本研究使用模糊控制理論設計出轉矩與磁通控制器應用於感應馬達直接轉矩控制。傳統的直接轉矩控制利用兩個PI控制器產生兩個電壓向量命令,而PI控制器因參數固定無法即時反應出系統之動態響應,為了改善傳統直接轉矩控制的缺點,將設計出兩個模糊控制器,分別為模糊轉矩控制器與模糊磁通控制器,用來取代傳統的PI控制器,改善磁通響應和降低轉矩漣波並提升系統動態性能。 另外,本研究結合了模糊控制理論與小腦模型控制器,並採用高斯函數作為歸屬函數,設計出適應性模糊小腦模型轉速控制器,此控制器具有結構簡單及學習快速等優點。此外,由於馬達會因為溫升效應造成定子電阻變動而影響動態性能,因此,本研究將參考模型適應系統理論和模糊控制理論結合設計出模糊定子電阻估測器,即時調適定子電阻值,以準確估測磁通量。 最後,本研究整合模糊轉矩和磁通控制器、適應性模糊小腦模型轉速控制器和模糊定子電阻估測器,以現實感應馬達無速度感測器直接轉矩控制驅動系統。經模擬及實驗結果證明,在馬達負載轉矩為8 Nm,轉速控制範圍在36 rpm至2000 rpm時,所提出方法皆具有優異的轉速動態響應。
The fuzzy control theory is used to design torque controller and flux controller these are applied to direct torque control (DTC) of induction motor. In the conventional DTC scheme, two PI controllers are used to generate the reference stator voltage vector. Then the parameters of PI controllers are stationary, therefore it can’t respond the dynamic performance. In order to improve the disadvantages of conventional DTC, this design of two fuzzy controllers (FC); Fuzzy torque controller (FTC) and fuzzy flux controller (FFC) are designed to substitute the original PI controllers, which improves the flux response and reduce the torque ripple for better system dynamic performance. The thesis also adopts cerebellar model articulation controller (CMAC) and fuzzy control theory with Gaussian as membership functions to design the adaptive fuzzy cerebellar model articulation controller (AFCMAC) speed controller. The mentioned AFCMAC has the advantages of simple structure and rapid learning ability. In addition, for the dynamic performance of motor is affected by stator resistance due to temperature effect, this research combines model reference adaptive system (MRAS) with fuzzy control theory to make the stator resistance fuzzy stator resistance estimator (FSRE) for real-time estimating and acquiring accurate flux linkage. Finally, this scheme integrated FTC, FFC, AFCMAC speed controller and FSRE to achieve the sensorless speed control for DTC of induction motor. Via the simuliaton and experimental results, the proposed direct torque control systems have excellent speed response and robustness within 36 rpm to 2000 rpm and 8 Nm load torque.