ABSTRACT
The use of photovoltaic (PV) systems has become popular worldwide as a potential renewable and sustainable energy alternative to traditional fossil fuels and other non-renewable energy sources. These systems generate electricity using solar energy and mitigate the effects of climate change. However, the uncertain nature of solar radiation and temperature creates unsteadiness in the response of the PV systems. Certain abnormalities usually occur in the PV systems’ characteristics due to the presence of these uncertainties. So, to design the PV system’s effective maximum power tracking mechanisms through direct current (DC) converter control, it is required to correctly investigate and identify all these abnormalities. With this motivation, this chapter proposes various systematic analytical models to detect any abnormalities (e.g., missing readings, garbage readings, outlier readings, redundant readings) that are hidden in the PV system’s characteristics to ensure optimal performance and prevent potential failures. These data-driven analytical models use a non-linear system identification method and detect all the possible abnormalities in the PV system characteristic data. To implement the proposed analytical models, a practical PV dataset “Photovoltaic Power and Weather Parameters” from SolarTech Lab is considered.
