Practical Applications of the Large-Eddy Simulation Technique for Wind Environment Assessment around New National Stadium, Japan (Tokyo Olympic Stadium)

In this study, we developed a new computational fluid dynamics (CFD) model called Airflow Analyst that deepens the affinity between CFD and geographic information system (GIS). First, a precise simulation of the surface-mounted cube was conducted. Validation testing based on the obtained data confirmed the predictive accuracy of Airflow Analyst. Second, New National Stadium Japan (Tokyo Olympic Stadium) was accurately reproduced in a computer, capturing the latest detailed urban area data for the base. For the target of the constructed 3D models, simulations with a large number of grid points/cells (CFD) were conducted. These simulations reproduced the complex turbulent flow fields both inside and outside the stadium. The experiment successfully reproduced the CFD simulation using a large number of grid points/cells, where the conditions of the wind flow ventilation from the sky were similar to those of the intended stadium design.


Introduction
Urban environments in our society are affected by the various impacts of wind, such as strong wind damage and ventilation. Thus, the built environment in urban design and regional planning needs to be assessed and optimized. Every step during the examination process for plans requires the involvement of citizens, administrators, and developers. Our research team was engaged in software de-velopment for urban planning and architecture design, focusing on new computational fluid dynamics (CFD) [1] with a high affinity for geographic information system (GIS) [2]. Recent research results in the fields we are interested in have been reviewed in the literature [1]. We have been developing Airflow Analyst [3] [4] [5] [6], which is extension software for ArcGIS, which is versatile GIS software. The biggest advantage of Airflow Analyst is that users are able to analyze the flow field characteristics or passive scalar transport and diffusion in regions with arbitrarily shaped objects without considering landform features or building clusters.
Airflow Analyst, running on the GIS system, has the following advantages: 1) Large labor savings in the preparation of three-dimensional (3D) data for urban areas due to its usage of geospatial information resources, which are circulated free-of-charge or sold; 2) Large-scale reductions in constructing 3D data planning due to the integration of computer-aided design (CAD) and building information modeling (BIM); 3) Instant simulation with intuitive operation due to the visual confirmation of scales on maps for computational wind directions, domains, and grid generation; 4) Visualizing the computational results of 3D data on a map and storing geographic information-spatial references by coordinates, thereby enabling superpositions of spatial analysis with other spatial information; and 5) The availability to distribute and share using superpositions with web mapping services.
Conventional wind-flow simulations for urban areas require specialized knowledge and skills, as well as enormous amounts of time and labor to create 3D models and generate grids integrating local terrain], building shapes, and design plans. To reflect wind-flow simulations in a plan or a design, however, a system with simple and intuitive operation is required to examine data immediately after being obtained. This system should reduce the work involved in the above CFD analysis, and planners should be able to appropriately perform numerical simulations.
Governments and open data communities, both inside and outside Japan, are rapidly and increasingly providing free 3D data management. Therefore, geospatial information infrastructures have been established with 3D urban data. Specifically, BIM and Construction Information Modeling (CIM) are employed, with enhanced data compatibility with GIS. BIM and CIM have also enabled 3D data usage in design planning. Along with this, advanced techniques have been developed to extract data from satellite and aerial photographs, as well as photogrammetric image data acquired by unmanned aerial vehicles (UAVs). A new data creation method has been proposed: point cloud data obtained via laser surveying can generate minute 3D data, and all these data can be analyzed via integration with the GIS.
Airflow Analyst is the first software program in the world to complete the process from grid generation to the visualization of data computation results.

Summary of the Airflow Analyst Software
For the numerical simulations, we used the Airflow Analyst software package, for which a collocated grid in a general curvilinear coordinate system was adopted. In this collocated grid, the velocity components and pressure are defined at the grid cell centers, and variables that result from multiplying the con- For the equations governing the flow, a filtered continuity equation for incompressible fluid (Equation (1)) and a filtered Navier-Stokes equation (Equation (2)) are used. When passive scalar transport and diffusion were considered, the standard convection-diffusion equation for a passive scalar was solved with the linkage of the above equations of flow field characteristics.
Since we investigated wind fields with mean wind speeds of 5 -10 m/s, the effects of the vertical thermal stratification of the atmosphere (atmospheric stability) were negligible. For the computational algorithm, a method similar to a fractional step (FS) method [8] was used, and a time marching method based on the Euler explicit method was adopted. Poisson's equation for pressure was solved using the successive over-relaxation (SOR) method. To discretize all the spatial terms except for the convective term in Equation (2), a second-order central difference scheme was applied. For the convective term, a third-order upwind difference scheme was applied. The interpolation technique by Kajishima [9] was used for the fourth-order central differencing that appears in the discretized form of the convective term. For the weighting of the numerical diffusion term in the convective term discretized by third-order upwind differencing, α = 0.5 was used instead of α = 3.0 from the Kawamura-Kuwahara scheme

