Technology that optimises the performance of Formula 1 racing cars is now solving problems for tall buildings.
More from: F1 technology advances tall building design
As tall buildings become ever taller and more complex architecturally, so understanding the effects of wind on these structures and their surroundings becomes a major headache for the designers. When the building is to be located in a congested urban environment the challenge can be even greater.
Planning authorities seek reassurance that wind spilling off a proposed building will not create uncomfortable or even dangerous conditions for pedestrians in nearby streets.
Until recently the only analytical tool available to design teams was wind tunnel testing of scale models, now a mature technology. However, scale effects inseparable from the use of such small models mean that the real world impact of the full size building is difficult to quantify with an acceptable degree of confidence.
Now, however, analytical methods developed for competition cars are being applied to other types of vehicle – and to applications far removed from road transport.
Bicester-based Wirth Research is leading the move into architectural computational fluid dynamics (CFD), using a supercomputer so powerful it requires its own electricity generator.
“In fact it’s more powerful than is allowed in Formula 1,” says Wirth Research engineering manager Rob Rowsell.
“Used in conjunction with our advanced moving platform simulator on racing car projects, it allows us to go straight from virtual design and analysis to component manufacture to track testing. Results from the CFD can be available overnight, and we can try out different variations quickly and cheaply.”
It’s more powerful than is allowed in Formula 1
Rob Rowsell Wirth Research
Wirth has never confined itself to F1. It has optimised designs for Le Mans sports cars and American Indycars as well, with many race victories to its credit. Recently, however, the company has been involved in other sectors, including road transport.
F1 typical CFD output
Rowsell reports: “We worked with Eddie Stobart’s HGVs, and came up with an aerodynamic package that reduced drag by 10% below the most effective commercially available alternative.
“And we’ve developed supermarket freezers, cutting energy use by nearly 20%.”
Expanding into the building design sector was a logical move, given the known shortcomings of the traditional wind tunnel testing. Wirth Chief CFD engineer Jon Winchester explains: “Typically a tall building wind tunnel model will be to 1/250 scale. A physical pressure probe will be sampling an area equivalent to 1.25m across at full scale.
“By contrast, high resolution CFD would be working with elements as little as 100mm across. This gives much more detailed results.”
In some cases building analyses will be based on more than 300M such elements, known as cells, which make up the ”mesh”. With car analyses these cells can be as little as 0.5mm across, and a typical CFD car model may have more than 500M cells.
Validating the results from a building analysis presents problems. A car design can be wind tunnel tested at full scale and put through exhaustive track testing. As this approach is obviously impractical for a super tall building, the Wirth team had to combine the faith in the technique it has developed over the years with virtual wind tunnel tests at 1/250 scale.
“If we get good correlation between the virtual results and the results from actual wind tunnel testing we are confident we can scale back up to full scale,” says Winchester.
One recent challenge was to build a virtual model of the business centre of Chicago and analyse the effect of wind on three super tall towers. Set among dense clusters of lesser skyscrapers, the lower sections of each super tall building are in the lee of its neighbours. Above this, however, the towers are fully exposed.
What the design teams needed was information on vortex shedding and the effect of neighbouring buildings. To achieve reliable results, the Wirth team carried out a transient analysis, otherwise known as an unsteady flow analysis.
This takes many times longer to run than a basic steady state analysis, which calculates average values. Transient analysis gives a more realistic picture of how the wind flow patterns vary over time and identifies the worst case scenarios much more accurately.
The different patterns of wind around each tower clearly demonstrate the effects of the individual architectural design. Closer to home, the Wirth team is now working with PLP Architecture on the detailed design of 22 Bishopsgate, a 278m, 62-storey tower located in the City of London’s existing tall building cluster.
F1 Tall Buildings London
It will be the second tallest building in London when completed. Construction of its planned predecessor, the Pinnacle, stopped in 2012, with only the first seven stories of the concrete core actually built.
Redesigned and renamed, the project is due for completion in 2019. It had to satisfy the planning authorities on a number of counts, not least the potential loss of natural light to surrounding buildings. Also vital was reassurance on the new building’s effect on pedestrian safety and comfort.
Downwash from the proposed building would be funnelled along adjoining roads and pavements. Initial CFD analysis indicated that limits set by the Lawson Comfort Criteria, the accepted guidance, would be exceeded frequently in certain areas. The problem was worst in Bishopsgate itself, but there were areas in adjoining streets that were also under threat.
“We came in later than we would really like,” Winchester reports. “There had already been a lot of wind tunnel testing done, which had fixed most of the issues, but couldn’t find a solution for the problems on Bishopsgate itself.
“Everything above the fourth floor was fixed, so we could only work below that level.”
Vanes or baffles to deflect the downwash and keep it above the heads of pedestrians were one possibility. These seemed to be most effective when located at the south western corner, the upwind corner in the prevailing winds. But how many baffles would be needed, and what shape?
Wind tunnel testing of all the options, even if possible, could take months. Instead, Wirth ran simulations of no fewer than 25 variations of baffle designs before coming up with the most effective. All in just three days.
The solution was a stack of six curved baffles, 8m above pavement level. “Then we handed it over to the architect to refine it for aesthetic reasons,” says Rowsell.
There are other aspects of building performance for which only CFD can yield reliable results. The effects of wind-driven rain are a particularly interesting area at the moment, Rowsell says.
This technology has been refined at the coal face of F1 design and can offer insight and a level of detail that no wind tunnel can achieve
Rob Rowsell, Wirth Research
How external winds affect the internal environment of buildings dependent on natural ventilation is another problem beyond the capability of a wind tunnel. So is mapping the spread of airborne pollutants under different wind regimes, and analysing smoke dispersion and fire routing. On a different scale, CFD can provide valuable inputs into landscape design and master planning.
Wirth’s secret is the sheer computer power it can deploy – a 45 teraflop cluster with more than 16 terabytes of memory. “This is orders of magnitude above the typical computer used in building design,” says Rowsell.
“This technology has been refined at the coal face of F1 design and can offer insight and a level of detail that no wind tunnel can achieve.
“We can study the effects of different wind angles and atmospheric and climatic conditions and seasonal variations. For tall buildings, this means potential aerodynamic problems can be identified and resolved during the design process.”
F1 tall buildings mitigation scheme