Title :
Neural control of the NASA Langley 16-foot transonic tunnel
Author :
Motter, Mark A. ; Principe, Jose C.
Author_Institution :
NASA Langley Res. Center, Hampton, VA, USA
Abstract :
Experimental results of controlling the Mach number in a transonic wind tunnel with a system of artificial neural networks are presented. Kohonen self-organizing maps are used to cluster the local tunnel dynamics and thereby predict the Mach number response to candidate control input sequences. The sequence minimizing the predicted error between the desired and actual Mach number is applied to the tunnel fan drive system. Comparison is made to gain scheduled automatic control currently in use
Keywords :
Mach number; dynamics; neurocontrollers; self-organising feature maps; wind tunnels; Kohonen self-organizing maps; Mach number; NASA Langley 16-foot transonic tunnel; candidate control input sequences; gain scheduled automatic control; local tunnel dynamics; neural control; predicted error minimisation; tunnel fan drive system; Artificial neural networks; Automatic control; Control systems; Delay lines; Error correction; History; NASA; Scheduling; Self organizing feature maps; Self-organizing networks;
Conference_Titel :
American Control Conference, 1997. Proceedings of the 1997
Conference_Location :
Albuquerque, NM
Print_ISBN :
0-7803-3832-4
DOI :
10.1109/ACC.1997.611883