Validation Testing of the Prediction Accuracy of the Airflow Analyst Software
Here, the results of validation testing are reported for the prediction accuracy of the airflow with a surface-mounted cube. Qualitative and quantitative assessments were conducted using flow visualization and turbulent flow measurement with a split-film anemometer. Figure 1 shows a bird's eye view of the thermally stratified wind tunnel used for the present study. Figure 2 shows a side view of the surface-mounted cube's setting state. Figure 3 depicts a close-up view of the surface-mounted cube in the wind tunnel's test section.
As shown in Figure 4, the U-shaped horseshoe vortex surrounding a surface-mounted cube is apparent [12]. In the present simulation, a horseshoe vortex was reproduced with a 3D structure, as shown in Figure 5. Figure 5 visualizes the flow by using the smoke-wire technique in the wind tunnel experiment, which used the flow field in the vicinity of the surface-mounted cube. We observed that a highly complex flow field formed in the vicinity of the surface-mounted cube. Figure 6 provides the results of the simulations. Comparing both images shown in Figure 6 and        present numerical simulation with those from the wind tunnel experiment. Satisfactory agreement was found between the results in the wind tunnel experiment and the simulations. Judging from these results, the effectiveness of Airflow Analyst was objectively confirmed.

Application of Airflow Analyst Software to More Realistic and Complex Situations
Next, we introduce simulations of the New National Stadium Japan (Tokyo Olympic Stadium). The results of this elaborate reproduction were obtained by CFD simulation with a large number of grid points/cells using the latest detailed data for urban areas.

Structured Mesh Generation Based on GIS Technique
Generally, four steps are required in conventional software to perform CFD simulations of urban areas, as shown in Figure 9. Process 1, inputting the 3D building data, requires considerable operating time. Here, we will explain the construction method of 3D modeling (and other) of the New National Stadium Japan (Tokyo Olympic Stadium). The Advanced World 3D Map (AW3D) data [7] were employed to express locations and the height of landscape characteristics, buildings, and trees. AW3D [7] is a 3D map created by imagery data from multiple high-resolution satellites in the USA. These data combine city images photographed from various angles without dead angles. Therefore, this technology can distinguish even a building or a tree with 3D images. The map reproduces accurate urban area configurations, providing individual height differences for a building whose roof shape is not horizontal. AW3D data provide global coverage, covering even area information where map data are difficult to obtain. The provided data are categorized into digital terrain models (DTM) for ground surface height with a five-meter resolution, architectural polygons for building shapes, and tree polygons for tree shapes. These data are provided in GIS file formats (tagged image file format (TIFF) format and the ESRI Shapefile format) by the universal transverse Mercator coordinate system (UTM). No processing is needed to directly use these data for numerical wind simulations.
We originally created 3D data based on the published planning drawings of the shape of the Olympic Stadium using 3D modeling software, Sketch UP. The 3D model, which we created, was input to the GIS according to the position and scale of the UTM coordinate system. Airflow Analyst can identify 3D model data from DIM (TIFF), ESRI Shapefile, CAD, and BIM. Then, the software enables mesh generation with ease. Consequently, planners are free from the laborious input operation used in preparing a wind flow analysis and are able to focus on the examination of computational results and reflect on the improvement of a proposed plan.   Figure 11 shows the computational domain and the other parameters set in this study. An enlarged view of the computational grid around the New National Stadium Japan is provided in Figure 12.    Figure 13 shows the air flow pattern in the vicinity of the ground surface (at a height of five meters). The target of the image in Figure 13(a) is only the stadium; it does not consider topography, building, or surface roughness, whereas Figure 13(b) does. That is to say, Figure 13(b) is the result of a simulation that is closer to reality. The simulation results in Figure 13(a), where only the stadium was considered, indicate that a wide variety of strong wind fields form in the side-peripheries. Wind speed was accelerated locally 1.2 -1.3 faster than the approaching wind speed (the region with color ranging from yellow to red in Figure 13). Accordingly, we recognized that the wind flow field, which formed inside the stadium (the blue region in Figure 13) showed complex changes. We confirmed that a complex wind flow field formed on the downstream side of the New National Stadium Japan (Tokyo Olympic Stadium). Focusing on Figure  13(b), where the topography, buildings, and surface roughness in the vicinity of the stadium were examined, we confirmed large reductions in the abovementioned sequence, such as the formation of a strong and complex wind field. Next, we discuss the flow visualization in the vertical plane of the Tokyo Olympic Stadium. As shown in Figure 14, natural and effective wind ventilation is accomplished with unique devices in the upper parts of the stadium, as expected (see the red circle) [13]. The results of the simulations that we conducted ( Figure 15) confirmed that both of the stadium reproductions efficiently took in air from the sky, although there were slight differences in the angles where wind from the sky was taken into the stadium from the upper to the lower portions. We are currently conducting a more quantitative evaluation of the difference in the angle at which the air above is taken into the stadium. Open Journal of Fluid Dynamics

