具全狀態回授之電子式馬達模擬器控制

Abstract

本論文提出將全狀態回授控制應用於電子式馬達模擬器之方法，其目的在於改善電子式馬達模擬器的電流暫態響應以提高馬達模擬器之準確度。傳統的馬達模擬器中，控制器參數由耦合電路之參數所決定，並未考慮到耦合電路在模擬馬達不同轉速時極點會跟著轉速變化，而使得控制器在高轉速時的響應會越來越差，甚至是失控。為了提高馬達模擬器的準確度與穩定性，需要更佳的控制策略抵銷轉速造成的影響。 為了改善傳統控制器的缺點，本論文提出基於電子式馬達模擬器的狀態空間模型設計的全狀態回授控制方法。藉由全狀態回授控制得到電流回授的補償增益，補償轉速跟電流之耦合的影響，使得馬達模擬器的極點位置固定在設計的位置。此方法之優點在於結合全狀態回授自由擺放極點之特性與比例積分控制器易設計之優點，簡化控制器設計流程並提升系統模擬精確度。本論文中詳細介紹並說明全狀態回授控制推導過程以及如何與比例積分控制器相結合，並且藉由電腦模擬以及硬體實驗結果，驗證本論文提出具全狀態回授的控制策略之表現。根據350V，1kW的硬體的實驗結果，與傳統PI控制策略相比，本論文所提出的具全狀態回授控制策略在不同的轉速下有一致的電流暫態與較短的安定時間表現。

In this thesis, the Full-State Feedback control, is proposed to improve the current dynamic response and to enhance the accuracy of the electric motor emulator (EME). The parameters of EME the controller are highly related to the characteristics of the coupling network in side the EME. When the EME operates at different rotating speeds, the poles of the transfer function of the coupling network will change. Usually, the drifting of the transfer function of the coupling networks is neglected. However, it will cause the current transient response to deviation under different rotating speeds or even lose control in very high-speed operations. To enhance the EME’s accuracy and stability, a better control method needs to be developed. To overcome the disadvantage of the conventional controller, a novel approach based on the full-state feedback control is proposed in this thesis. The current compensation gain is calculated by the EME’s state-space model with full-state feedback, which can compensate for the coupling term related to the rotational speed and current. Eventually, the poles and zeros of the EME control loop can remain unchanged for different operating conditions. The proposed control method combines the conventional proportional-integral(PI) controller with the full-state feedback to achieve the pole-placement flexibility but not drifting. Thus it simplifies the design of the controller and improves the accuracy of the EME. Details of the mathematical derivation of full-state feedback and the integration with the PI controller are provided in this thesis. Computer simulation and hardware experimental results are shown to verify the performance of the proposed control method. Based on 350V,1kW hardware experimental results, compared with the conventional PI control method, the proposed control method with full-state feedback has the same current transient performances and shorter settling times in different rotational speeds.