An adaptive neurocontroller for vibration suppression and shape control of a flexible beam

This paper presents research into an adaptive nonlinear neural network control algorithm that can be used with smart structure actuators and sensors to control the shape and suppress the vibrations of flexible beams. The algorithm presented couples an explicitly model-based adaptive component, which employs time-varying estimates of the beam material properties, with an adaptive neural network, learning in real-time, which is used to estimate the additional actuator moments needed to offset the effects of periodic external disturbances. Experimental results are given for a thin beam, with bonded piezoceramic sensors and actuators, demonstrating the ability of the algorithm to track desired bending profiles and reject the vibrations caused by external disturbances, as well as to maintain this performance despite changes in the material properties of the structure or in the properties of the external disturbance.