The concept and implementation of Smart Cities is an important approach to improve decision making as well as quality of life of the growing urban population. An essential part of this is the presentation of data from different sources within a digital city model. Wind flow at building scale has a strong impact on many health and energy issues in a city. For the analysis of urban wind, Computational Fluid Dynamics (CFD) has become an established tool, but requires specialist knowledge to prepare the geometric input during a time-consuming process. Results are available only as predefined selections of pictures or videos. In this article, a continuous, semi-automated workflow is presented, which ❶ speeds-up the preparation of CFD simulation models using a largely automated geometry optimization; and ❷ enables web-based interactive exploration of urban wind simulations to a large and diverse audience, including experts and layman. Results are evaluated based on a case study using a part of a district in Stuttgart in terms of: ➀ time saving of the CFD model preparation workflow (85% faster than the manual method), ➁ response time measurements of different data formats within the Smart City platform (3D Tiles loaded 30% faster than geoJSON using the same data representations) and ➂ protocols (3DPS provided much higher flexibility than static and 3D container API), as well as ➃ subjective user experience analysis of various visualization schemes of urban wind. Time saving for the model optimization may, however, vary depending on the data quality and the extent of the study area.

Air pollution from motor vehicles in densely populated cities remains an urgent problem. In order to analyse and predict air quality at city scale, in-situ monitoring systems of pollutant concentrations and meteorological parameters are used, supplemented by mathematical prediction models and numerical simulations. However, local pollution processes in urban areas located close to roads often cannot be studied efficiently by monitoring due to the lack of a detailed sensor network. Numerical modelling at the urban micro-scale can provide detailed information on the pollutant transport around buildings, but requires high resolution building geometry and approaches to describe boundary and initial conditions for meteorological parameters, pollution sources and traffic intensity.In the present study, the authors focus on the numerical modelling of scenarios of pollutant transport from vehicles in the vicinity of an urban environment fragment in Stuttgart, Germany. The modelling methodology is based on a combination of classical CFD methods, sensor data on traffic intensity and vehicle statistics, and meteorological data. The traffic statistics used in the simulation include the number of vehicles per hour, divided into 11 vehicle classes, and vehicle speed data. The numerical simulation was performed using the 3D Reynolds-averaged Navier-Stokes equations with the k-ε turbulence model, supplemented by gaseous pollutant and particulate matter transport models. Ansys Fluent 2023R1 is used as the main CFD modelling tool.Non-stationary measured data of traffic on the considered road section were analysed and used to improve the numerical model. This allows the description of the daily air pollution dynamics within the urban fragment. This detailed simulation of unsteady aerodynamics and air quality at the urban micro-scale can be used to focus on indoor/outdoor air exchange problems, to optimise the transport system, and to contribute to the provision of sustainability and climate safety technologies in cities.