Predicting Renewable Energy Resources using Machine Learning for Wireless Sensor Networks

Wireless Sensor Network (WSN) nodes rely on batteries that are hazardous and need constant replacement. Therefore, we propose WSNs with solar energy harvesters that scavenge energy from the Sun. The key issue with these harvesters is that solar energy is intermittent. Consequently, we propose machine learning (ML) algorithms that enable WSN nodes to accurately predict the amount of solar irradiance, so that the node can intelligently manage its own energy. Our ML models were based on historical weather datasets from California (USA) and Delhi (India) for the period between 2010 to 2020. In addition, we performed data pre-processing, followed by feature engineering, identification of outliers and grid search to determine the most optimized ML model. In comparison with the linear regression model, the support vector regression (SVR) model showed accurate forecasting of solar irradiance. Moreover, it was also found that the models with time duration of 1 year and 1 month has much better forecasting results than 10 years and 1 week, with both root square mean error (RMSE) and mean absolute error (MAE) less than 7% for Sacramento, California, USA.