模糊型PI速度控制器在交流永磁馬達之應用

Abstract

一般在設計PI控制器時其參數調整的方法有：試誤法或參數鑑別得知系統參數後再依古典控制法確定參數。但是試誤法調整相當耗時，而參數鑑別又牽扯到是否能精確得知參數以求得精確的控制。本文嘗試以模糊機制使控制器具備自動調整參數的能力以期望其達到智慧型控制器的雛形。 本文以向量控制器為基本架構，配合模糊手法使控制器參數變為可自我調整，接著再以另一層的模糊機制達到修正參數至我們所設定的狀態。廣義而言，這樣的修正功能或許稱不上具備學習能力，但是就實際效果而言，這樣的設計卻可使輸出響應有向理想響應逼近的趨勢。 由於此系統架構及控制器的設計步驟都是很淺顯的，因此我們並不深入探究控制器的理論部份。但這不表示這樣的控制器經不起考驗，由模擬和實驗結果，我們可相信其實用上的可行性。

In general, we design a PI controller by: try-and-error or parameters-identified method. But try-and-error always takes much time to make a fine tuning, and identify-method is difficult to realize precisely.So this thesis attempts to make a controller to achive the goals of self-tuning and self-modification by a 2-stage fuzzy structure. The basic structure of controller is F.O.C.(Field-Oriented Controller). By fuzzy design procedure, we make the controller to have the ability of self-tuning, and then we apply another fuzzy mechanism to achive the goal of parameter modification. Strictly, this may not be a learning mechanism, but it with the ability of modification to the trend that we specified in advance. In this thesis, we focus on the practicability of controller instead of the theoretical topic. It doesn't mean that the design is not good. By the conclusions of simulation and experiment, we see the effects of this design.