Trafalgar Square · wind speed-up at 1.5 m · 4 directions

Method

Archi Automate output, unretouched. sim_wind ran its quick-look preset on the OpenStreetMap context: one steady RANS solution per direction in OpenFOAM v2606, which is installed separately, about 10 minutes per direction (8.7 to 11.7 minutes, measured while another simulation was running on the same PC). The viewer draws the speed-up at 1.5 m and moves particles along the computed flow. Speed-up is the wind at 1.5 m divided by the approaching wind at the same height; 1.0 means the same speed as the approaching wind.

How it was made: the context is the OpenStreetMap import made for the sun study of this square. sim_wind turned the 526 elements within 250 m into solid massing, the tallest 54 m, on flat ground, and meshed each direction on its own: 323,300 to 393,406 cells, 500 iterations each. It samples the speed-up every 2 m over 250 × 250 m, 1.5 m above the ground. A first run, with 180° in place of 300°, stopped when 180° failed to mesh; this is the second run.

The particles start empty when viewer_record begins, so each direction was recorded for 14 seconds and the first 4 seconds were cut, without re-encoding. The viewer’s own direction sweep restarts the particles at every switch, so the four-direction clip joins the four recordings, 2.5 seconds each, with ffmpeg. The square tile shows the prevailing-direction clip whole, on a dark plate.

Limits

• Screening study, steady RANS — not a planning report.
• Four directions only: no yearly statistic, no comfort rating and no gusts.
• Speed-up is a ratio to the undisturbed wind at the same height, not m/s.
• Massing from OpenStreetMap: 184 of the 574 buildings and building parts in the imported context have no height there and use a 9 m default, St Martin-in-the-Fields among them. Terrain is not modelled: the ground is flat. Nothing farther than 250 m from the centre is modelled.
• Next to buildings, a few sample points fall outside the mesh and show no value (69 to 131 of 15,876 per direction).
• No accuracy is claimed for this square.