Exploiting Digitalization of Solar PV Plants Using Machine Learning: Digital Twin Concept for Operation

Tolga Yalçin, Pol Paradell Solà, Paschalia Stefanidou-Voziki, Jose Luis Domínguez-García, Tugce Demirdelen

The rapid development of digital technologies and solutions is disrupting the energy sector. In this regard, digitalization is a facilitator and enabler for integrating renewable energies, management and operation. Among these, advanced monitoring techniques and artificial intelligence may be applied in solar PV plants to improve their operation and efficiency and detect potential malfunctions at an early stage. This paper proposes a Digital Twin DT concept, mainly focused on O&M, to obtain more information about the system by using several artificial intelligence boxes. Furthermore, it includes the development of several machine learning (ML) algorithms capable of reproducing the expected behavior of the solar PV plant and detecting the malfunctioning of different components. In this regard, this allows for reducing downtime and optimizing asset management. In this paper, different ML techniques are used and compared to optimize the selected methods for enhanced response. The paper presents all stages of the developed Digital Twin, including ML model development with an accuracy of 98.3% of the whole DT, and finally, a communication and visualization platform. The different responses and comparisons have been made using a model based on MATLAB/Simulink using different cases and system conditions.