本論文旨在實現並聯式主動電力濾波器(SAPF)之負載電流補償，透過同步參考座標(SRF)法進行鎖相，在直流鏈電壓控制上，由於直流儲能電容必須自行充放電來提供工作電壓給SAPF以及補償功率元件切換損失，如此會造成直流鏈電壓波動，因此本論文透過模糊類神經網路(Fuzzy Neural Network, FNN)來取代傳統比例積分(Proportional-Intergral, PI)控制，以提升SAPF的工作效率。 在電力系統設計方面，本論文採用三相四線式系統，並透過OPAL-RT即時模擬器來協助完成硬體迴圈，首先使用Matlab/Simulink模擬並聯式主動電力濾波器、三相電源端、非線性負載以及間諧波負載的架構，再利用Texas Instruments公司所生產之型號TMS28335數位訊號處理器(DSP)進行本論文所使用的控制策略來達到補償諧波電流之作用，並且最後再分析補償後的總諧波電流失真，其數值須符合IEEE Std. 519-2014之規範以及現今台灣電力公司諧波管制標準。 ;This paper aims to realize the load current compensation of the shunt active power filter (SAPF). The phase-locking is performed through the synchronous reference frame (SRF) method. In the DC link voltage control, the DC energy storage capacitor must be charged and discharged to provide work. The voltage gives SAPF and compensates the switching loss of the power components, which will cause the DC link voltage to fluctuate. Therefore, this paper uses fuzzy neural network(FNN) to replace the traditional proportional-integral (PI) control to improve SAPF work efficiency. In terms of power system design, this paper uses a three-phase four-wire system, and through the OPAL-RT real-time simulator to help complete the Hardware-in-the-loop(HIL). First, use the Matlab/Simulink to simulate shunt active power filter, three-phase power supply, non-linear load and inter-harmonic load, and then use the model TMS28335 digital signal processor (DSP) produced by Texas Instruments to carry out the control strategy used in this paper to achieve the effect of compensating harmonic current. Finally, after analyzing and compensating the total harmonic current distortion, its value must comply with the IEEE Std. 519-2014 specifications and the current harmonic control standards of the Taiwan Power Company.
