Neural network procedures were explored with the objectives to control a delta wing in the non-linear domain. Results show that neural networks are well suited to represent non-linear computed lift and pressure forces on delta wings. A model problem, the flare maneuver, suggests that neural network controllers can perform well for non-linear systems provided that suitable input and training data are provided to it.

Neural network procedures were explored with the objectives to control a delta wing in the non-linear domain. Results show that neural networks are well suited to represent non-linear computed lift and pressure forces on delta wings. A model problem, the flare maneuver, suggests that neural network controllers can perform well for non-linear systems provided that suitable input and training data are provided to it.

Experimental study of thrust-induced effects on a pitching-up 60 deg. delta wing was conducted. A downward vectored trailing edge jet has a significant effect in delaying the vortex breakdown. Strong asymmetric vortex bursting can be induced using asymmetric jet control. Detailed PIV measurements are taken to characterize the vortex flow field with and without jet control. A parallel experiment of the vortex flow over a 75 deg. delta wing is made to further the understanding of th vortex breakdown phenomenon. Related investigations of dynamic stall phenomena over a pitching airfoil are also summarized. Concurrently, considerable efforts have been made in the development of an innovative computational scheme. A grid-free vortex distribution technique is introduced to simulate the diffusion process of a vortex-dominated flow. This allows the readily extension of the 2-D vortex method to be used in a 3-D flow field such as a pitching-up delta wing configuration. Several computational examples will be discussed.