Paul N. Roschke

Neural Network Modeling of a Magnetorheological Damper

The magnetorheological (MR) damper is a newly developed semiactive control device that possesses unique advantages such as low power requirement and adequately fast response rate. The device has been previously tested in a laboratory to determine its dynamic properties and characterized by a system of nonlinear differential equations. This paper presents an alternative representation of the damper in terms of a multilayer perceptron neural network. A neural network model with 6 input neurons, one output neuron and twelve neurons in the hidden layer is used to simulate the dynamic behavior of the MR damper. Training of the model is done by a Gauss-Newton based Levenberg-Marquardt method using data generated from the numerical simulation of the nonlinear differential equations. An optimal brain surgeon strategy is adopted to prune the weights and optimize the neural networks. An optimal neural network is presented that satisfactorily represents dynamic behavior of the MR damper.

Robust Control for Structural Systems with Parametric and Unstructured Uncertainties

Tall, slender structures and long bridges inherit numerous uncertainties due to model errors, stress calculations, material properties, and load environments and may undergo large forces from natural hazards such as earthquakes and strong wind events. This paper develops a robust active control approach with para metric uncertainty in the system and control input and unstructured uncertainty in the disturbance input ma trices based on an uncertain structural system. A special single-valued decomposition (SVD) is applied to structured uncertain structures. The robust control law provides robust relative stability, an H ∞-norm distur bance attenuation, and H 2 optimality. The H ∞ norm of the transfer function from the external disturbance forces (e.g., earthquake, wind, etc.) to the observed system states is restricted by a prescribed attenuation in dex δ. Preservation of robust H 2 optimality of uncertain structural systems is discussed. This paper considers both structured uncertainties and norm-bounded unstructured uncertainties. Numerical simulations that use the robust controller show significant reduction in vibrations.

Neuro-fuzzy control of structures using acceleration feedback

This paper described a new approach for the reduction of environmentally induced vibration in constructed facilities by way of a neuro-fuzzy technique. The new control technique is presented and tested in a numerical study that involves two types of building models. The energy of each building is dissipated through magnetorheological (MR) dampers whose damping properties are continuously updated by a fuzzy controller. This semi-active control scheme relies on the development of a correlation between the accelerations of the building (controller input) and the voltage applied to the MR damper (controller output). This correlation forms the basis for the development of an intelligent neuro-fuzzy control strategy. To establish a context for assessing the effectiveness of the semi-active control scheme, responses to earthquake excitation are compared with passive strategies that have similar authority for control. According to numerical simulation, MR dampers are less effective control mechanisms than passive dampers with respect to a single degree of freedom (DOF) building model. On the other hand, MR dampers are predicted to be superior when used with multiple DOF structures for reduction of lateral acceleration.

Fuzzy Control of Base-Isolation System Using Multi-Objective Genetic Algorithm

Smart base-isolation strategies are being widely investigated as a way to reduce structural damage caused by severe loads. This study uses a friction pendulum system (FPS) as the isolator and a magnetorheological (MR) damper as the supplemental damping device of a smart base-isolation system. Neuro-fuzzy models are used to represent dynamic behavior of the MR damper and FPS. A fuzzy logic controller (FLC) is used to modulate the MR damper so as to minimize structural acceleration while maintaining acceptable base displacement levels. To this end, a multi-objective optimization scheme that uses a nondominated multi-objective genetic algorithm (NSGA-II) is used to optimize parameters of membership functions and find appropriate fuzzy rules. To demonstrate the effectiveness of the proposed multi-objective genetic algorithm for FLC, a numerical study of a smart base-isolation system is conducted using several historical earthquakes. The findings show that the proposed method can find optimal fuzzy rules and that the NSGA-II-optimized FLC outperformed a passive control strategy, a conventional semiactive control algorithm and a human-designed FLC.

A comparative study in the semi-active control of isolated structures

A comparative analytical and experimental study of several algorithms for the control of seismically excited floor- and base-isolated structures is pursued in the current study. A hybrid isolation system that is comprised of a bidirectional roller–pendulum system (RPS) and augmented by controllable magnetorheological (MR) dampers is proposed to reduce the potential for damage to structures and sensitive equipment. Bidirectional motions are intelligently ameliorated in real time by the modulation of MR damper resistance. A Bouc–Wen model is adopted in numerical and experimental trials to predict behavior of the MR dampers. Three contrasting control techniques are examined. They include neural network control, LQR/clipped optimal control with variable gains and fuzzy logic control. Each control scheme is a candidate for mitigating the response of a superstructure to near- and far-field seismic loadings. Minimization of displacement and acceleration responses of the structure are considered in the formulation of each approach to control. Results of the numerical and large-scale experimental efforts reveal that the response of the isolated structure is effectively alleviated by all of the considered control methods, although they do not perform equally well. The LQR/clipped optimal controller with variable gains is superior to the other controllers in 50% of the investigated cases, while the fuzzy logic controller performs well for earthquakes with large accelerations. Neural network control is found to be effective in minimizing the acceleration of the superstructure that is subject to moderate excitation.

