The research project FELINES aims at designing a protection system capable of sensing electromagnetic fields that are preliminary to a lightning event, and consequently disconnect part (or all) of the electric infrastructure under its protection. These fields are generated by the so-called preliminary breakdown pulses, localized events taking place during the first phases of the lightning
inception. 

Few study in the literature exist about the modelling and measurements of such phenomenon, and one of the key points of the project is to use them (or develop new ones) to perform a thorough study about the electromagnetic quantities that characterize the phenomenon itself, that can be measured and exploited for the protection system. At the same time a complete study relative to the effects of direct and indirect lightning on specific electric infrastructures will be carried out, with the goal of quantitatively understanding how the key parameters of a lightning can be a real danger or not. The activity above described is the basis for the development of the so-called early detection algorithm: a Machine Learning (ML) based procedure that should be able to classify the received (and transduced) electromagnetic signals into dangerous or not dangerous. The early detection algorithm will be the core of the protection system, and will be completely developed from scratch
during the project. 

All the components of the system (sensors, transducers, algorithm) previously selected, modeled and developed will contribute to the creation of a digital twin, that will be the final outcome of the project. As a matter of fact, the digital twin will serve as a demonstrator, that will simulate the preliminary breakdown pulse, the way it is measured from the system and the final decision taken by the ML based algorithm, that will result either in a “non dangerous” decision or in a “dangerous” decision, leading to an activation signal sent to the relative protections.