本研究主要實現一具有可變滑動面之滑動模式控制器(sliding mode controller, SMC),應用於感應馬達無速度感測器之直接式磁場導向控制系統。而為改善驅動器之性能和強健性,將模糊理論結合滑動模式控制,以模糊滑動面取代固定斜率常數滑動面,藉此縮短迫近模式下系統軌跡到達滑動面之時間,同時減少高頻切跳的現象。此外,基於參考模型適應性系統理論,設計適應性虛擬降階型磁通估測器,用來估測馬達磁通值,以達成適應性向量控制之目的。為增加磁通估測的強健性,利用模糊小腦模型理論設計出適應性模糊小腦模型定轉子電阻估測器,以實現感應馬達線上參數估測之能力。 最後,將模糊滑動模式控制器及適應性模糊小腦模型定轉子電阻估測器植入感應馬達向量控制系統實驗架構中。並經實驗和模擬結果得知,馬達操作於36 rpm至2000 rpm轉速範圍且8Nm負載的條件下,系統的動態表現不但優於固定斜率參數的系統,亦能有效提升系統抗參數擾動之能力。
In this thesis, a sliding mode controller (SMC) with fuzzy sliding surface for speed-sensorless direct field orientation control system of induction motor is implemented. Combine sliding mode control with fuzzy theory to improve the driver’s performance and robustness. In place of using the fixed sliding surface, the fuzzy-varying sliding surface makes the shorter reaching time of system trajectory, and reduces the chattering of SMC. To achieve adaptive vector control, the flux is estimated by the adaptive pseudo-reduced-order flux observer (APRO), which is based on model reference adaptive system theory. Under the structure, the fuzzy cerebellar model articulation theory is applied to the stator and rotor resistance estimators, which are realized for on-line parameter identification of induction motor. The experimental results indicate that, in contract to the fixed slop constant of the sliding surface, the fuzzy sliding mode controller not only performs well dynamic responses in a wide speed range (36 rpm to 2000 rpm) with 8-Nm torque load, but also has better robustness against parameter variations of induction motor.