Dalrymple and Martin (1996) showed that the plane wave approximation in the channel is good
for most cases, which can provide a reasonably accurate solution with far less computational effort
than full wave solutions. For a normal incident wave, there is a slight difference between full wave
solution and plane wave approximation in the gap or the slit, so that we can neglect all evanescent
waves in the gap or the slit.
According to plane wave approximation, the velocity potential in the gap can be denoted as: