一般功率調節系統(Power Conditioning System, PCS)的主要功能為改善微電網系統(Microgrid, MG)之穩定性與負載平衡性，本文設計一多功能功率調節系統，除具有一般功率調節策略外也具有電壓補償之功能，當發生電力事件時可有效改善負載電力品質。另外，由於電力系統之動態特性，本文提出一智慧型控制法稱為模糊派翠類神經網路，藉由模糊推論可有效處理不確定資訊，再利用派翠網路之故障辨識能力搭配類神經網路的線上學習法則，最後結合雙二階廣義積分器，以此改善傳統比例積分控制器的應用於系統上的響應結果。 本文透過MATLAB SIMULINK模擬所提出之系統架構與控制法，並比較傳統PI控制與本文提出之模糊派翠類神經網路應用於本文系統上響應結果的差異。此外，利用OPAL-RT即時模擬器搭配數位訊號處理器所建構之之硬體迴圈測試環境，觀察智慧型控制法在即時響應結果上的差異，並進一步驗證本文系統之可行性。最後，由實驗結果證實本文提出之多功能功率調節系統能夠有效改善負載端之電力品質與穩定性。 ;The main function of conventional Power Conditioning System(PCS) is to improve the Microgrid (MG) stability and load balance. This paper design a Multi-functional Power Conditioning System(MFPCS), it has not only the conventional power conditioning strategy, but also the voltage compensation function which can improve the power quality of load. Beside, because the dynamic characteristic of power system, this paper propose the intelligent controller called Fuzzy Neural Petri Net (FNPN), which can effectively processing uncertain situation by fuzzy inference, identify system fault with petri net and neural network online learning algorithm, combine the dual sequence order generalized integrator to improve system response control by the traditional Proportional Integral(PI) controller. This paper use MATLAB SIMULINK simulate the proposed system construct and control method, compare response difference with PI and FNPN control. Beside, this paper also use HIL environment by OPAL-RT and digital signal processer to experiment real time response, verify system feasibility proposed in this paper. Finally, the experimental results confirm that the multi-functional power conditioning system proposed in this paper can effectively improve the power quality and stability of the load.
