Simulation-Based Data-Driven Wind Engineering—Analyzing the Influence of Building Proximity and Skyways on Pedestrian Comfort

Pedestrian wind comfort and safety is an important factor in urban city planning. Building structures and details can significantly impact the wind environment, causing undesirable accelerations of the wind close to the ground. Commonly used methods to analyze the wind environment include wind tunnel experiments and computational fluid dynamics simulations. Recently, machine learning techniques have also been proposed as an alternative method. These methods require a large amount of training data to build a prediction model, and existing simulation models can be suitable for the generation of such data. This work demonstrates a methodology to perform a large number of computational fluid dynamics simulations to generate training data for wind engineering problems. The methodology is then applied to analyze the influence of building proximity on pedestrian wind comfort. Additionally, the influence of a skyway, or pedestrian bridge, between the buildings is included. A comprehensive set of simulations is performed varying the surrounding building size and the skyway elevation, span and height. A hybrid approach is then applied, where a data-driven approach is used to identify dominating parameters and to build a fast predictive model, while the simulation data is used to gain further insight into the physical phenomena.