Variant Mode Data Analytics in Predicting the Radiation Effect on Solar Power Generation using Machine Learning Algorithms

Power harvesting using solar power is the recent trend and innovations happening in deploying many types of equipment working with solar power. This is harmless and greatly reduces pollution and is eco-friendly. The government also provides more concessions for establishing these solar power harvesting methods. There are two subsystems in solar power generation like sensor management systems. The subsystems have to be managed by predicting the power generation and identifying the right time for panel cleaning, and maintenance. In solar power generation systems, it is necessary to identify the faulty equipment and replace it for robust power generation. In the proposed article we are predicting the effect of ambient temperature, and module temperature on radiation of the solar power generation system using the Weka machine learning tool using algorithms like SMOreg, Linear regression, KNN, and Multilayer Perceptron. The prediction model predicts the solar power radiation with Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) of 0.0294 and 0.0558 of the Ambient and module temperature respectively. The prediction of radiation in the solar power plant will be helpful in grid maintenance, efficient use of accessories, identifying and servicing the sub-optimally performing unit to increase the daily yield, and reducing the operational cost.