论文部分内容阅读
针对液压伺服系统常规“白箱”建模由于参数无法精确获得导致所得模型精度不高及“黑箱”建模所得模型内部结构未知的问题,本文提出基于ODE参数辨识的液压伺服系统“灰箱”建模。首先,建立了工程上实用的系统状态空间模型,根据系统特征确定了待辨识参数,将模型辨识问题转化为常微分方程(ODE)参数辨识问题;然后,采用正弦扫频信号作为激励信号和基于带边界约束的信赖域优化算法的初值问题方法进行参数辨识;为了和ODE参数辨识结果进行对比,本文同时采用系统的频率响应数据和最小二乘法辨识得到系统的“黑箱”传递函数模型;最后,通过大量实验验证了辨识模型的精确度。实验结果表明,本文提出的基于信赖域算法的液压伺服系统模型辨识方法可以有效处理参数的边界问题,使辨识模型既具有实际的物理意义,又与实际系统高度符合。
Because of the inaccuracy of the parameters, the accuracy of the obtained model is not high enough and the internal structure of the model obtained by the “black box” modeling is unknown, this paper presents a hydraulic servo system based on ODE parameter identification “Gray box ” modeling. Firstly, an engineering practical system state-space model is established. According to the characteristics of the system, the parameters to be identified are identified, and the problem of model identification is transformed into the problem of ODE parameter identification. Then, the sinusoidal sweep signal is used as the excitation signal and based on In order to compare with the ODE parameter identification results, the system’s “black box” transfer function model is identified by using the system’s frequency response data and the least squares method at the same time, and the initial value problem of the trust region optimization algorithm with boundary constraints is identified. Finally, a large number of experiments verify the accuracy of the identification model. The experimental results show that the proposed model identification method of hydraulic servo system based on trust region algorithm can effectively deal with the boundary problem of parameters, so that the identification model has both practical physical meaning and high conformity with the actual system.