3D Tuned Mass Damper: Taming a Swaying Skyscraper

A Blender-built glass tower in wind: vortex shedding, resonance and a roof damper tuned by the Den Hartog rules, with live sway, acceleration and three numerical checks.

Switch to 2D

Presets

Playback

Parameters

Camera

Layers

Quality

Model check

The science

What is modelled

A 80 m glass tower, 24 m wide, swaying in its first bending mode. A steady wind sheds vortices alternately from the two downwind corners, which pushes the tower sideways with an alternating lift force. When the shedding frequency matches the tower's natural frequency the sway builds up (resonance). A tuned mass damper is a heavy block on springs and dampers near the roof, tuned so that it moves against the tower and takes the energy out of the sway.

Equations

  • Tower, one degree of freedom (tip coordinate x, generalised mass M): M x″ + c1 x′ + k x = F(t) − [cd(x′−y′) + kd(x−y)], with k = Mωn² and c1 = 2ζ1Mωn.
  • Damper mass md = μM: md y″ = cd(x′−y′) + kd(x−y), kd = md(ftωn)², cd = 2ζdmdftωn.
  • Vortex shedding: fs = St·U/D with Strouhal number St ≈ 0.2 for a square section; lift amplitude F0 = ½ρU² D H CL g where g = 0.363 is the integral of the mode shape 1 − cos(πz/2H) over the height.
  • Steady state (phasors, Ω = 2πfs): X = F0 / [k − Ω²M + iΩc1 − Ω²md(kd+iΩcd)/(kd+iΩcd−Ω²md)].
  • Den Hartog optimum: ft = 1/(1+μ), ζd = √(3μ / (8(1+μ)³)). For an undamped tower the peak amplification is then √(1 + 2/μ), against 1/(2ζ1) without the damper.
  • Peak acceleration at the tip: a = Ω² |X|.

Checks run in the page

  • Den Hartog's closed form √(1+2/μ) against a numerical frequency sweep of the full two-mass equations.
  • A long RK4 time integration from rest (tower and damper) against the phasor formula, with and without the damper.
  • An energy balance over whole shedding periods: work done by the wind equals the energy dissipated in the tower and in the damper.

Honest simplifications

One sway mode, one direction, linear springs and dampers. Real vortex shedding locks onto the tower's frequency over a band of wind speeds and the lift force has a random part; here the force is a pure sine at St·U/D and the lift coefficient is a constant (0.25), so the amplitudes are illustrative rather than design values. There is no turbulence, no along-wind response, no torsion and no limit on the damper stroke. The sway and the damper motion are drawn exaggerated (slider) because real amplitudes of a few centimetres would be invisible; the red line and the graphs use the same exaggeration, the numbers do not. The 0.15 m/s² comfort line is an illustrative round figure, not a code value. The streaks and rings show where the wind goes and are schematic. This is educational, not a structural design tool.

What to try

  • Start with the resonant wind and no damper, then press the preset with the 1 % optimum damper: the sway drops by about 80 %.
  • Detune the damper (tuning 0.8) or switch its damping off: most of the benefit disappears.
  • Raise the mass ratio to 3 %: a heavier damper works over a wider range of wind speeds (third graph).
  • Move the wind speed away from 30 m/s and watch the shedding frequency leave the tower's frequency.
Loading model...
Loading Blender model...
Tip displacement against time, with and without the damper
Frequency response: amplification against fs / fn
Peak tip acceleration against wind speed
Made with 🟠 Blender · Three.js