Robust vibration control of smart composite beams using neural networks was studied. Linear quadratic Gaussian with loop transfer recovery (LQG/LTR) methodology was used to design a robust controller on the basis of the state space model of the system. The state space model of the system was obtained using the finite-element method and mode superposition. The finite-element model was based on a higher-order shear deformation theory which included the lateral strains. The mode superposition method was used to transform the coupled finite-element equations of motion in the physical coordinates into a set of reduced uncoupled equations in the modal coordinates. The performance of the LQG/LTR controller was verified for various arbitrary initial conditions. A system of neural networks was then trained to emulate the robust controller. The neural network system was trained using the backpropagation algorithm. After suitable training, the NN (neural network) controller was shown to effectively control the vibrations of the composite beam. A robustness study including the effects of tip mass, structural parameter variation, and loss of a sensor input was performed. The NN controller is shown to provide robustness and control capabilities equivalent to that of the LQG/LTR controller.

A modal dynamic model was developed for the active vibration control of laminated doubly curved shells with piezoelectric sensors and actuators. The dynamic effects of the mass and stiffness of the piezoelectric patches were considered in the model. Finite element equations of motion were developed based on shear deformation theory and implemented for an isoparametric shell element. The mode superposition method was used to transform the coupled finite element equations into a set of uncoupled equations in the modal coordinates. A robust controller was developed using Linear Quadratic Gaussian with Loop Transfer Recovery (LQG/LTR) design methodology to calculate the gain and actuator voltage requirements. A Neural Network controller was then designed and trained offline to emulate the performance of the LQG/LTR controller. Numerical results have been presented for a flat plate and a spherical shell showing the variation in initial conditions and structural parameters. The neural network controller was shown to effectively emulate the LQG/LTR controller.

No abstract