GA-based optimum design of a shape memory alloy device for seismic response mitigation

Damping systems discussed in this work are optimized so that a three-story steel frame structure and its shape memory alloy (SMA) bracing system minimize response metrics due to a custom-tailored earthquake excitation. Multiple-objective numerical optimization that simultaneously minimizes displacements and accelerations of the structure is carried out with a genetic algorithm (GA) in order to optimize SMA bracing elements within the structure. After design of an optimal SMA damping system is complete, full-scale experimental shake table tests are conducted on a large-scale steel frame that is equipped with the optimal SMA devices. A fuzzy inference system is developed from data collected during the testing to simulate the dynamic material response of the SMA bracing subcomponents. Finally, nonlinear analyses of a three-story braced frame are carried out to evaluate the performance of comparable SMA and commonly used steel braces under dynamic loading conditions and to assess the effectiveness of GA-optimized SMA bracing design as compared to alternative designs of SMA braces. It is shown that peak displacement of a structure can be reduced without causing significant acceleration response amplification through a judicious selection of physical characteristics of the SMA devices. Also, SMA devices provide a recentering mechanism for the structure to return to its original position after a seismic event.

Active control of a tall structure excited by wind

Abstract A control strategy is developed for application to a flexible laboratory structure excited by simulated wind forces for the purpose of minimizing along-wind accelerations. Static and dynamic characteristics of the structure are identified through a modal analysis method that formulates a linear model of the system. Actual wind speed data is used to produce a simulated wind loading by means of drag forces. An LQG/LTR control strategy based on acceleration feedback is used in conjunction with a magnetorheological (MR) damper to reduce structural response. When a strong wind loading is applied to the structure, the control force notably reduces simulated peak floor accelerations.

GA-fuzzy control of smart base isolated benchmark building using supervisory control technique

The effectiveness of a supervisory fuzzy control technique for reduction of seismic response of a smart base isolation system is investigated in this study. To this end, a first generation, base isolated, benchmark building is employed for numerical simulation. The benchmark structure under consideration has eight stories and an irregular plan. Furthermore it is equipped with low damping elastomeric bearings and magnetorheological (MR) dampers for seismic protection. The proposed control technique employs a hierarchical structure of fuzzy logic controllers (FLC) consisting of two lower-level controllers (sub-FLC) and a higher-level supervisory controller. One sub-FLC has been optimized for near-fault earthquakes and the other sub-FLC is well-suited for far-fault earthquakes. These sub-FLCs are optimized by use of a multi-objective genetic algorithm. Four objectives, i.e. reduction of peak superstructure acceleration, peak isolation system deformation, RMS superstructure acceleration and RMS isolation system deformation are used in a multi-objective optimization process. When an earthquake is applied to the benchmark building, each of the sub-FLCs provides different command voltages for the semi-active controllers and the supervisory fuzzy controller appropriately combines the two command voltages based on a fuzzy inference system in real time. Results from numerical simulations demonstrate that isolation system deformation as well as superstructure responses can be effectively reduced using the proposed supervisory fuzzy control technique in comparison with a sample clipped optimal controller.

Neuro-fuzzy control of structures using magnetorheological dampers

The paper describes a novel approach for reduction of environmentally induced vibration in constructed facilities by way of a neuro-fuzzy technique. The control technique is presented and tested in a numerical study that involves two types of building models. The energy of each building is dissipated through magnetorheological (MR) dampers whose damping properties are continuously updated by a fuzzy controller. This semi-active control scheme relies on development of a correlation between accelerations of the building (controller input) and voltage applied to the MR damper (controller output). This correlation forms the basis for development of an intelligent neuro-fuzzy control strategy. To establish a context for assessing the effectiveness of the semi-active control scheme, responses to earthquake excitation are compared with passive strategies that possess similar authority for control. According to numerical simulation, MR dampers are less effective control mechanisms than passive dampers with respect to a single degree of freedom (DOF) building model. On the other hand, MR dampers are predicted to be superior when used with multiple DOF structures for reduction of lateral acceleration.

Design, fabrication, testing, and fuzzy modeling of a large magnetorheological damper for vibration control in a railcar

This paper presents the procedure used for design, fabrication, testing, and numerical modeling of a magnetorheological (MR) damper that is to be applied for vibration control in a 70-ton railcar. MR dampers are semiactive vibration control devices whose damping characteristics can be modified in real time by varying an applied current. Design parameters for the MR damper are estimated from those exhibited by a linear viscous damper that exerts the necessary force required to limit vertical vibrations of the rail truck within acceptable limits. An MR damper is fabricated by modifying the piston of a standard hydraulic damper to function as a solenoid. The assembled MR damper is tested in a uniaxial testing machine by subjecting it to sinusoidal and random displacements while simultaneously varying the current flowing in the solenoid. A variable magnetic field is applied to the MR fluid that fills the damper cavity and the resisting force exerted by the damper is recorded. Data collected in the laboratory are used to train a fuzzy model of the MR damper that characterizes its behavior. Results indicate that a fuzzy model of the MR damper can predict its behavior with a sufficient degree of accuracy while requiring minimal computational time.