Machine Learning based Voltage Regulation for PV System using Buck Converter

This paper presents the issue of voltage instability in a solar-powered system due to fluctuations in solar irradiance. These fluctuations in solar irradiance cause oscillations in output voltage and reduce the efficiency of the overall system. To overcome this issue, the paper proposes a solar-powered DC voltage regulation system that employs machine learning techniques to maintain the system’s operational stability. The primary source of energy is from a PV system and it includes a buck converter that regulates the output voltage. A microcontroller, integrated with a machine learning model using K-Nearest Neighbors (KNN) algorithm, continuously monitors the solar irradiance, load requirements, and system voltage and adjust the buck converter’s duty cycle. By leveraging the historical data and predictive analytics, the proposed system can expect fluctuations in solar irradiance and load, thus enabling effective voltage regulation. This system reduces dependency on grid power. It enhances energy efficiency and system operational stability.