Generative Tall Building
Approximating Computational Fluid Dynamics for Generative Tall Building Design
Samuel Wilkinson1 and Sean Hanna2
2University College London
Background literature review, methodology, results, and analysis are presented for a novel approach to approximating wind pressure on tall buildings for the application of generative design exploration and optimisation.
The predictions are approximations of time-averaged computational fluid dynamics (CFD) data with the aim of maintaining simulation accuracy but with improved speed.
This is achieved through the use of a back-propagation artificial neural network (ANN) with vertex-based shape features as input and pressure as output.
Although computational fluid dynamics (CFD) has existed now for over 50 years and parametric CAD for over 30, both have seen an increased interest in architectural practice over the last decade.
However, in computational design, especially in generative exploration or optimisation, CFD remains a challenging simulation tool to integrate.
There are at least three reasons for this: firstly, the cost of expertise and software is high; secondly, the relationship between a design and its fluid environment is complex, often subtle, and esoteric; and thirdly, the time required to achieve accurate results is typically greater than that available, namely at early project stages when the guidance provided by the simulation is most valuable.
Effects of the wind upon buildings are numerous: for pedestrian comfort in surrounding proximity; ventilation and therefore thermal comfort and indoor air quality; and structural performance.
Wind loads, along with seismic, are the two primary external forces that increase with building height. Therefore tall buildings have been identified as a focal typology for this and a number of reasons.
The aerodynamic shape has a primary impact on these forces and therefore subsequently on the overall structural, material, energy, and financial performance.