Research On Design And Application Of Model Parameter Adaptive ESO Based On Frequency Domain Index

As a class of modeless control scheme with strong adaptability and strong antiinterference ability,Active Disturbance Rejection Control technology observes the total disturbance other than double-integral and compensates it through extended state observer.Regardless of the form of the external disturbance and the plant model,it compensates the plant as double-integral form.Therefore,the performance of ADRC is directly determined by the accuracy of the total disturbance observation of ESO.The ESO design method is complex and the actual model of the plant is very different from double-integral.High-frequency components and noise are introduced into the system when ESO work.In order to achieve better disturbance observation and compensation effect,this thesis proposes a novel parameter tuning method of the ESO,and improves the classic ESO structure using adaptive adjustment method.The main work of the thesis includes the following aspects.Firstly,we propose a parameter tuning method of the extended state observer,which can expand the disturbance observation bandwidth while suppress the measurement noise effectively.The equivalence relation between the classical perturbation observer and the linear extended state observer is deduced,and the parameter tuning problem of the third-order ESO is transformed into the design problem of a low pass filter.Then a Chebyshev filter is employed to improve the performance of the ESO.The parameter design method is extended to the nonlinear ESO.The simulation results show that compared with the existing parameter tuning method,the proposed method endows the ESO with better noise suppression ability and observation performance.Secondly,two improved methods of ESO structure are proposed.The structure and parameters of ESO based on model information is designed,and the simulation results show advantages in reducing compensation rate and improving control performance.Modified recursive empirical Frequency-domain optimal parameter estimation method is applied to get model information as an online identification algorithm,and the parameters of ESO based on model information can be adaptively adjusted with information from MREFOP method,which is called as model parameter-based adaptive ESO.The simulation results show its effectiveness and advantages.Finally,the ESO design methods proposed in this thesis,combined with a frequency domain method,is applied to load simulator for validation.In order to achieve better control performance,plant characteristics and the main control issues of load simulator are analyzed.Then the design of the linear ESO,non-linear ESO and the ESO based on model information is completed which,together with frequency domain control scheme,are applied to the actual load simulator system respectively.The comparative experiments results show that the proposed one,compared with the existing ESO designing method,endows the ESO with better noise suppression ability and observation performance with less perturbation compensation.Moreover,the performance of load simulator is improved.