How web-based data visualisation shapes cities

Introduction

Advances in measurement and analysis technology and the increasing accuracy of CAE simulations have led to deeper integrations of measurement and analysis data. In the field of CAE, there has been a growing trend towards the extraction of real-world data from high-precision measurement devices to inform highly accurate simulations and analyse results.  By applying highly accurate real-world measurement data to simulation software and visualising analysis results, high-precision measurement data and analysis data are increasingly integrated, and measurement data can then be used to validate the results of virtual simulations. 

Hexagon provides solutions for measurement hardware and design and engineering software, as well as development, sales, and technical support for CAD, CAM and CAE software. Hexagon aims to help users uncover deep insights through the fusion of measurement and analysis.

This paper presents an overview of the simulation of wind environments using Hexagon solutions as an example of the fusion of measurement and analysis. The measurement data includes a 3D model of the Tokyo urban area based on aerial surveys collected by Hexagon’s Geosystems division. The wind environment simulation leverages visualisation software from Cradle CFD, a thermal-fluid analysis software developed by Hexagon’s Manufacturing Intelligence Division. This example demonstrates how the measurement of 3D urban models, wind environment simulations and their visualisation can be achieved using highly accurate solutions from Hexagon.

Wind environment simulation methods 

The phenomenon of strong winds blowing near the ground surface due to the construction of high-rise buildings is known as building wind and is a cause of wind damage in cities. When a new building is constructed, it is necessary to understand how changes in the wind environment around the building will affect nearby residents, inform them of changes to their environment, and take countermeasures in advance if issues are predicted. Some local authorities, particularly when constructing new high-rise buildings, are enacting building requirements that factor the wind environment into construction. Builders may also be required to predict the loads applied to building surfaces by wind. Simulation is often required to accurately predict the effects of wind on the local environment.
 

Wind environment calculation methods　

There are two methods for simulating the wind environment: wind tunnel testing and numerical fluid analysis. In the past, wind tunnel testing was the predominant method, but the rapid development of CAE simulation capabilities means the less expensive method of numerical fluid analysis has become more common. Using the numerical fluid analysis method, CAD data of buildings and the topography of urban areas is aggregated, and the analysis area is divided into calculation elements for analysis.

Vertical distribution of wind speed and surface roughness classification

Fig. 1 Vertical distribution of wind speed

As shown in Fig. 1, natural wind varies in terms of wind speed and turbulence intensity in a power-law distribution in the height direction. For this reason, wind tunnel experiments and numerical fluid dynamics methods need to reproduce wind conditions similar to naturally occurring wind. 

Wind properties are strongly influenced by ground surface roughness. Surface roughness can be expressed in terms of the magnitude of the volume density of buildings and other structures and varies according to differences in regional categories such as maritime, coastal, rural, suburban low-rise residential areas, urban and metropolitan areas. Surface roughness is also determined by the thickness of the turbulence boundary layer and the magnitude of the turbulence. In the construction sector, differences in surface roughness are classified as I-V, as shown in Table 1.

Table 1: Ground surface roughness classification

Wind environment assessment

When assessing the impact of building winds, it is necessary to consider the impact of all wind directions based on their frequency of occurrence, rather than only on the magnitude of wind speeds occurring from a particular wind direction at a particular location. For this reason, evaluation methods based on the frequency of occurrence of wind speeds are commonly used in Japan for wind environment assessment. This method includes the method employed by Murakami et al [1] and the method used by the Wind Engineering Research Institute [2], both of which are often used in Japan [3]. Both express the frequency of occurrence of wind speeds using Weibull distribution, which is a type of probability distribution [4]. The coefficient used in this process is the Weibull parameter, which is calculated based on meteorological observations over a period of three to ten years and takes different values in different regions. Both methods mainly target pedestrian experience, and a different method is required when assessing wind impact on buildings such as houses, or trains and cars in operation.

The method proposed by Murakami uses a wind environment assessment scale developed by Dr Shuzo Murakami based on an analysis of daily maximum instantaneous wind speeds and their impact on the living environment. In this method, the wind environment is assessed by the wind speed ratio at 1.5 m above ground level and the Weibull parameter at the reference point. The probability of exceeding the maximum daily average wind speed is calculated and the results are used to classify the areas on a scale of 1-3 in order of susceptibility to wind, as shown in Table 2. The portion of the area that exceeds rank 3 is generally considered to be rank 4, and this idea is also used in this paper.

Table 2 Wind environment assessment indices by Murakami et al.

