Intelligent Methods And Experiments For Nonlinear Structural Vibration Control

The current development of structural control has often been restricted to linear structures. However due to the inelastic deformation under intense ground shaking, it has theoretical and practical value to study the control of structural nonlinear vibration. The application of intelligent control algorithm in structural nonlinear vibration control is emphasized in this dissertation. The main research works are outlined as following:1.Aimed at controlling structural nonlinear vibration using interstory controller, an adaptive fuzzy sliding mode (AFSM) control algorithm is proposed to control structural nonlinear vibration, combining the advantage of sliding mode control and adaptive fuzzy control. Furthermore, the chattering phenomenon of sliding mode control is attenuated by an adaptive fuzzy system. The adaptation law is derived from Lyapunov direct method. To obtain all states of the structure, a dynamical neural network (DNN) observer is designed taking the advantage of dynamical neural network to approximate the arbitrary dynamic system. Numerical simulation is conducted on the nonlinear Benchmark structure. Robust analysis is also performed. The results show that AFSM control is suitable and robust for the control of structural nonlinear vibration. The dynamical neural network can also estimate the total states of structural nonlinear vibration, which leads to the output control of structural nonlinear vibration using the intelligent control algorithm.2.An active control force based on the AFSM control algorithm is analyzed and proved to be suitable for semi-active control device. Magnetorheological (MR) damper is used to trace the active control force using the developed semi-active control algorithm. Simulation of MR damper control using AFSM control is conducted on the nonlinear Benchmark structure. The control force is applied to each floor through an MR damper. The characteristic of active control force, semi-active control force and control effect are analyzed. The energy analysis is also performed. Furthermore, the placement of MR damper is optimized using a genetic algorithm. Based on the characteristic of semi-active control a two-phase optimization progress using genetic algorithm is developed and verified via numerical simulations. The optimal placement of MR damper for the control of structural nonlinear vibration is presented. On the other hand, considering fault tolerance control for structural nonlinear vibration, the dynamical neural network is used to identify the nonlinear structure with faults. Based on this dynamical neural network model, the corresponding fault tolerance controller is proposed to control structural nonlinear vibration and demonstrated by numerical simulations.3.To utilize active mass damper (AMD) to control the structural nonlinear vibration, a specified fuzzy controller is proposed and analyzed. To overcome the difficulty for excessive inputs to fuzzy controller, the generalized fuzzy input based on linear-quadratic cost is adopted and verified by the numerical simulation. The in-depth analysis of AMD control is performed considering the analysis of simulation results. The reasonable proposals for AMD control of structural nonlinear vibration are presented. Furthermore, the innovative AMD-ID (interstory damper) controller is developed to control structural nonlinear vibration. Numerical results verify the effectiveness of proposed controller.4.The experimental study of vibration control of a nonlinear beam with piezoelectric material considering geometric nonlinearity and model uncertainty is conducted. The geometric nonlinearity is theoretically and experimentally analyzed for large deflection of beam vibration. Additional masses are added to the tip of beam to realize the mass uncertainty. The general fuzzy control and adaptive fuzzy sliding mode control are compared using experimental data based on mass uncertainty. The experimental results show that adaptive fuzzy sliding mode controller can perform very well with nonlinearity and model uncertainty.5.To meet the demand on the cost and repeatability of test model of nonlinear structure, the MR rotary brake is proposed and used to mimic the plastic hinge to establish a nonlinear vibration model. The hysteretic loops of different nonlinear behaviors are realized through the control of voltage input to the MR rotary brake. Moreover, the MR damper is incorporated into the test model to implement the semi-active control. The performance of DNN observer and AFSM controller are verified. As a result, the test data demonstrates that the intelligent algorithms proposed by this dissertation are effective for the control of structural nonlinear vibration.