Conclusions
In this study, we developed a new computational fluid dynamics (CFD) model called Airflow Analyst that deepens the affinity between CFD and geographic information system (GIS). Here, we demonstrated the promising effectiveness of Airflow Analyst as follows. First, a precise simulation of the surface-mounted cube was conducted. Validation testing based on the obtained data confirmed the predictive accuracy of Airflow Analyst. Second, New National Stadium Japan We will continue to develop Airflow Analyst, which is a promising analysis platform for this Smart City Project.
Airflow analysis, which is continuing to develop, is used for analyses in diverse fields. Here, we introduce some results from previous studies. First is the case of Shiki Hall of Ito Campus, Kyushu University, which was built in the center zone of the campus ( Figure A1). We elaborated upon an investigation of wind environment with a one-meter spatial resolution ( Figure A2). This hall is a circular Figure A1. Shiiki Hall at Ito Campus, Kyushu University, Japan. were analyzed in detail, assuming cases with a northwest wind whose occurrence frequency is high during winter. This analysis confirmed that a strong wind region was not generated inside the Galleria ( Figure A3(A)).
Next, a simulation result for Fukuoka, Japan is reported. The simulation shown in Figure A4, which was performed in 2011, reproduces the complex air flow behavior generated in the urban Fukuoka area, creating individual building models of building shape clusters, with an approximately three-meter spatial resolution. The findings of this study were thought to be able to effectively use   To support the Epsilon Launch Vehicle, we conducted a CFD simulation with a large number of grid points/cells with the Uchinoura Space Center as a target. The wind direction of the target was northwest, which was recognized as the direction of the prevailing wind during the winter. The results of the numerical analysis clarified that when the northwest wind blew for a long period, the area around Uchinoura Space Center was severely affected by the separated flow from Mt. Kunimi-yama (887 m elevation) located on the upstream side of the center ( Figure A8). We expect further analyses of the obtained computational results to be used for the improvement of wind speed monitoring at the rocket launch site.
In Figures A9-A11, we introduce simulations of the University of Tokyo Atacama Observatory (TAO) Project (August 2014-March 2016), at the request of the University of Tokyo. TAO is a project to construct a 6.5m infrared-optimized telescope at the summit of Cerro Chajnantor (5640 m elevation), located in the Atacama Desert in Northern Chile, and to promote astronomical observations that aim to elucidate the origins of the Galaxy and the planets. This project is led by the Institute of Astronomy (IoA) at the University of Tokyo, Japan.
First, a numerical wind flow condition simulation was performed for a vast area, with the Cerro Chajnantor as the subject, assessing the strength and the vertical distribution of the wind intruding into the Tokyo Atacama Observatory (TAO, Figure A10). Second, TAO was accurately reproduced with a 0.15 m spatial resolution, and an enormous amount of numerical data were obtained from the CFD simulation with a large number of grid points/cells. By analyzing the enormous number of numerical data obtained from the large-scale simulation, we successfully reproduced the changes in complex air flows, which formed inside and outside the telescope's dome ( Figure A11). The number of computational grid points was approximately 100 million in this simulation. We will continue to develop this research to examine the obtained numerical data. We plan to establish a system to control the airflow inside the dome systematically.
The last case that we introduce in this study is a joint investigation between       saying from Antarctica suggests, "Who controls the wind is who controls Antarctica". The findings of our joint investigation contribute to producing more efficient construction plans for the Antarctic Syowa Station.