Empowering Random Forest to predict operation of Dual axial and fixed tilt PV panel

Abstract

Optimizing photovoltaic (PV) panel performance is becoming more and more important as the need for renewable energy sources rises. This is especially true when it comes to minimizing the negative impacts of shadowing, which can reduce energy generation. With an emphasis on efficiency under varied circumstances, this study investigates the use of a Random Forest model to forecast the performance of both dual-axis and fixed-tilt $P V$ panels. Important factors that significantly affect $P V$ output are examined, including latitude, solar elevation, azimuth angles, and ideal tilt. Using a hardware prototype that has servomotors and light-dependent resistors (LDRs), dual-axis tracking is simulated for the best possible solar panel placement. The study uses actual temperature, voltage, and humidity data to compare the energy production and efficiency of dual-axis and fixed-tilt panels. Using real-world temperature, voltage, and humidity data gathered by the Node-MCU module and analyzed by a cloud-based system, the study compares the energy production and efficiency of dual-axis and fixed-tilt panels. According to simulation data, dual-axis tracking achieves up to 88% of the system’s maximum capability, providing a noticeable efficiency boost. Furthermore, Random Forest classification is used to forecast PV system performance under different circumstances, emphasizing the superiority of dualaxis systems in increasing energy production and reducing the effects of shadowing.