本研究利用直接轉矩控制(direct torque control, DTC)理論實現感應馬達速度控制,其優點為動態響應快、系統架構簡單和容易計算。並且本研究應用空間電壓向量調變技術,來改善傳統直接轉矩控制系統的轉矩漣波及噪音等問題。 另外,本研究結合小腦模型控制器與模糊控制理論,設計出適應性模糊小腦模型轉速控制器(AFCMAC),此控制器具有結構簡單及學習快速等優點。為了克服馬達溫度隨著運轉時間增加而上升,造成馬達定子電阻值變動的影響,本研究將參考模型適應系統(model reference adaptive system, MRAS)理論和模糊控制理論結合設計出模糊定子電阻估測器(fuzzy stator resistance estimator, FSRE),即時調適定子電阻值,以精準估測磁通量。最後,本研究以轉速估測器來實現無轉速量測器直接轉矩控制系統,除了可以節省成本外,還可避免破壞馬達之結構。 經模擬及實驗結果證明,在馬達負載轉矩為8Nm,轉速控制範圍在36rpm至1800rpm時,具模糊定子電阻估測器之直接轉矩控制系統皆具有優異的轉速動態響應,且感應馬達控制系統在參數變動時都能維持良好的強健性。
In this thesis, the direct torque control (DTC) scheme is used to fullfil the speed control for induction motors. As compared to the field oriented control (FOC), the proposed controller has the following advantages: quick dynamic response, simple structure, and low computation complexity. Also, the space voltage vector modulation is adopted to replace the conventional hysteresis-switching to reduce the torque ripples and noise influences. The proposed speed control is an adaptive fuzzy cerebellar model articulation controller (AFCMAC) which is designed based on the cerebellar model articulation controller (CMAC) and fuzzy control theory. AFCMAC has the advantages of simple structure and rapid learning ability. In addition, the fuzzy stator resistance estimator (FSRE) is proposed to accurately estimate the stator resistances and stator flux in the temperature varation. The speed estimator is utilized to realize the speed sensor-less control and achieve the advantages of cost-effectiveness and structure robustness. The results of simulations and experiments show that, with 8 Nm load torque, the proposed direct torque control system with AFCMAC and FSRE achieves excellent speed response and robustness against parameter variations within a speed range between 36 rpm to 1800 rpm.