Experimental and computational assessment of a shape memory alloy based morphing wing incorporating linear and non-linear control

2018, American Institute of Aeronautics and Astronautics Inc, AIAA. All rights reserved. The use of morphing wings can improve flight performance under specific flight conditions and therefore serve as viable concepts for improving aircraft performance. Shape memory alloy (SMA) composite actuators placed on the wing skin can generate the necessary strain to obtain a desired configuration. However, a closed-loop control scheme is necessary to fully explore the potential of morphing wings that utilize non-linear actuators. The performance of a PID and a reinforcement learning controller are explored for a morphing wing with an SMA composite actuator on the bottom surface. The controllers are used to obtain any desired configuration on the space of attainable options for an SMA-based, 3D printed morphing wing in a wind tunnel environment. Thermocouples and laser sensors are utilized to measure temperature and displacement that are key metrics in the control algorithms. Results show that a PID scheme is capable of obtaining all desired displacements regardless of wind tunnel velocity, and angle of attack. As for the reinforcement learning algorithm, the neural network approximately learned the constitutive model of the morphing wing after more than 1000 cycles. Therefore, both control schemes are valid for this application, but, according to preliminary results, PID is advantageous because it does not require extensive training.

Experimental and computational assessment of a shape memory alloy based morphing wing incorporating linear and non-linear control Conference Paper