Abstract: Aiming at the problem of inaccurate identification of nonlinear disturbance parameters in electric cylinder(EMA) lifting servo mechanisms using the limited memory interval cyclic least squares(CLS) method, a limited memory interval cyclic least squares method based on disturbance decoupling(DCLS) is proposed. Firstly, the complete and simplified dynamic model of the mechanism, as well as the reasons why nonlinear parameters are difficult to identify, are introduced. Then, the limitations of the CLS identification method in identifying and fitting nonlinear disturbance parameters are analyzed, and the DCLS identification method isproposed through the unbalanced torque separation testing method based on dynamic balance. Finally, based on the experimental platform, excitation signals, and identification loop, precise identification of nonlinear models and disturbance parameters within the mechanism's travel range is achieved. The experimental results show that the root means square errors of the speed and displacement responses identified by the DCLS identification method have decreased by 47.6% and 48.9% respectively compared to the CLS identification method. The DCLS identification method can solve the problem of mutual coupling when identifying disturbance parameters using the CLS identification method, and provide reference for high-precision modeling of EMA servo lifting mechanisms.

Key words: EMA lifting mechanism, disturbance decoupling, memory interval, cyclic least square method