MMC-based PV-Fed STATCOM with Hybrid GA-RBFNN for PQ Enhancement

Abstract

Power quality (PQ) is a consequential factor, which is highly influenced by the functioning of the transmission and distribution network. In the distribution systems, voltage swells, sags, harmonics and flickers are regarded as major power quality issues. The Static Synchronous Compensator (STATCOM) is one among the specialized power devices, which has gained considerable interest for its capacity of enhancing the performance of power systems. This paper analyses the power quality issues of a photovoltaic (PV) system adopting a STATCOM-based five-level Modular Multilevel Converter (MMC). An interleaved quadratic boost converter (QBC) is employed in this work, which escalates the PV output in a wider range. An Adaptive Neuro-Fuzzy Inference System (ANFIS) controller is employed for obtaining the fast, efficient and flexible control of the DC-DC converter. The MMC plays a crucial part in mitigating power quality issues as it has the beneficial impacts such as scalability, power quality and modularity. In this approach, a five-level MMC is implemented, which transfers active power to the load by maintaining the power factor constant. The reference current generation of MMC is carried out by a hybrid Genetic Algorithm-Radial Basis Function Neural Network (GA-RBFNN) algorithm and the pulses of MMC are generated through a hysteresis controller. It provides improved training speed and convergence along with enhanced operating efficiency of the network. The output of five-level MMC is fed to a transformer-coupled LC filter. The simulations are carried out in Matlab and a minimized Total Harmonic Distortion (THD) of 1.12% is obtained.