity, color temperature, and fixture position for each individual light bulb, and the positions of furniture and shading devices, and the material types of interior envelopes, the new parameters are saved in a text file (Fig. 5a). The developed visual script (Fig. 5b) automatically updates the lighting properties in Autodesk Revit (Fig. 5c) using unique IDs of model objects.

The applicability of the BLDF system is validated in a design assessment and renovation project. The case study was performed in an office at Osaka University, Japan. Figure 9 shows 3D BIM model and 2D plan of existing fixtures position of the case study room. The BIM model of the room was created using Autodesk Revit Architecture 2015 (Fig. 9a). The BIM model contains details of the lighting system, and geometric and non-geometric information of components, e.g., building envelops and furniture. Lighting properties, such as the intensity and the color temperature, are configured based on the specifications of the lamps. Figure 9b shows a 2D plan of existing lighting fixtures. The room has 16 lighting fixtures with 32 tubular fluorescent (T8) lamps of 32 W.

The goal is to facilitate lighting design by allowing users to interact with the design and to perceive and experience the effects of the modifications simulated by the system. In order to integrate BIM with a game engine, the geometry information of building elements with their reference IDs are transferred from the BIM application (i.e., Revit) to the game engine (i.e., Unreal Engine). The BIM geometry data are transformed to a static mesh using 3ds Max. Figure 10 shows the process of transferring the model, in which the FBX file format is used to export the model from Revit into Unreal Engine. Regardless of the data export format, some important information, e.g., material textures, color temperature, and light intensity, is lost while exporting data to the game engine. The lost features must be redefined in the game environment.

In the experiment, users were invited to use the BLDF system with an HMD (Oculus Rift Kit 2, OLED display, 1920 × 1080 pixels per eye with 100^° nominal field of view) to visualize the design through a first-person perspective. In the first-person perspective with an HMD, a mouse is used to interact with game objects and a keyboard is used to navigate in the virtual environment. The user can move the game objects and change the lighting design properties by clicking on the GUI buttons. The goal is to reduce the energy consumption while providing adequate illumination on desks in an aesthetically pleasant environment. In order to achieve this goal, the following six test cases were designed. The test cases are based on three fundamental aspects of lighting design of buildings: (1) evaluating the illuminance level, (2) calculating the energy consumption of the lighting system, and (3) visualizing lighting appearances.