【作者】 黄治潭；

【导师】 张孝祖；

【作者基本信息】 江苏大学 ， 车辆工程， 2003， 硕士

【摘要】 车辆悬架系统的设计必须适应行驶平顺性和操纵稳定性的要求，传统的被动悬架系统已无法从结构设计上使车辆具有良好的平顺性，而半主动悬架由于其悬架参数如阻尼、弹簧刚度等具有可调性，使得它可以很好地满足车辆行驶过程的需要，同时与主动悬架相比，由于工作中几乎不消耗发动机的功率，结构简单，造价较低，因此受到车辆工程界的广泛重视。 本文建立了4自由度1／2车辆双轴非线性模型，推导出了阻尼可变形式的系统状态方程，然后提出了一种新的半主动悬架控制方法，即模糊神经网络控制，它采用BP神经网络离线训练模糊控制器的规则，神经网络学习的结果蕴涵了模糊控制规则。采用Matlab软件中的Similink及其S函数编程，通过计算机仿真验证了这种控制算法的有效性，并与被动悬架的性能指标进行了对比，证明采用本文的半主动悬架的控制算法进行仿真，确实能提高车辆悬架的综合性能。最后进行了实车试验，验证了本文的仿真程序是正确的。更多还原

【Abstract】 The design of vehicle suspension system must be adapted to the requirement of the ride comfort and the control stability. It’s difficult for traditional passive suspension to enhance the ride comfort. However, semi-active suspension can meet the ride comfort well because its suspension parameters such as damp coefficient or spring firmness can be adjustable.Compared with active suspension, semi-active suspension is more inexpensive and has simpler structure. Moreover, it hardly consumes engine power when it works. Therefore, semi-active suspension has been attached much importance by the automobile engineering industry.This paper establishes a 1/2 nonlinear double axles model which consists of 4 degrees, then deduces its state equation when its damping coefficient is adjustable. Then it puts forward a new way of fuzzy neural network(FNN) control to adapt the damp of the semi-active suspension. This control method applies BP neural network to train the rule of FNN off line and the result of the BP neural network contains the rule of FNN.The author use the software Matlab and Simulink as well as it’s S-function to program and finally validates the efficiency of the control arithmetic through simulation. Compared with passive suspension, semi-active suspension can enhance the integrate performance of the vehicle.Finally, the author did the experiments to validate that the program of the simulation is right.更多还原