The aerodynamic properties of tennis balls.

Several experimental procedures were developed to enhance the understanding of the
aerodynamic properties of tennis balls. Four test methods were tried as quantitative
assessments of the aerodynamic forces that act on tennis balls, whilst an additional two
methods were introduced for qualitative purposes. A computational trajectory model was
developed to predict the effect of any modifications to tennis balls proposed in the study.
The test methods adopted utilised two different wind tunnels, projection devices, dropper
devices, aerodynamic load cells and motion analysis techniques using high-speed digital
cameras. Several different tennis balls were tested: some had the nap modified to
investigate changes in aerodynamic forces that may occur during play, others were
oversized to investigate the options available for slowing the game down.
CD and CL profiles were obtained for a normal sized ball with unmodified nap and then
used to develop a set of equations that enable the CD and CL of a tennis ball to be
calculated at any speed and spin rate. When used in a trajectory model, a 6.5% larger ball
was shown to decelerate 5% faster than a normal sized ball when projected with the same
initial elevation angle, speed and spin rate. This results in the larger ball landing 1.5 metres
shorter and taking more than 19ms longer to arrive at the receiver.
Initial testing showed that the CD of all tennis balls with unmodified naps was similar and
remained constant at around 0.53 up to a wind speed of around 63ms'. The nap of the
tennis ball was modified to represent early wear characteristics (fluffed) and extensive wear
characteristics (shaved). It was found that the CD of a ball with a fluffed nap is higher than
that of a ball with an unmodified nap, which in turn is greater than the CD of a ball with a
shaved nap. The CD of a ball almost twice the size of a normal tennis ball was found to be
independent of Reynolds number up to 5x105, which is clear evidence that the boundary
layer around a tennis ball turns turbulent at a low Reynolds number. The ball with the
shaved nap was shown to be similar to a classic rough ball however, with boundary layer
transition occurring at a low Reynolds number.
The flow around a tennis ball was assessedu sing pressurep rofiles and smokep articles, and
the separation of flow for all balls was shown to be near the poles. Pressure profile testing
provided clear separation details, and showed how the pressure around the ball differs for
subcritical and postcritical Reynolds number regimes. Flow through and over the fibres
causes the elevated CD over and above that associated with separation at the apex of a
sphere.