A Smart Control of Solar Panel Mechanism with Arduino Based Data Logger for Neural Network Analysis

Abstract

A solar panel mechanism is a system of devices that can track the sun's movement and adjust the solar panels accordingly. In this experiment, the amount of solar irradiation that can be captured and converted into electrical energy by a solar panel and pyranometer is being investigated. Preliminary study has been done to gather solar irradiance data from the pyranometer and solar panel for comparison. The data gathered was then used to perform a forecast using neural network analysis. Two configurations were used in this experiment where both labelled as "Auto-detect" and "Passive." Auto-detect refers to the placement of the solar panel on a 3D-printed base with a servo motor that regulates rotations in response to inputs from Light Dependent Resistor (LDR) sensor. This arrangement will keep an accurate orientation between solar panel and solar energy source. While in passive configuration, the solar panels are manually positioned. From this study, the maximum sun irradiance measurements made with a pyranometer were 0.9986 W/m2 in auto-detect mode and 0.9962 W/m2 in passive mode. Auto-Detect Configuration has 72.46% efficiency while Passive Configuration: 58.46%. For the neural network analysis, it has been found that the prediction has low accuracy. Therefore, additional improvement techniques can be used in future work.