Târgu Jiu, Romania · Brâncuși, 1938 · 29.3 m of steel and cast iron
The Endless Column what its beads do in the wind
Under the 15 cast-iron beads of the Endless Column (Coloana fără sfârșit) by the Romanian sculptor Constantin Brâncuși, in Târgu Jiu, stands a 28.90 m steel spine, 38 centimetres wide at the top. Without the beads, by Giovanni Solari's calculations, the spine could start to gallop, swinging ever wider, in a 20 m/s wind, below the 30.6 m/s design wind. With the beads, the threshold rises to 95 m/s. Increase the wind and compare the column with the bare spine.
The galloping thresholds are those Giovanni Solari estimated from forces measured in the wind tunnel, for the most exposed direction of each version. The motion on screen is illustrative.
The spine inside
The Endless Column, raised in 1937–1938 at Târgu Jiu, is 29.26 m tall according to Solari: 15 full cast-iron modules, each measuring 45 × 90 × 180 cm, and an end piece at the base and at the top, threaded on a 28.90 m steel spine. The spine is not a solid bar: four steel angles tied with batten plates at the top, clad in plates that thicken towards the base, riveted and bolted, with concrete inside up to 5.50 m.
The model built from those dimensions gives the bare spine a natural frequency of 0.79–0.90 Hz, depending on what the battens and rivets weigh. On 3 November 2000, during the restoration, the measured frequency was 0.830 Hz. The second bending frequency comes out 4.31 times the first in the model; measured, 4.50 times.
Galloping
As the spine moves sideways, it changes the angle at which the wind meets it. If the resulting force puts more energy into each swing than damping removes, the oscillation grows: that is galloping. Giovanni Solari measured the forces on models of the column and of the spine in a wind tunnel, then calculated the speed at which galloping begins: 20 m/s for the bare spine and 15 m/s for the spine wrapped in sheeting, as it stood during the works. At the top of the column, the mean wind speed has a 2% chance of exceeding 30.6 m/s in any year.
The calculation here, with the model's spine mass and the frequency and damping measured in 2000, gives the bare spine a threshold of 20–26 m/s, close to Solari's. Without the beads, by Solari's account, the spine would have faced a very high risk of collapse by galloping.
What the beads do
The beads add mass: 0.86 t for each full module. With that mass alone, the model spine would vibrate at 0.43–0.45 Hz; measured with the beads fitted, on 14 November 2000, it vibrated at 0.513 Hz. The difference shows that the modules are not just a weight: they make the column about 40% stiffer in bending, which Solari modelled with an equivalent elastic modulus. In the 2000 tests the damping ratio roughly doubled, from 0.82% to 1.59–1.93%, after the joints between the modules had been sealed with silicone.
With the published module mass and the measured frequency and damping, Den Hartog's formula puts the whole column's galloping threshold at 96–103 m/s for wind 5° off the faces and 207–222 m/s for wind square to them; Solari obtained 95 and 217 m/s. The shape helps too: the modules, narrowing and widening every 90 centimetres, make the vortex wake irregular, with its frequencies spread over a band. Around 3 m/s, vortex shedding can excite the column's first mode, but its mass and damping, a Scruton number of 119, keep the motion small. René Gabbai calculated that even a round or square column with the same mass and damping would stay stable. The Column's shadow, over a year, is on another page.
Sources and method
The model divides the spine into 116 beam elements, fixed at the base, each with the section from the 1937 drawings as given by Solari: four angles, outer and inner plates thickening towards the base, concrete up to 5.50 m (counted as mass only). The steel's elastic modulus is taken as 206 GPa. The batten plates, rivets and bolts are not itemised; their mass is estimated at 0–30% of the spine's steel mass. Each full cast-iron module weighs 0.86 t, according to the UNESCO nomination dossier. Den Hartog's formula gives the critical wind speed for galloping: four times the equivalent mass per metre (weighted by the mode shape, (z/h)^1.75), the damping ratio and the angular frequency, divided by the air density (1.25 kg/m³), the reference width and the absolute value of the aerodynamic coefficient from the tunnel. The formula assumes a single mode and wind uniform up the height.
- Giovanni Solari, “Brâncuși Endless Column: A Masterpiece of Art and Engineering”, International Journal of High-Rise Buildings 2(3), 2013, pp. 193–212. ctbuh.org · koreascience.kr The spine's sections (p. 198), the frequencies and damping measured in 2000, the tunnel coefficients, the galloping thresholds, the design wind.
- Dan Lungu (Technical University of Civil Engineering, Bucharest) and colleagues, 2005: the vibration measurements of November 2000, cited by Solari. The periods of 1.205 s, 0.268 s and 1.949 s.
- UNESCO nomination dossier for the Târgu Jiu ensemble, p. 13 (table). whc.unesco.org The mass of a full module: 0.86 t.
- René D. Gabbai, “On the Aeroelastic Indifference of Brancusi's Endless Column”, Journal of Engineering Mechanics 134(6), 2008, pp. 462–465. nist.gov The stability of round or square columns with the same mass and damping.