The method by the Wind Engineering Research Institute uses a wind environment assessment index created by analysing average wind speeds and corresponding streetscapes. In this method, the wind environment is assessed by the wind speed ratio at 5 m above ground level and the Weibull parameter at the reference point. The cumulative frequency of the average wind speed is calculated and the area is classified from the range of wind speeds where the cumulative frequency is 55% and 95%. This results in a classification that scales from A to D in order of susceptibility to wind, as shown in Table 3. The difference between the method by Murakami et al. and the method by the Wind Engineering Research Institute is that the former method evaluates wind gusts by their frequency of occurrence and the latter by the magnitude of the average wind speed.

Table 3: Wind environment assessment indicators from the Institute of Wind Engineering.

Wind environment simulation case studies

3D city models from aerial surveys

Hexagon's Geosystems division provides geospatial data for the 23 wards of Tokyo. The dataset includes a 3D city model containing orthoimagery with a ground resolution of 7.5 cm, LiDAR point cloud data with a point density of over 20 points/m2 and textured, highly accurate object-based structures. The data was collected using the Leica CityMapper-2 aerial survey sensor, with simultaneous measurements of nadir and oblique imagery, as well as LiDAR data. The data can be used for a wide range of applications, including high-resolution maps for autonomous driving, property management, applications for local government, 5G network planning, and disaster mitigation and disaster information applications such as flood simulation and emergency response.

In this wind environment simulation, 3D mesh data generated using measured LiDAR point cloud data is used as urban geometry data. The data covers an area of approximately 3 km square around Tokyo Station, and an overview of the data used in this study is shown in Fig. 2. Fig. 2(a) shows the shape data including textures and Fig. 2(b) shows the shape data only. The data format is OBJ format. The mapped area of this simulation is within the boxed area of Fig. 3.

Analysis model creation, condition setting, and solver calculations

Cradle CFD's STREAM is a general-purpose thermal-fluid analysis software used mainly in the electronics and building and civil engineering industries. The main feature of this software is that it uses an orthogonal structural grid for the geometry of the calculation elements, which means that there are few situations where model modifications are required, even when complex models are used, and the difficulty level of mesh division is not affected by the model geometry or scale. The urban model in this simulation is also large-scale, making STREAM the ideal solution.

The workflow in this simulation is shown in Figure 4. First, the geometry information of the city model is directly loaded into the STREAM pre-processor to set up the analytical model. The analytical model is shown in Fig. 5(a) and Fig. 5(b) for the horizontal and vertical directions respectively. When creating the analytical model, a sufficient aiding area of 1.5 km each in the horizontal direction and approximately 0.5 km in the vertical direction of the urban model is set up, as shown in Fig. 5. In the solver calculations, the incompressible Navier Stokes equations are solved using the finite volume method, and the standard k-ε model of RANS (Reynolds Averaged Navier-Stokes) is used to model the turbulence for steady-state flow-only analysis. For the inflow boundary condition, a power-law flow velocity boundary condition is used, which takes into account that the wind velocity varies with power relative to the height.

The magnitude of the inflow wind speed is 2.9 m/s, which is the average for the period 1991-2020 in the Tokyo region at the observed elevation [5]. For the discretisation of the advection term, the high-order accuracy QUICK (Quadratic Upstream Interpolation for Convective Kinematics) method is used. The computational elements used in this simulation are shown in Fig. 6(a) and Fig. 6(b) for the isometric and vertical views, respectively. 

The size of the computational elements ranges from 1.5 to 6 m. The number of calculation elements is about 75 million elements and the calculation time is about 50 minutes per wind direction using MPI parallel calculation with 144 Intel Xeon Gold 6140 2.3 GHz cores. WindTool (bundled with STREAM) is used to calculate the wind environment metrics, which can be selected between the Murakami method and the Wind Engineering Research Institute's method, which is used in this simulation. WindTool automatically creates the 16 wind direction setting files required for calculating wind environment evaluation indices from the base pre-processor file and calculates the wind environment indices after performing solver calculations. The tool can be used to efficiently perform everything from the creation of preprocessor files to the tabulation of analysis results.

Visualisation of simulation results

The results of the wind environment assessment index using the method of Murakami et al. using the wind environment simulation results for 16 wind directions using the visualisation software scPOST included with STREAM are shown in Figure 7(a).

The results for the wind direction with the highest frequency of occurrence in the Tokyo region, north-northwest, are also shown in Fig. 7(b). The wind environment assessment index and wind speed contour at 1.5 m from the ground are plotted for each of the two. Fig. 7 is drawn around a 120 m high high-rise building in Kanda Nishiki-cho, Chiyoda-ku, Tokyo, and the analysis results focused on this building are presented here.

