| High-speed railway plays an important role in China’s economic development. However, the increase of the train’s speed will amplify the train’s vibration significantly, which will induce an obvious decrease of the ride stability and ride quality. At the same time, as cars becoming an indispensable meaning of transportation in people’s travelling, people have made high demands on the cars’ NVH (Noise, Vibration, Harshness) property, which is directly related to the ride comfort of the cars. A suspension is one of the key components in improving the ride comfort and driving safety. Magnetorheological damper-based (MR damper-based) semi-active suspensions have become a hot point in the field of vehicle vibration control, because they possess similar performance with active suspensions and have some attractive characters such as simple structure, low energy consumption, fast response, and high fail-safe. However, the MR semi-active suspension technology is far from mature and some theoretical and practical issues still need to be further studied. In this dissertation, in order to reduce the vibration of cars and high-speed trains, the dynamic models of MR dampers and semi-active control strategies are studied by adopting the methods of theoretical analysis, numerical simulation and experimental test. The main contributions of this dissertation are as follows:1. Dynamic modeling of magnetorheological dampers. Currently, only a few dynamic models of MR dampers have considered the influences of the excitation properties and even fewer inverse models which consider the hysteresis nonlinearity and are suitable for the actual application. Thus, to meet the requirements of the actual application including model accuracy, simplicity, adaptability and reversibility, this work builds an excitation-current-dependent forward and inverse dynamic model for MR dampers, respectively. These could provide some bases for the design of effective semi-active controllers.The dynamic properties of two kinds of different structure and size of MR dampers, including an inner-pass MR damper and a by-pass MR damper, are tested by using MTS test system. The influences of the force-velocity properties on the control current and the excitation property are analyzed. Based on the test data, an extended hyperbolic tangent function-based model and a simplified hyperbolic tangent function-based model are proposed. Then, a hyperbolic tangent function-based model, a phenomenon model, an extended nonlinear hysteretic biviscous model, and a generalized sigmoid hysteresis model are established. The parameters of all models are identified by using the genetic algorithm (GA). Systematic comparisons of all models are carried out from the aspects of the model accuracy, complexity and reversibility. Results show that the extended hyperbolic tangent function-based model possesses the highest model accuracy, and the simplified hyperbolic tangent function-based model is easy to be reversed. The inverse model of the simplified hyperbolic tangent function-based model is analytically obtained. The adaptive neural-fuzzy inference system (ANFIS) is also adopted to establish the inverse model of MR dampers. Comparisons between the two inverse models are carried out from the aspects of the model accuracy and modeling difficulty.2. Researches on the semi-active control of automotive suspensions. As the MR damper-based vehicle suspension is a complex nonlinear system, the design of the semi-active controller is still a challenge. So some existing semi-active controllers are analyzed and improved, and then a LQG-Fuzzy semi-active controller is proposed. A generic simulation platform and experimental platform are established to assess the control effects of the established semi-active controllers. The results could provide theoretical and experimental bases for their application in automotive suspensions.A1/4vehicle suspension dynamic model and a road random irregularity model are established. An On-Off controller and a Fuzzy controller are designed, which are not necessary to know the precise model of the controlled objects. By adopting the inverse model of the simplified hyperbolic tangent function-based model, a Skyhook semi-active controller and a LQG semi-active controller are built, in which the hysteresis nonlinearity of MR dampers are considered. Combining with LQG control algorithm and fuzzy control algorithm, a LQG-Fuzzy controller is proposed. The control effects of all five controllers are assessed by simulation. The influences of the control effects on the variation of the sprung mass and the property of MR dampers are also studied. Then, a simulated1/4vehicle suspension experimental platform is established. The programs of all the controllers are compiled by using MATLAB software-based rapid control prototyping system. The experimental comparisons of the five semi-active controllers are carried out. Simulation and experiment results indicate that all the five semi-active controllers can suppress the acceleration of the sprung mass and the deflection of the suspension, and the proposed LQG-Fuzzy semi-active controller possesses relatively high robustness. 3. Researches on the robust semi-active control of high-speed trains. Because of the increase of the train’s speed, the robustness of the control system should be improved. The current control strategies are mainly concentrated in the field of classical control. The in-depth researches about robust control and hysteresis nonlinear suppression are still lack. So the H∞robust control algorithm is applied into high-speed trains in this section, and a H∞-ANFIS robust semi-active controller is proposed. This work provides a new method for the semi-active control of high-speed trains.A17-degree-of-freedom (DOF) model for a full-scale railway vehicle is developed and random track irregularities are modeled. By adopting the H∞robust control theory to calculate the expected control force and the ANFIS technology to calculate the required control current, a H∞-ANFIS semi-active controller for the high-speed train’s suspension is proposed. Moreover, an On-Off controller and a Fuzzy controller are also designed. The control effects of all three semi-active controllers are evaluated through simulation. The influences of the control effects on the time delay are studied, and the fail-safe of the MR damper-based semi-active suspension system is also analyzed. Results show that all the three semi-active controllers can significantly reduce the lateral vibration of the train’s car body, among which the H∞-ANFIS semi-active controller possesses the best control effect. However, the vibration of the bogies could be increased to some extent. The vibration of the wheelsets is almost the same as the passive suspension. With increasing of the delay time, the control effects of all three semi-active controllers reduce, and the time delay has the biggest impact on the H∞-ANFIS semi-active controller.This paper is focus on improving the deficiencies in the present magnetorheological semi-active suspension. Some forward models and inverse models of MR dampers are established, and a variety of semi-active controllers are designed. Simulation and experimental researches are carried out to valid these semi-active controllers. These studies can provide some theoretical and experimental bases for the application of MR dampers in the semi-active suspensions of vehicles. |
PDF Full Text Request