Abstract: In order to identify hysteresis characteristics of piezoelectric actuators in the processes of applications, the hybrid modeling method of Preisach model and LS-SVM was adopted. Initial weighted coefficients in α-β plane were calculated based on Preisach model  through first order reverse curves. Centroid of hysteresis units and initial weighted coefficients were used as inputs and outputs of the LS-SVM which optimized by genetic algorithm. The finer weighted coefficients were calculated based on the optimized model and desired hysteric model was set up. Through the contrast between results of simulated model and the experimental results, errors stay between 0.12%~2.01%. Results show that the established model may accurately describe the hysteresis non-linearity of piezoelectric actuators and the proposed method is a practicable method.

Key words: piezoelectric actuator, hysteresis characteristics, parameter identification, least squares support vector machine(LS-SVM)

摘要： 为了辨识压电作动器应用中迟滞非线性特性，采用Preisach模型与最小二乘支持向量机（LS-SVM）的混合建模法构建压电作动器模型以表征其迟滞非线性。在Preisach模型基础上采用一阶回转曲线法求得α-β平面内的迟滞单元加权系数，以迟滞单元形心和加权系数作为训练模型的输入输出，利用遗传算法对模型参数进行了辨识。数值仿真验证了建立的模型能精确描述压电作动器迟滞非线性特性，仿真结果与实验结果的相对误差在0.12%~2.01%之间，证实了模型的有效性。

关键词: 压电作动器, 迟滞特性, 参数辨识, 最小二乘支持向量机