MULTIVARIABLE NEURAL-NETWORK-BASED CONTROLLERS FOR SMART STRUCTURES

This paper details identification and robust control of smart structur es using artificial neural networks. To demonstrate the use of artific ial neural networks in the control of smart structural systems, two sm art structure test articles were fabricated. Active materials like pie zoelectric (PZT), polyvinylidene (PVDF) and shape memory alloys (SMA) were used as actuators and sensors. The Eigensystem Realization Algori thm (ERA), a structural identification method has been utilized to det ermine a minimal order discrete time state space model of the test art icles. The ERA requires the Markov parameters of the physical system. A neural network based method has been developed to estimate the Marko v parameters of a multi input multi output system from experimental te st data. The accelerated adaptive learning rate algorithm and the adap tive activation function were utilized to improve the learning charact eristics of the network and reduce the learning time. The identified m odels were used to design a robust controllers for vibration suppressi on of smart structures using a modified Linear Quadratic Gaussian with Loop Transfer Recovery (LQG/LTR) method. This control design methodol ogy has better loop transfer recovery properties while accommodating t he limited control force available from the SMA and the PZT actuators. This controller was copied into a feedforward neural network using th e connectionist approach. This neural network controller was implement ed using a PC based data acquisition system. The closed loop performan ce and robustness properties of the conventional and the neural networ k based controller are compared experimentally.