Wind Load Pressure

The starting point: dynamic pressure

Every wind-load calculation begins with the same scalar. Imagine standing in front of a perfectly flat wall while the wind blows directly at it. The air ahead of you carries kinetic energy ½ρV² per unit volume. Up against the wall the air must slow to zero, and Bernoulli's equation forces that kinetic energy to appear as a rise in static pressure. The increase is the dynamic pressure:

q = ½ ρ V²        (Pa, with ρ in kg/m³ and V in m/s)

At sea-level, ρ ≈ 1.225 kg/m³. Plug in a few wind speeds and the engineering picture springs out:

Wind speed V	= km/h	q = ½ρV²	Per m²	Felt as
5 m/s	18	15 Pa	1.5 kgf	Pleasant breeze
15 m/s	54	138 Pa	14 kgf	Strong wind, hard to walk
30 m/s	108	551 Pa	56 kgf	Severe storm / weak hurricane
50 m/s	180	1531 Pa	156 kgf	Strong hurricane / typhoon
75 m/s	270	3445 Pa	351 kgf	Cat-5 hurricane / EF-3 tornado

Notice how brutally the numbers grow. Wind force is quadratic in speed: doubling V quadruples q. A modest typhoon at 50 m/s puts almost three times the pressure on a wall as a 30 m/s storm. The whole structural engineering of tall buildings is dominated by the rare-event tail of this curve.

Pressure coefficient — how the wind decides where to push

Dynamic pressure q is what the upstream air "carries". The actual pressure on any particular face of the building is

p = q × C_p

where C_p is a dimensionless pressure coefficient that depends on where on the building you are. At a perfect stagnation point on the windward wall, all the kinetic energy is converted, so C_p = +1. Real buildings see C_p ≈ +0.8 on the windward wall (the flow is not quite normal everywhere on the face). At the side walls the flow accelerates around the corner, the local static pressure drops, and C_p falls to about −0.7. Behind the building, separated flow leaves a low-pressure wake — C_p ≈ −0.5 on the leeward wall, and as low as −1.4 at roof corners and edges. Cladding designers care most about those local minima; that is where panels get sucked off.

Location	C_p (typical)	What it means
Windward wall	+0.8	Air pushes inward
Side wall (mid-face)	−0.7	Air sucks outward
Leeward wall	−0.5	Wake suction
Flat roof corner	−2.0 (worst case)	Conical vortex lifts the roof
Hip roof, slope > 30°	+0.4 / −0.7	Windward slope pushed, leeward sucked
Long-low building, end zones	−1.4	Local cladding peak suction

Total wind force on an area A normal to the face is then

F = q × C_p × A

and for the along-wind shear on the building as a whole you add the windward push to the leeward suction (which have the same sign on the building since one pushes east and one is sucked east).

Wind grows with height: the boundary-layer profile

Wind speed at 10 m elevation in open terrain is not the same as wind speed at the 87th floor. Surface friction creates an atmospheric boundary layer, and within it the mean wind speed follows a power law (or logarithmic) profile:

V(z) = V_ref × (z / z_ref)^α

with the exponent α depending on terrain roughness. ASCE 7-22 defines three exposure categories:

Exposure	Terrain	α (power-law)	z_g (gradient)
B	Urban / suburban, with closely spaced obstructions	~0.22	365 m
C	Open terrain with scattered obstructions	~0.15	274 m
D	Flat unobstructed (water, mudflats)	~0.11	213 m

The shape of the curve matters for tall buildings. A 300 m tower in exposure B sees a wind speed at the top about 1.9× the reference value at 10 m; the corresponding dynamic pressure scales as that ratio squared — about 3.6× the ground-level q. That is why design pressures at the top of a skyscraper can run two to four times those near the base.

Putting it together: the ASCE 7-22 design equation

Modern codes wrap all the corrections into a single design formula. ASCE 7-22 for the main wind-force resisting system:

p = q_z × G × C_p          (windward)
p = q_h × G × C_p          (leeward, sidewall, roof)

q_z = 0.613 × K_z × K_zt × K_d × K_e × V²        (Pa, V in m/s)

The constants are:

• V — basic wind speed from the ASCE hazard map. Risk-category II buildings use a 700-year return period speed.
• K_z — velocity-pressure exposure coefficient, encoding V(z)² / V_ref². Tabulated by height and exposure.
• K_zt — topographic factor for hills and escarpments (1.0 on flat ground; up to about 1.6 on a sharp ridge crest).
• K_d — wind directionality factor (≈ 0.85 for buildings); accounts for the fact that the worst wind speed and the worst direction rarely coincide.
• K_e — ground elevation factor (high-altitude air is thinner; K_e = 1.0 at sea level, < 1 at altitude).
• G — gust-effect factor, ≈ 0.85–1.0 for rigid buildings, computed dynamically for flexible ones.
• C_p — external pressure coefficient (the value of which depends on face and aspect ratio).

