电动缸举升伺服机构动力学建模与参数辨识

Modeling & identification of EMA lifting servo mechanism

中文关键词  电动缸举升机构  正交投影定理  记忆区间  循环最小二乘法

英文关键词  EMA lifting mechanism  Orthogonal projection theorem  Memory interval  Cyclic least square method

中文摘要

      针对电动缸举升伺服机构动力学参数辨识困难的问题, 本文提出了一种限定记忆区间的循环最小二乘辨 识法, 用于机构全行程内非线性模型、扰动参数的辨识拟合. 首先, 分析了机构完整模型的非线性参数成因, 并提出 了相应的简化模型; 然后, 分析了最小二乘法在平均位姿下的参数近似辨识特性, 并提出了限定记忆区间的循环最 小二乘法; 最后, 通过构造输入输出信号和选择辨识伺服环路, 使得辨识过程不会超过机构的行程范围. 实验结果表 明: 所提辨识法使得正弦速度和位移响应均方根误差相比最小二乘法分别下降了59.3%和84.2%. 所提辨识法可给 电动缸举升伺服机构的高精度动力学建模, 控制器结合观测器的复合控制策略的实现提供参考.

英文摘要

      To solve the problem of difficult identification of dynamic parameters of electric mechanical actuator (EMA) lifting servo mechanism, a cyclic least square identification method with limited memory interval is proposed, which is used to identify and fit the nonlinear model and disturbance parameters of the mechanism in the whole stroke. Firstly, the cause of nonlinear parameters of the complete mechanism model is analyzed, and the corresponding simplified model is proposed; Then, the parameter approximate identification characteristics of the least squares method under the average pose are analyzed, and the cyclic least squares method with limited memory interval is proposed; Finally, by constructing the input and output signals and selecting the identification servo loop, the identification process will not exceed the travel range of the mechanism. The experimental results show that the proposed identification method makes the root mean square error of sinusoidal velocity and displacement response reduce by 59.3% and 84.2% respectively compared with the least square method. The proposed identification method can provide reference for high-precision dynamic modeling of EMA lifting servo mechanisms and the implementation of composite control strategies of controllers and observers