The area around the building of interest has been evaluated as a wind environment evaluation index with ranks 1 to 3, and the influence of building winds due to the building can be confirmed in the rank 2 and 3 areas. In the results for the wind direction north-north-west, the wind speed around the building of interest is higher than the surrounding area. The wind environment assessment indicators obtained in this way can be used as a reference to consider building wind measures, such as the installation of fences or planting, if necessary. Such comparisons can be achieved by editing the analysis model on the STREAM pre-processor and calculating the solver again.

When visualising with scPOST, all analysis data in the analysis domain and all measurement data required for visualisation must be loaded into the software in advance. For this reason, high-performance computing is generally required to quickly visualise the results of a large-scale analysis such as this simulation. In recent years, we have been researching and developing visualisation software that can render measurement data and CAE analysis results on a web browser as well as on Windows and Mac operating systems. In particular, visualising the results on a web browser has the great advantage of lightweight computing resources, as the server performs the data storage and analysis calculations and has a function to change the visualisation precision and fidelity as required for the display area. An example of visualisation using this software is shown in Figure 8. 

The wind direction in this figure is south-south-west. In Figure 8 (a), the streamlines resulting from the simulation are superimposed on the texture data in Figure 2 (a). In Fig. 8 (b), the streamlines around the building of interest are displayed and the wind loads on the building of interest are drawn in Fig. 8 (c). Using this software, the wind flow can be rendered with smooth viewpoint movements even on low-end computers. Furthermore, using this visualisation software technology platform, it is also possible to register the viewpoint movements of pedestrians, cars and other vehicles and reproduce urban mobility in software.

An example of this can be accessed in video format via the QR code in Fig. 8(a). In this animation, the viewpoint is shifted from the vicinity of Tokyo Station to the vicinity of Kanda Nishiki-cho while displaying the streamlines of the analysis results. It then moves to the inside of the building of interest, where the results of the air-conditioning analysis inside the building can also be seen. This makes it possible to share the movement of occupants and others among stakeholders, such as the client. The above procedure enables the analysis and visualisation of wind environment simulations using highly accurate measurement data. In addition to this, when constructing a new building, the analysis model can be edited on the STREAM pre-processor and solver calculations can be performed again to compare the analysis results.

Conclusion

This paper introduces a wind environment simulation using Hexagon solutions, including a 3D model of the Tokyo urban area based on aerial survey, thermal-fluid analysis software and visualisation software. An example of visualisation of wind environment assessment indices is presented through wind environment simulation using a wide-area urban model that was measured with high accuracy. The visualisation is capable of being deployed on a web browser. This capability, which is currently under research and development, can be used to draw the wind flow together with the measured data in a smooth operation and also allows multiple people to share the viewpoint movement of people, vehicles, etc. The solution described in this paper can be used not only for building wind analysis but also for heat island measurement and flooding analysis.

References

[1] Shuzo Murakami, Yoshiteru Iwasa, Yasunari Morikawa, A Study on Wind Environment Survey and Evaluation Scale by Occupants' Logbooks : A Study on Wind Characteristics and Wind Environment Evaluation in Low-rise Urban Areas - III, Transactions of the Architectural Institute of Japan, Vol.325, pp.74-84, (1983)
[2] Osamu Nakamura, Masaaki Yoshida, Keiji Yokoya, Junji Katagiri, Wind properties in urban areas - mainly from the cumulative frequency of wind speed, Proc. 9th Symposium on Wind Engineering, pp.73-78, (1986)
[3] Wind Engineering Research Institute, Fundamentals of Building Wind, Kajima Publishing House, (2005).
[4] Shuzo Murakami, Kunio Fujii, Study on wind properties and wind environment evaluation in low-rise areas in urban areas - I : Statistical properties of wind in low-rise areas in urban areas by long-term observation, Transactions of the Architectural Institute of Japan, Vol.310, pp.88-97, (1981).
[5] Japan Meteorological Agency website, Historical weather data search, https://www.data.jma.go.jp/stats/etrn/index.php


This paper is a translation of the article published in the Journal of the Japan Society for Computational Engineering and Science.

Katsuya Ichihashi, Jun Eto and Takao Itami, "Simulation and Visualization of Wind Environment Using 3D City Model of Tokyo Urban Area Derived from Aerial Data", Vol.29, No.1 (2024), (In Japanese)