ARTIFICIAL INTELLIGENCE TECHNIQUE TO ENHANCE POWER QUALITY AND STABILITY OF THE GRID-CONNECTED PHOTOVOLTAIC SYSTEM

Abstract

Massive deployment of grid-connected photovoltaic (PV) systems has raised considerable issues pertaining to the power quality, stability of voltages, and fault ride-through. Harmonic distortion, voltage sag on grid faults, and inadequate dynamic response under variable irradiance conditions remain among the issues that limit reliable penetration of PV to the modern power networks. In this paper, an artificial intelligence (AI) control system has been introduced, which uses an ‘Artificial Neural Network (ANN)’ to improve the quality of power and stability in a grid-connected PV system. The suggested solution combines an ANN-based fault detection unit and a Distribution Static Compensator (DSTATCOM) to deliver rapid voltage support in case of grid disturbances. An elaborate mathematical model of the PV array, DCDC ‘boost converter, voltage source inverter (VSI), and grid’ interface is designed and simulated. The ANN is trained under a supervised learning method to effectively differentiate between normal and faulty operating conditions. Simulation outcomes show great enhancement of voltage stability, harmonic reduction, and fault ride-throughs. ANN has a regression accuracy rate of 99.8% and is able to detect faults very quickly to implement a reactive power compensation in a timely manner. The solution suggested is a very efficient and smart way of enhancing the quality of the power and providing a resilient grid integration of the photovoltaic systems in the smart grid setup.