Eurocode 1 (EN 1991-1-4) lays the same ingredients out slightly differently — basic wind velocity v_b, mean velocity v_m(z), turbulence intensity I_v(z), peak velocity pressure q_p — but the result on a real structure typically agrees to 10–20% with the ASCE value.

Worked example: a 200 m tower in a 50 m/s storm

A square plan tower, 50 m × 50 m × 200 m tall, in ASCE Exposure C, in a storm with basic V = 50 m/s. We want the design pressure on the top face and the along-wind base shear.

Step 1 — height profile. K_z at z = 200 m, Exposure C:
        K_z = 2.01 × (z/z_g)^(2α) = 2.01 × (200/274)^(0.30) ≈ 1.83

Step 2 — velocity pressure at the top:
        q_h = 0.613 × K_z × K_zt × K_d × K_e × V²
            = 0.613 × 1.83 × 1.0 × 0.85 × 1.0 × 50²
            = 2384 Pa  (≈ 2.4 kPa)

Step 3 — design pressure on each face (G = 0.85):
        Windward  p = 2384 × 0.85 × (+0.8) = +1621 Pa  (push)
        Leeward   p = 2384 × 0.85 × (−0.5) = −1013 Pa  (suction)

Step 4 — along-wind base shear per unit height:
        f = (p_W − p_L) × B = (1621 − (−1013)) × 50 m
          = 2634 Pa × 50 m = 131.7 kN/m

Step 5 — total base shear (uniform approximation):
        F = f × H = 131.7 × 200 = 26.3 MN

That's ~2700 tonnes-force pushing horizontally on the foundation.

This is the kind of number that turns into a core wall thickness on a structural drawing. The same equation, run with C_p on roof corners and at cladding panel scale, sets the suction every glass panel and roof flashing must resist.

Vortex shedding and the across-wind problem

Static along-wind force is only half the story. When wind flows around a bluff body (square, rectangular, or circular cross-section), the boundary layer separates and rolls into discrete vortices that shed alternately from each side — the Kármán vortex street. Each vortex carries circulation, and shedding it imparts an impulse to the body normal to the wind. The shedding frequency is set by the Strouhal number:

f_s = St × V / D     (St ≈ 0.2 for square sections and cylinders in the relevant Re range)

Trouble arrives when f_s matches one of the building's own natural frequencies. The wind then drives the structure at resonance, and across-wind motion grows until non-linearity or damping limits it. The critical wind speed for lock-in is

V_cr = f_n × D / St

For a 500 m square-plan tower of natural frequency 0.1 Hz and side D = 60 m, V_cr ≈ 30 m/s — a commonplace storm wind. The cure is to (a) raise damping, (b) detune the geometry so vortex shedding is incoherent, or (c) add a tuned mass damper.

Tacoma Narrows, 1940

The 1940 collapse of the Tacoma Narrows suspension bridge is the textbook cautionary tale. The deck oscillated in a torsional mode in winds of only 19 m/s; popular accounts blame "resonance with vortex shedding", but the modern engineering consensus, beginning with the work of Robert Scanlan in the 1970s, is that the failure was aerodynamic flutter — a self-excited coupling between torsion and the lift on the H-shaped deck, distinct from but related to classical vortex-induced vibration. Either way, the lesson stuck: tall, slender or thin-section structures must be designed for fluid-structure interaction, not just static pressure.

BAE / Ferrybridge cooling towers, 1965

Three of the eight 114 m reinforced-concrete cooling towers at the Ferrybridge C power station collapsed in November 1965 during a gale at 41 m/s. Buffeting and shielding effects between adjacent towers produced higher local pressures than the original single-tower design considered. The inquiry led to substantial revisions in UK wind-loading guidance for shell structures and to formal recognition of group effects and aeroelastic instabilities in the British codes.

Three dynamic effects, three mitigations

Effect	Mechanism	Mitigation
Along-wind buffeting	Random gust pressure on the windward face	Gust factor G in the static design; lateral stiffness
Across-wind vortex shedding	Vortex street normal to wind direction	Aerodynamic shaping (chamfered/rounded corners, setbacks, porosity), tuned mass damper
Torsional galloping / flutter	Self-excited oscillation when aerodynamic damping turns negative	Avoid resonance with section redesign; stay above critical speed; tuned-liquid dampers

The hierarchy is roughly: static along-wind is the smallest concern (well-understood, conservative codes); across-wind vortex shedding governs most tall-tower design; torsional flutter is a check, not a design driver, for buildings (but it is the driver for long-span bridges).

