本文提出一多模式動態電壓補償器(Multi-Mode Dynamic Voltage Restorer, MMDVR)透過與主動式濾波器性能相結合，在電力事件發生時透過電壓補償以及孤島模式(Island Mode)控制，穩定各區域內負載側的供電品質。另外為了解決電網系統在控制上之動態響應，本文提出一遞迴式小波模糊派翠類神經網路智慧型控制器，透過遞迴結構的鑑別能力搭配類神經的線上學習結構可有效解決系統非線性參數及外部干擾，最後結合解耦合雙同步參考軸鎖相迴路，以改善傳統比例積分控制器之響應結果。 本文使用MATLAB/Simulink模擬本文所提出之多模式動態電壓補償器架構應用於以三區微電網之分散式架構來取代傳統電力系統，相較於傳統個別微電網，其特點為輸出電力可自給自發電，降低輸配電線路的損失。並比較傳統比例積分控制器與本文提出之遞迴式小波模糊派翠類神經網路在系統控制上之動態響應結果的不同。最後，透過模擬軟體、比較結果驗證了本文所提出之多模式動態電壓補償器搭配智慧型控制器能有效增強電網韌性與動態響應性能。;This thesis proposes the multi-mode dynamic voltage restorer (MMDVR) by combining the performance of the series active power filter, through voltage compensation and island mode control when a power event occurs, to stabilize the power supply quality at the load. In addition, in order to solve the dynamic response of the power grid system in the control, this paper propose the Recursive Wavelet Fuzzy Neural Petri Network (RWFNPN) intelligent controller, which can effectively solve the nonlinear parameters and external disturbances of the system through the discrimination ability of the recursive structure and the online learning structure of the neural network. Finally, combined with the Decoupling Dual Synchronous Reference Frame Phase Lock Loop (DDSRF-PLL) to improve the response results. This thesis use MATLAB/Simulink to simulate the proposed multi-mode dynamic voltage compensator structure proposed in this paper, which is applied to replace the traditional power system with the distributed structure of the three-zone microgrids. Compare with the traditional individual microgrid, it’s characteristic is that the output power can be self-generated, reducing the loss of distribution lines. And compare the dynamic response results of the traditional PI controller and the RWFNPN control. Finally, verified that the multi-mode dynamic voltage restorer can effectively enhance the resilience and dynamic response performance of the power grid.
