The effect of surface geometry on the aerodynamic behaviour of a football

AbstractSeveral studies have investigated the effect of surface features on the aerodynamic behaviour of a sphere or a football. Research on footballs typically compare real footballs with highly complex seam geometry and surface texture, making it difficult to identify which features influence the aerodynamic behaviour. This study commissioned many 3D printed footballs with regularly increasing seam length and surface texture to undertake a designed experiment to study these changes in a controlled fashion. Each ball was tested in a wind tunnel in non-spinning cases, or spinning about a vertical axis, at a range of speeds and key aerodynamic parameters were extracted from the data. Several methods were employed to characterise the roughness of each ball, and these roughness metrics were statistically tested for correlations with selected aerodynamic parameters. Using these relationships provides design guidance to football manufacturers to understand how modifying their surface geometry would influence the ball’s aerodynamic behaviour. In general, increasing the volume of roughness of the ball, measured as the change in volume of the ball introduced through seams or texture compared to a smooth sphere, decreased the critical Reynolds number. Balls with larger texture elements, particularly those with protrusive texture, had a much lower critical Reynolds number than other balls with the same absolute volumes of roughness. The post-critical drag coefficient did not significantly correlate with any of the roughness features of the balls. In a spinning case, the balls with high roughness generated a higher side force, this relationship plateaued at a certain level of roughness. The reverse Magnus behaviour changed significantly with the surface roughness; as the overall roughness volume of the ball increased, the Reynolds number at which the reverse Magnus changed to a conventional Magnus effect decreased. The large protrusive texture elements were effective at preventing a reverse Magnus effect from occurring at all in the tested Reynolds number range.