Battery Monitoring and Energy Forecasting for an Off-Grid Solar Photovoltaic Installation

This document details the design and implementation for an off-grid PV energy management system with key objectives being battery monitoring and energy forecasting. Coulomb counting is used to estimate the state of charge of a 24 V, 100 Ah lead-acid battery bank. An ACS712 hall effect current sensor module is used to implement the coulomb counter. A random forest regression model is used to predict solar irradiance with a root mean square error of 15.4%, when comparing actual Johannesburg irradiance data to predicted data. Energy forecasting is successfully done with a root mean square error of 6.2% for a three-day continuous test. It is concluded that the proposed solution successfully demonstrates a working principle for an energy management system, however, further improvements are needed to make the forecast model as accurate as possible.