In this work, the concept of an Artificial Intelligence-based (AI) Digital Twin (DT) of an aircraft system is introduced, with the goal to improve the corresponding MRO Operations. More specifically, the current study aims to obtaining knowledge on the optimal placement of sensors in an ideal Power Electronics Cooling System (PECS) of a modern airliner, aiming to improve input data as a basis for an AI-based DT. The three main fluid parameters to be measured directly or indirectly at various physical locations at the PECS are mass flow rate, temperature and static pressure. The physics-based model can then be combined with a Machine Learning (ML) model, such as a Random Forest (RF), with a multitude of decision trees. Following, the AI system determines whether the PECS operations is considered normal, aiming to optimize the performance of the system and to maximize the Useful Remaining Life (URL). The suggested AI-DT approach is based both on data-driven and physics-based models, an approach which results in increased reliability and availability, reducing possible Aircraft on Ground (AOG) events. Subsequently, the enhanced prediction capability results in the optimization of the maintenance processes and in reduced operational costs.