本計畫預計將完成5kW直流/交流併網型轉換器之設計與製作，其中閉迴路控制器將採用比例積分微分型類神經網路(Proportional-Integral-Differential Neural Network, PIDNN)控制器並以DSP TMS320F2812進行實現。為使併網型轉換器能順利與市電併聯，將設計同步併網控制法與主動式孤島偵測法。另一方面，為達成有效的電能管理分配以及提升分散式發電系統使用效率，將設計一微電網控制器，以CAN Bus作為分散式發電系統內部和DSP TMS320F2812間之資料傳輸。待電力資訊收集完畢後使用微電網控制器之GSM無線通訊模組將微電網控制器之資訊傳送至市電端之區域調度中心，以利區域調度中心對各分散式發電系統作有效的能源分配與管理。本計畫最後將藉由IEEE1547及UL1741所提供之測試方式以測試所發展之併網型轉換器，使其具備孤島運轉之相關功能，並能符合相關之併聯標準與規範。 The objective of this project is to accomplish the design and manufacture of 5kW DC/AC grid-connected converter, in which the proportional-integral-differential neural network (PIDNN) controller will be adopted for the closed-loop controller and be implemented by means of DSP TMS320F2812. In order to connect the grid-connected converter with the grid smoothly, the methods of synchronous grid-connected control and active islanding detection algorithm will also be designed. Moreover, to improve the efficiency of power management for the distributed generation (DG) system, a microgrid controller will be designed and the CAN Bus will be adopted for internal data transmission between the DG system and the DSP TMS320F2812. After the power information is collected completely, the data will be sent to the area dispatched control center (ADCC) of the grid side utilizing the GSM wireless module installed in the microgrid controller for the efficient power allotment and management of the DG system. Finally, in order to possess the related function of islanding mode and meet the relevant interconnection standards, the project will verify the developed grid-connected converter by using the offered test methodology in IEEE1547 and UL1741 and to meet the standard. 研究期間：9901 ~ 9912
