Chapter nine - Weather-based solar power generation prediction and anomaly detection

Abstract

Leveraging the renewable energy resources has become a necessity with the depletion of the nonrenewable, especially the conventional energy resources. Additionally, the involvement of artificial intelligence has started playing an important role in the advancement of the renewable energy sectors. In this chapter, we focus on the prediction of solar power generation by using the regression strategy. Artificial neural networks make the prediction task easier and more accurate as compared to the traditional methods. The prediction of power generation is carried out based on the weather conditions like ambient temperature, module temperature and irradiation. A high R2 value ensures the effectiveness of our regression model. Also, the problem of anomalies in the prediction of power generation is dealt with in this work. The anomalies can occur due to dirt on the panel, or faulty panels, and we detect that using our regression model. Multiple occurrences of anomalies from a particular inverter help to identify the malfunctioning inverters.