AI-assisted power quality improvement in PV systems using RNN-based DSTATCOM and HiClampCasBoost converter

The rapid deployment of Photovoltaic (PV) systems in modern power distribution networks has caused major issues related to voltage fluctuations, reactive power unbalance, harmonics, and less efficient systems. It is critical to protect Power Quality (PQ) under these conditions to ensure safe and consistent operation of grid. Flexible AC Transmission System (FACTS) devices are used to mitigate those problems encountered in the network. Hence, this work proposed a grid-tied PV system integrated with Distribution Static Compensator (D-STATCOM) device that help accurate voltage sag, swell, and harmonic distortion at the distribution level. In this work, the interaction of D-STATCOM with the grid is controlled by a Recurrent Neural Network (RNN) based controller with PWM signals adjusted for voltage stability, sag, swell and load disturbances compensation. Further, a high-gain HiClampCasBoost Converter (HiClampCasBoost) converts low PV output voltage and regulates the DC-link, assuring constant output of the D-STATCOM while under voltage conditions prevail. Additionally, a Adaptive Neuro-Fuzzy Inference System (ANFIS) based Maximum Power Point Tracking (MPPT) controller, optimized using the Hummingbird Optimization Algorithm (HOA) for tracking Maximum Power Point (MPP) of PV array under dynamic solar conditions to produce maximum energy from PV array whilst ensuring D-STATCOM DC-link stability. The MATLAB simulations show that the developed system enhances PQ and grid stability in comparison to conventional methods. The proposed system has several benefits, such as superior voltage regulation, reduced THD of 1.63%, superior converter efficiency of 96%, tracking efficiency of 99% superior dynamic operation performance, etc.