Adaline-based direct power control of grid-connected photovoltaic system to improve power quality

This paper presents a grid-connected photovoltaic system in which a two-level voltage source inverter operates as a shunt active power filter. A hybrid control strategy combining an Adaptive Linear Neuron (Adaline) with Direct Power Control (DPC) is proposed. In this control, Adaline extracts reference powers, while DPC generates the optimal switching signals through a switching table, without the need for a separate Pulse Width Modulation (PWM) or hysteresis modulation stage. The system performance is evaluated with MATLAB/Simulink simulations under various conditions, including linear and nonlinear loads as well as constant and variable irradiance. Simulation results confirm that the photovoltaic (PV) system effectively delivers active power to the grid while compensating for harmonic and reactive currents. The total harmonic distortion with the Adaline-DPC remains below 5% of the IEEE-519 standard limit. Although the Adaline-DPC has a higher computational burden, it outperforms the conventional hybrid DPC methods such as conventional DPC (CDPC), HSF instantaneous active and reactive power control with DPC (HSFDPC), SRF control with DPC (SRFDPC), and HSFSRF control with DPC (HSFSRF-DPC) by providing lower THD under both ideal and nonideal source and load conditions.