本文將模糊概念導入具有監督控制效能之小腦模型控制器中以設計一適應性模糊小腦模型控制器,此控制器不但保留原先小腦模型控制器快速學習、結構簡單和非線性之學習能力等優點外,在控制上亦具有模糊控制之特性。此外本研究之適應性模糊小腦模型控制器並聯一監督控制器,透過所設定的能量函數上限,可以將能量函數有效的抑制在上限內。 本研究之感應馬達向量控制系統結合直接轉子磁場導向與適應性控制法則,利用適應性虛擬降階型磁通估測器估測轉子磁通,不但解決了全階型磁通估測器置根比例常數須隨轉速命令變動而調整之缺點,同時可減少運算時間。而為了達到無轉速量測器控制與適應轉子電阻的變化,本研究採用小腦模型PI控制器估測轉速與轉子電阻。 經由模擬與實驗結果證明,此一適應性模糊小腦模型控制器植入感應馬達向量控制系統中,運轉在2%到100%之額定轉速,負載為8Nm時轉速不但能快速響應,同時在馬達參數變動之環境下仍具有很好的強健性,且監督控制器的確可以發揮其監督的功用。
In this thesis, an adaptive fuzzy cerebellar model articulation controller (AFCMAC) and the corresponding speed control of induction motor (IM) are proposed. The controller is developed by integrating the cerebellar model articulation controller (CMAC)-based supervisory control and the fuzzy control. The advantages of the AFCMAC include rapid learning ability, simple structure, and non-linear function learning capability. With the cooperation between AFCMAC and supervisory controller, as well as the setting of a upper limit of energy function, energy function can be effectively controlled within the upper limit. The vector control of the motor drive proposed in this thesis combines direct field orientation control (DFOC) and the adaptive control law. An adaptive pseudo-reduced-order flux observer (APRO) is used to estimate the rotor flux and solve the problem that the controller gains of pole assignment must follow the speed command, which often happens when the adaptive full-order flux observer (AFO) is applied. Meanwhile, the computation time of control algorithm can be decreased. In addition, to establish the speed sensor-less control and overcome the drift effect of rotor resistance, the cerebellar model articulation PI controllers (CMAPIC) are adopted to estimate speed and rotor resistance separately. Under the operation conditions that the rate speed range varies from 2% to 100% with 8-Nm start-up torque load, experimental results indicate that the speeds not only have quick response, and also remain robust in the environment of varying motor parameters with the proposed AFCMAC.