伺服系统弹性负载的闭环辨识方法EI北大核心CSCD

为有效解决机械谐振问题,伺服系统弹性负载的辨识是非常关键的步骤.本文以工业中最常见的双惯量系统作为辨识对象设计闭环辨识方法,使用伪随机二进制序列作为激励并采集电机电流转速信号.在此基础上,使用最小二乘法拟合系统的自回归移动平均模型,并提高模型阶次以保证拟合精度.为抑制采样噪声的影响,提出基于平衡截断的模型降阶方法,根据Hankel奇异值大小判断系统阶次并提取主要模态.最后,通过仿真和实验进行验证,结果表明:相比于传统辨识方法,本文所提出的辨识方法能够有效抑制噪声干扰,具有更高的精度.

Closed-loop identification method for servo elastic load

The identification for the servo elastic load is an essential step to solve mechanical resonance problem.This paper designs a closed-loop identification method for the two-mass system, which is the most common in industrialapplication. The current and velocity signals of the motor are collected while the pseudo-random binary sequence is usedto stimulate system. On this basis, the least squares method is applied to fit the auto-regressive and moving average model,using a higher fitting order to ensure the accuracy. In order to suppress the influence of sampling noise, a balanced truncationbased model reduction method is proposed, which judges the order of the system and extracts dominant states according tothe Hankel singular value. In the end, the proposed method is verified by simulation and experiment. The results show that:compared with the traditional identification method, the proposed identification method can effectively suppress noise andhas higher accuracy.