磁流变阻尼器逆向模型的建模、优化与仿真

Modeling,Optimization and Simulation of a Magnetorheological Damper Inverse Model

在材料特性实验台上对磁流变(magnetorheological,简称MR)阻尼器的阻尼特性进行了测试,识别了BoucWen模型的未知参数。利用BP神经网络技术建立了MR阻尼器非线性逆向模型,并利用遗传算法高效的全局优化能力对MR阻尼器神经网络模型的结构、权值和阈值进行优化。将所建逆向模型应用于铁道车辆的半主动振动控制中进行仿真。分析结果表明,优化后的神经网络模型预测精度和泛化能力均得到显著提升,半主动控制效果明显,验证了该优化方法的有效性。

The damping characteristics of Magnetorheological(MR)dampers is tested on material testing system(MTS),and the unknown parameters of Bouc-Wen model are identified.A nonlinear inverse model of MR dampers is built by using BP neural network technology.In order to improve prediction accuracy and generalization ability of the inverse model,the structure,weights and threshold values of the model are optimized by using genetic algorithm(GA)theories for their rapid local searching ability.The inverse model is applied in the semi-active control system of railway vehicle for simulation analysis.The simulation results show that the prediction accuracy and generalization ability of the inverse model optimized are improved significantly.The vibration is controlled effectively,and the optimization method is valid.