Introduction
Work by the University of Southampton’s Aerodynamics and Flight Mechanics research group (AFM) has led to advances in the field of Computational Fluid Dynamics (CFD) applied to urban meteorology. New techniques have been incorporated in commercial software releases and adopted by leading design and engineering firms, giving UK businesses a competitive edge over their international competitors. This research is increasingly influencing, for example, the design of wind-sensitive structures by ensuring faster, cheaper and more precise prediction of factors such as peak wind loading on them and pollutant dispersion around them.
Context
The AFM research group at Southampton is a recognised leader in CFD. It leads the UK Turbulence Consortium, which coordinates the use of Britain’s supercomputers for large-scale CFD simulations and seeks to maximise the potential of ever-increasing computer power in modelling turbulent flows. Supported by UK research councils, the European Union and industry, AFM has been at the forefront of efforts to apply CFD to wind flows in urban environments, developing methods sufficiently accurate to compute the turbulence and consequent heat transfer and pollutant dispersion within cities.
Research challenge
Large-eddy simulation (LES) is a CFD technique in which only the smallest eddies are estimated – all the larger ones in the turbulent wind flow are computed exactly. It has been a major component in the science of modelling turbulent flows for some decades. Accurate and increasingly cost-effective, LES is now being used in a range of research fields and related industry sectors, including the clean environment, safety assessment, construction and aerospace. But for modelling environmental flows like local winds and the consequent pollutant dispersal within one part of a city, it is necessary to specify exact conditions at all points where the wind enters the physical area being studied and this poses a serious problem.
Our Solution
In 2008, Dr Zheng-Tong Xie and Professor Ian Castro published a paper that offered the first detailed description of a novel way of specifying the inflow boundary conditions for turbulence simulations using LES. Their proposal was based on a new Hybrid Forward Stepwise (HFS) method, a filter-based concept that was shown to be far more efficient and exact than previous methodologies.
The Southampton engineers argued that the shortcomings of earlier techniques, including the use of white noise and the need to run ‘precursor’ simulations, not only tended to deliver notably inaccurate results but added to complexity and computational times. As well as the obvious benefits of improved precision, the comparative speed of their HFS method – around 500 times faster than the first filter-based approach in 2003 – gave rise to clear implications for the cost-effectiveness of simulations.
The new method’s high levels of efficiency, demonstrated in a series of tests that simulated flows over smooth walls and arrays of staggered cubes, were in large part derived from employing a digital filter technique. This represented an important step towards addressing the longstanding problems associated with applying LES to areas consisting of streets, parks and other urban features.
Further research demonstrated the power of HFS by computing various kinds of generic and real-life urban-type flows. These included an analysis of the Marylebone Road area of London, which led to improved predictions of pollutant dispersion.
Our impact
The results of AFM research have been taken up by leading engineering simulation companies. This collaboration has led to the software companies offering better and more efficient designs with superior safety assessments and improved modelling capacities.
CD-adapco, the world’s largest independent provider of engineering simulation software, support and services, based in the UK has implemented AFM’s new inflow turbulence generation method as a plug-in for its CFD software which is specifically geared towards satisfying the growing legislative emphasis on sustainability and emissions.
Major design and engineering firm Arup invited AFM to help with the preliminary estimation of wind loads on the Gerald Desmond Bridge Replacement, a new structure with a 300-metre main span and a unique cross-section in California. Arup has begun to adopt AFM’s methods in all LES applications for commercial projects, beginning with a stadium being designed in Doha, Qatar, for the 2022 World Cup.
Wider outreach activity has included lectures and presentations to the Atmospheric Dispersion Modelling Liaison Committee, the Environment Agency and the UK Wind Engineering Society.
