This paper proposes a single-phase shunt active filter (ShAF) combined with photovoltaic (PV) to enhance power quality performance by reducing source current harmonics and compensating for reactive power in a single-phase 220-Volt distribution system with a frequency of 50 Hz connected to a non-linear load. The PV panel consists of several PV modules with a maximum power of 600 W each. An adaptive neuro-fuzzy inference system (ANFIS) controls the voltage in the DC link capacitor circuit in the ShAF. This method is proposed to overcome the weakness of the Fuzzy Sugeno method in neural-network-based learning capabilities to determine the fuzzy rules of the input membership functions (MFs) and the weakness of the proportional-integral (PI) control in determining proportional and integral constants using trial and error method. The single-phase system is connected to a non-linear load with a combination, i.e. without ShAF, using ShAF, and using ShAF-PV, respectively, with a total of seven cases. Based on the three proposed control methods and model configurations, the ShAF-PV circuit with ANFIS control is able to result in the best performance because it is able to produce the lowest source current THD. The single-phase system using ShAF-PV with ANFIS control is also capable of injecting the largest reactive power compared to the ShAF and ShAF configurations with PI and Fuzzy-Sugeno control. The increase in reactive power in the ShAF-PV is further able to compensate for the reactive power, so it is able to suppress and reduce the source reactive power significantly.
Fitriani et al. (Fri,) studied this question.