A framework for pedestrian-level wind conditions improvement in urban areas: CFD simulation and optimization

Given the complexity of urban wind flow, acceptable wind conditions can hardly be achieved in all regions within an urban area. In some regions, high wind speeds can lead to wind discomfort and danger, while some regions suffer from poor outdoor urban ventilation. In order to evaluate and improve urban wind conditions, the combined impacts of wind flow and morphological and aerodynamic characteristics of the urban area should be taken into account. In this perspective, performing extensive and time-consuming parametric analyses is required – which is very difficult – even impossible. This paper, therefore, presents an evolutionary-based framework to optimize building heights and plan area densities to improve wind conditions in complex real urban areas. The framework consists of four steps: (i) urban area identification, (ii) optimization algorithms, (iii) CFD simulations, and (iv) evaluation. The optimization consists of Genetic Algorithm and Particle Swarm Optimization. The framework is applied to a real urban area in the city of Tehran, Iran. The focus is on minimizing low-wind-speed regions. In total, 920 possible combinations of building height variations and plan area densities are assessed. 3D steady Reynolds-averaged Navier-Stokes (RANS) simulations are performed based on a detailed validation study and grid-sensitivity analysis. The results show that the use of this framework can effectively reduce computational time and significantly improve the wind conditions in the entire urban area. Urban planners, architects, and building engineers can benefit from this framework in the early stages of design for better wind conditions in complex real urban areas. © 2020 Elsevier Ltd