Why the Gap Between Two Towers Is Not a Wind Tunnel

Rand Simulation — Applications Engineering AI
Built environment · Ansys Fluent · 7 min read

AI disclosure: RandSim Labs is an experimental AI-driven engineering simulation platform. Content on this site, including simulations, analyses, figures, and written materials, may be generated or assisted by AI using licensed Ansys tools. AI-generated content may contain errors and is provided for educational, informational, and demonstration purposes only. Users should independently verify all results before relying on them for engineering, design, manufacturing, safety, or other production decisions.

Two glass towers in a 50 mph wind, streamlines bending around them and slowing in the sheltered gap, wind-amplification field on the street — A 50 mph wind meeting two 120 m towers, solved in Ansys Fluent. Streamlines are colored by head-height speed — fast (red) as they bend *around* the towers, slowing to cyan as they enter the gap. The street is painted with the wind-amplification field: deep blue where the wind is calmed, red at the windward corners where it speeds up.

The result: everyone expects the narrow slot between two skyscrapers to act like a wind tunnel — pinch the air, speed it up. We drove a 50 mph wind through two 120 m towers in Ansys Fluent and it does the opposite. The street between them is a sheltered pocket, running about a third slower than the open approach. The wind that actually knocks people over is out at the building corners, where it accelerates past 1.5 × and crosses the line into “unsafe for vulnerable pedestrians.” A screen strung across the gap would guard the calmest square meters on the block.

The intuition everyone has

Anyone who has rounded the base of a downtown tower on a blustery day knows the ambush: the wind that was a nuisance on the open street suddenly tries to take your umbrella, your hat, and your footing. The folk explanation is a Venturi — squeeze a moving fluid through a narrow gap and it has to speed up, the way a garden hose does when you thumb the end. So the slot between two towers ought to be the worst place of all.

It is a clean intuition, and for a sealed pipe it is correct. For open air around buildings it is not. Blocken, Carmeliet and Stathopoulos showed in 2008 that as the passage between two buildings narrows, more of the approaching air simply escapes over the top and around the sides rather than forcing itself through the gap — so the speed-up never climbs the way the pipe formula predicts. They even found diverging passages more severe than converging ones, the exact opposite of the textbook picture. We wanted to see that play out in a solve.

What we put in the wind

Two square towers — 40 m on a side, 120 m tall, roughly the height of Leeds’ Bridgewater Place — standing 25 m apart, a realistic downtown street gap of about a fifth of a building height. The wind is a 50 mph gale, applied the way a wind engineer would rather than as a uniform block: an atmospheric-boundary-layer profile pinned to 50 mph at the standard 10 m weather-mast height. Over the modeled urban ground roughness that leaves about 11 m/s — a stiff 25 mph — at head height (1.75 m) in the open, before the towers touch it.

The solver is Ansys Fluent: incompressible air, k-ω SST turbulence, a 1.8-million-cell mesh with prism layers packed tight against the ground and the towers so the wind is actually resolved at head height and not sampled 10 m up. The subtle part is how you define “amplification” honestly. We solved the same domain twice — once with the towers, once completely empty — and report R, the local wind speed divided by the empty-domain speed at the very same spot. That is the standard Blocken amplification factor, and it cancels any drift in the approaching profile. The proof it is working: far from the towers, R lands at 1.00.

Plan view of wind amplification at head height: blue sheltered gap and wake, red corner jets — Wind amplification R at head height (1.75 m), plan view. Blue is calmer than the open approach, red is faster; white is R = 1. The gap and the long wake behind the towers are deeply sheltered; the only red is at the windward corners, where the streamlines wrap the sharp edge into a jet.

The gap is a shelter, not a tunnel

Down the 25 m street between the towers, the mean head-height wind comes out at R ≈ 0.66 — about a third slower than the undisturbed approach. The pair of towers blocks far more air than it funnels, and the surplus spills up and over rather than accelerating through the slot. Directly behind them sits a broad, calm wake. If you were looking for somewhere to get out of the wind on that gusty day, the gap between the towers is one of the best spots on the block — the precise opposite of the wind-tunnel fear.

The real hazard is the corners

The red in the map tells the true story, and it is all at the windward corners. There the flow attaches to the front face, races to the sharp vertical edge, and separates into a narrow, high-speed jet. Amplification peaks at R ≈ 1.53 — a head-height speed near 16.9 m/s, about 38 mph. On the Lawson comfort scale (sitting 4, standing 6, walking 8, brisk walking 10 m/s) that clears the S15 threshold: unsafe for vulnerable pedestrians — the elderly, children, cyclists. And because this is a steady, mean-speed solve, the real gusts a person feels there are higher still.

Pedestrian comfort map in Lawson categories: uncomfortable in the open, calm in the wake, unsafe patches at the corners — The same plane read as absolute comfort. A 50 mph gale is already “uncomfortable” at head height everywhere in the open (orange); the towers carve out calm zones in their shelter (green) and small *unsafe* patches (red) at the windward corners.

This is not an abstract worry. Bridgewater Place in Leeds, a 112 m slab, funnelled street-level gusts of 67–79 mph around its base; in 2011 a pedestrian was killed when a lorry was blown over in that wind. The remedy that was eventually built was not a panel across a gap — it was roughly 6 m perforated baffles, vertical screens and a 65 m canopy placed at the corners and along the affected street.

So — would a wind screen be required?

Yes, but not where the intuition points. A screen hung across the gap would shelter the calmest part of the site while doing nothing for the wind that actually endangers people. That wind is at the corners and along the flanks, and that is where baffles, canopies, colonnades or dense planting earn their keep — exactly the Bridgewater lesson. The first job is knowing which square meters are dangerous, and that is what the simulation is for: it turns “the corner feels windy” into a map you can design against.

Honest scope. This is a steady Reynolds-averaged (k-ω SST) solve for a single wind direction aligned with the street, reporting mean-speed amplification. RANS does not resolve the turbulent gusts that do much of the real harm — measured peaks near towers run roughly 2–3× the approach wind, so the corner figures here are a mean-speed floor, not the worst a person feels. A smooth solved ground does not perfectly hold the wind profile over a domain this long, which is exactly why amplification is taken against an empty-domain reference that develops the same way. The towers are generic squares, not a specific building, and there is no local wind-rose or weather record here, so this is the physics of one strong-wind day rather than a formal Lawson compliance assessment. None of that moves the headline: the gap shelters, and the corners bite.

Planning a tower, a plaza, or a canopy and need to know where the wind actually goes? Pedestrian comfort is a solve, not a guess, and the dangerous ground is rarely where intuition puts it. We run urban wind and pedestrian-comfort CFD in Ansys Fluent, geometry through comfort map. Rand Simulation — innovation through insight.