A generalizing fuzzy model for shallow cavity flows under different Mach regimes

Abstract

Modeling of a flow passing over a shallow cavity is an interesting problem as the internal dynamics are inextricably intertwined due to the Navier-Stokes equations. The origin of the cavity flow identification problem is the desire for reducing the skin friction in aerial vehicles through an active control effort. As there is not a well developed closed-loop scheme, towards the goal of feedback control, the first step becomes to develop a suitable dynamic model imitating the behavior of the plant under certain operating conditions. For this purpose, we present an approach exploiting the fuzzy inference mechanisms. Fuzzy logic is a practical tool for expressing human expertise in the form of if-then statements. The idea in the fuzzy identification is to perform local observations from the flow field and to find a fuzzy interpolation scheme over the sensory information. The results we have observed indicate that a classical non-adaptive fuzzy model is able to perform one step ahead prediction of the critically essential behavior observed at the cavity floor.