基于遗传算法的磁流变阻尼器Bouc-Wen模型参数辨识

Parameter identification of Bouc-wen model for MR damper based on genetic algorithm

在除试验数据外无任何先验知识的条件下如何识别Bouc-Wen模型的参数是一个亟待解决的问题。在对磁流变阻尼器进行力学特性试验的基础上,采用遗传算法对磁流变阻尼器Bouc-Wen模型进行参数辨识,并采用缩小参数取值范围的方法逐渐提高遗传算法的识别精度。通过分析参数值与电流间的特征曲线,采用曲线拟合的方法确定它们之间的函数关系,再利用遗传算法得到具体的函数表达式。最后用不同幅值和频率的激励试验数据对识别结果和拟合结果进行了验证。结果表明:利用缩小取值范围的方法得到的Bouc-Wen模型参数识别结果在全局最优解的附近,拟合结果和辨识出的模型均能满足要求。

A Bouc-Wen dynamic model of magnetorheological（MR） dampers was considered and its parameters were obtained by using genetic algorithm（GA） based on mechanical characteristic experiments of a MR damper.Identification precision of GA was improved by gradually shrinking linear ranges of these parameters.Functional relationships between some parameters and applied current were determined by using curve fitting through analyzing their characteristic curves.The specific function expressions were derived by using GA.The identification results and curve fitting results were verified with experiment data in different amplitudes and frequencies.It is shown that the identification results by the method proposed are close to global optimum values,and the model identified can meet the need of engineering applications.