Tuned mass dampers and architectural shape control

Two engineering responses dominate at the very tallest scale.

Taipei 101 (508 m). A 660-tonne polished-steel sphere hangs from eight steel cables between the 87th and 92nd floors, visible to visitors as part of the building's design language. Tuned to the tower's first sway period of ~6.8 s, the pendulum swings out of phase with the deflection; viscous dampers around its base bleed energy as heat. The system absorbs roughly 40% of typhoon-driven sway acceleration, keeping the building comfortably within the human-comfort limit of ~0.15 m/s² peak acceleration over a 10-minute window. In Typhoon Soudelor (2015) the damper recorded a peak swing of about 100 cm — the largest ever observed in a TMD, but well within its 1.5 m clearance.

Burj Khalifa (828 m). No TMD. Instead, the geometry itself was tuned in wind-tunnel testing at RWDI: a three-lobed plan that steps inward in 27 setbacks as the building rises. Each setback breaks the coherence of vortex shedding — the shedding frequency, which depends on local cross-section dimension D, varies discontinuously with height, so no single wavelength of disturbance can resonate with the tower. Wind-tunnel comparisons with an equivalent prismatic 800 m tower show roughly a 25% reduction in peak across-wind moment.

Shanghai Tower (632 m). A 120° twist along the height plus a tapered, rounded triangular plan with an outer glass skin separated by a 1–10 m gap from the inner structural cylinder. The twist breaks the spanwise coherence of vortex shedding; the gap allows the outer skin to act as a windbreak for the working structure. Across-wind load reduction is reported at ~24% with the addition of a 1000-tonne eddy-current TMD as backup.

Citicorp Center (1977). The first building to use a tuned mass damper for wind-induced sway — a 400-tonne concrete block sliding on an oil film at the 59th floor, controlled by hydraulic pistons. Notably retrofitted with structural reinforcements in 1978 after engineer William LeMessurier realised the bolted (not welded) connections combined with cross-wind loading from a particular quartering wind direction could overstress the brace welds.

Cladding pressures vs main-system pressures

One subtle but critical point. The "main wind-force resisting system" (MWFRS) — the columns, cores, and bracing that carry the whole building's wind load — sees area-averaged pressures: large faces over which local C_p peaks have been averaged out by the gust factor's spatial correlation. Individual cladding panels see local pressures that can be much higher because they sample one small spot at one instant. ASCE 7-22 publishes a separate table of "components and cladding" pressure coefficients for this reason; for a flat roof corner zone, GC_p can reach −2.6, against an MWFRS-equivalent of perhaps −1.0. Many cladding failures in hurricanes (Andrew 1992, Wilma 2005, Harvey 2017) traced back to neglecting this distinction.

When you need a wind tunnel

The ASCE-7 analytical procedure is calibrated for "regular" rectangular buildings up to ~120 m. Above that height — or for any structure with significant slenderness (H/√(B×D) > 4), important reentrant corners, asymmetry, fundamental period exceeding 1 s, or unusual aerodynamic shape — codes require either a wind-tunnel test or full computational fluid dynamics study. The standard test uses a 1:300 to 1:500 boundary-layer scale model with surface pressure taps, instrumented in a wind tunnel that reproduces the atmospheric boundary-layer turbulence appropriate to the site. Aeroelastic models (flexible scale models matched in stiffness and damping) are used for the most slender or unconventional towers. Most super-tall buildings spend six to eighteen months in tunnel iterations before any structural design is frozen.

Common pitfalls

• Forgetting suction on the leeward and side faces. The total along-wind force is windward push plus leeward suction (they add). Designers who only count the windward face underestimate by up to a factor of two.
• Using V at ground level for the whole height. Wind speed grows with z; pressure grows with V². A uniform-pressure assumption is conservative for the base but unconservative for the top.
• Treating the gust factor as a constant. G ≈ 0.85 is the "rigid" approximation. For flexible structures with the first natural frequency below 1 Hz, the dynamic G_f can exceed 1.5; ASCE 7 has a separate procedure.
• Confusing 3-second gust and 10-minute mean. ASCE uses a 3-second gust speed in its hazard maps; Eurocode uses a 10-minute mean. The conversion factor is roughly 1.4–1.5; mixing them gives wrong design pressures by a factor of about two.
• Ignoring across-wind for slender towers. Static along-wind analysis can pass while the building fails serviceability on occupant comfort because of across-wind vortex-induced acceleration.
• Trusting the code envelope on unusual shapes. Code coefficients are tabulated for rectangular and a few standard shapes. A re-entrant L-plan, an irregular cross-section, or a stepped tower needs wind-tunnel verification — code C_p values can be wrong by ±50% on such geometries.