Learning, processing and communication in smart energy grids

The introduction of renewables and thegoal of increasing efficiency and reliability is revolutionizing the energygrids. Energy grids must become intelligent and be able to manage energy flowsand flexibly adapt to changes in production and demand. This paradigm hasfostered the creation of new business models on top of technology advances. Thephysical infrastructure must embed sensing, communication and processingcapabilities. This talk provides an overview of relevant applications (frommetering, to diagnostics, to energy management). It then covers relevant newwireless (among which long range radio), and power line communication (PLC)technologies to provide massive connectivity. New applications and recentresults of PLC that enable grid diagnostics by processing high frequencysignals will be presented. In this context, we also elaborate on the concept oflearning the system model and discuss how top-down phenomenological approachesfollowed by machine learning paradigms can be merged with bottom-up physicaldomain approaches in different layers of PLC for smart grids.