Study On Vibration Control Method Of Piezoelectric Intelligent Plate Based On Neural Network-Genetic Algorithm

Vibration active control is a new and developing subject in recent two decades.It mainly researches the theories,the methoes and applications of vibration active control. Active vibration control for intelligent structures containing piezoceramics is a new and challenging field in the project of vibration control.This thesis mainly deals with reasonable mathematical models, identification and active vibration control of cantilever plate. Neural networks control method appling to vibration control of cantilever plate and building a vibration active control system based on neural network-genetic algorithm.In the study,the flexural rigidity of composite plate which includes the stiffness of piezoelectic material is derived based on fist order shear deformation theory.The dynamic equation of a plate with surface bonded or piezoelectics embedded is deduced by using Hamilton's principle.In order to get large driving force and efficient sensor signal, according to piezoelectric laminated plate(beam) modal theory of optimal position theory, actuators and sensors are placed at the locations where the strain is local maximum along the plate.Through numerical examples verify the effectiveness of mechanical modeling method.In the thesis the vibrating characters of piezoelectric laminated plate are analyzed, a new active control system, the genetic algorithm-neural network system is proposed. In this system the genetic algorithm(GA) is used to determine the control force and the neural network(NN) is applied to simulate vibrating characters of laminated plate. The system well combines the GA controller and NN identifier. It is a new control system that has a great future.