Neural Network Based Adaptive Flow Control for Maneuvering Vehicles

Abstract : This report documents a phase I STIR effort with the objective of developing and demonstrating effective nonlinear adaptive control of the aerodynamic flow about a dynamic body using a distributed array of synthetic jets for actuation. Design of a wind-tunnel test apparatus is presented. Motion of the model is constrained to two degrees of freedom. A conventional elevator is used to trim the model and change its dynamic characteristics. Position control of the model is achieved by an adaptive outer loop controller. This outer loop commands the flow control actuators. A dynamic simulation model of the wind tunnel apparatus is presented, as are designs for both the inner and outer loop controllers. The outer loop design is adaptive. A non-minimum phase transfer function is presented to model the active flow control actuators, and includes possible coupling effects between actuation, the dynamics of flow field, and the rigid body dynamics of the model. The outcomes of simulation studies are presented. The parameters were selected to have an adverse effect on the closed loop response, therefore representing a hypothetical worst-case situation. These results demonstrate successful adaptive control of the simulated wind tunnel test article employing flow devices for actuation.