Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/2761
標題: 應用模糊控制及類神經網路控制於壓電元件追蹤定位之研究
作者: 謝承勳
關鍵字: 模糊控制;類神經網路;壓電;追蹤定位;精密測量;ZMI-1000系統
出版社: 機械工程學系
摘要: 
本研究主要目的是針對壓電致動器的追蹤定位技術,作一深入的探討,並整合ZMI-1000系統、壓電元件模組、不同的控制理論及LabVIEW圖控軟體,擬建立較佳的壓電元件自動化測量及分析的追蹤定位控制技術。
由於壓電致動器可提供極微小之位移、驅動容易、體積小、出力大,對於高精密的定位控制有極佳的效果,然其本身之遲滯及非線性效應卻又大大的影響到其精密定位之能力。本研究中將以高精度之差頻雷射干涉儀進行壓電元件之位移測量並輔以模糊控制及類神經網路控制作為壓電元件之追蹤定位控制,以改善壓電元件的遲滯及非線性現象,期能達到精密追蹤定位之目標,並建立完整的人機介面,且能應用在其它型式的壓電致動器上而不需繁雜的修改。
由於壓電致動器的數學模式極難求取,所以我們為避免以系統鑑別的方式來對受控系統建模,而採用自組織模糊控制系統及啟發式類神經網路學習控制系統來做最佳的控制,且增加變速率自組織模糊控制及啟發式類神經網路進階參數自調學習控制的方法,比較這些控制方法應用在壓電動態追蹤定位上的優劣性,包括學習訓練時間比較及追蹤控制的比較,實際控制效果也相當不錯,對於壓電致動器不同頻率波形追蹤、不同波形目標的追蹤定位也有不錯的效果,輔以LabVIEW圖控程式軟體,建立容易使用及自動化測量並完成分析之人機介面,提供了幾種控制壓電元件追蹤定位方法的選擇。

The goal of this research is to propose a control algorithm for piezoelectric actuator position tracking technique. We have combined the high precision PZT equipment and some control theory to build up a precise position tracking technique.
Piezoelectric actuator can act as precise positioning equipment for its slight displacement, high efficiency and large output forces characteristics. However it still exists some inherent drawbacks, the hysteresis and nonlinearity, that may affect its precise positioning ability. This research utilized a heterodyne interferometer, served as high precision measurement tool, to monitor the piezoelectric actuator movement and feed the signal back to the motion control system. In our control system we used the Fuzzy and Neural-Network control techniques to improve the hysteresis and nonlinearity problems of the piezoelectric actuator, and have achieved accurate track position. By the way, the piezoelectric control system mentioned above can apply to other type of piezoelectric actuator without large modifications.
The mathematical modeling of piezoelectric actuator evaluation is an extremely difficult work to do. To keep off the complex and difficult system identification process, we choose Fuzzy and Neutral-Network control algorithms to enhance the PZT position tracking accuracy. Several approaches have been implemented into the PZT control system and function very well, including Self-organizing Fuzzy Logic Control, Heuristic Neural Network Learning Control, Different speed Self-organizing Fuzzy Logic Control and Heuristic Neural-Network Enhance Parameter Self-Tuning Learning Control. The comparative results between the control theory mentioned above are presented. The experiments also show that the proposed methods performed very well in the piezoelectric actuator position tracking control. We provide several algorithms effective in track position of piezoelectric actuator implemented with a PC based hardware computer and LabVIEW programming software environment.
URI: http://hdl.handle.net/11455/2761
Appears in Collections:機械工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.