Abstract: In order to improve the accuracy of polynomial dynamics model of MRD，a polynomial modeling method was proposed based on PSO algorithm.MRD test platform was built.Traditional polynomial model and Chebyshev polynomial model were identified and compared by using the measured experimental data of mechanics properties.Combining PSO algorithm with Lagrange interpolation equation and measured data to optimize interpolation nodes，the polynomial dynamics model was constructed after the simplification of MATLAB Simplify function.The flow and main steps of PSO algorithm for optimizing interpolation nodes were researched，and the average cumulative relative errors between PSO algorithm and Chebyshev model were analyzed based on 12th order polynomial modeling.The results show that the relative errors of PSO polynomial modeling method is 47.0% less than that of Chebyshev polynomial model under sinusoidal excitation frequency of 1Hz，amplitude of 15mm and current of 0~1.5 A.The PSO polynomial modeling method may well reflect the dynamics characteristics of MRD and meet the needs of practical engineering applications.

Key words: magnetorheological damper（MRD), particle swarm optimization（PSO), Lagrange interpolation, fitness function, dynamics model

摘要： 为提高磁流变阻尼器（MRD）多项式动力学模型精度，提出了基于粒子群优化（PSO）算法的多项式模型建模方法。搭建了MRD试验平台，利用测得的力学特性数据，辨识并对比分析了传统多项式与Chebyshev多项式模型；运用PSO算法，结合Lagrange插值方程和实测数据优化插值节点，经MATLAB Simplify函数化简，构建了多项式动力学模型；研究优化插值节点的PSO算法流程及主要步骤，分析比较PSO算法与Chebyshev模型的12阶多项式建模平均累积相对误差。研究表明，在正弦激励频率1 Hz、振幅15 mm、电流0~1.5 A工况下，PSO多项式建模方法比Chebyshev多项式模型相对误差下降了47.0%，能较好地反映MRD动力学特性，满足实际工程应用需要。

关键词: 磁流变阻尼器, 粒子群优化, Lagrange插值, 适应度函数, 动力学模型