Power Electronics As Edge Computing Grid Components

Objetivo

Increasing electricity generation with renewables and enabling the transportation sector's electrification are urgently needed to decarbonise our society. Power Electronic Converters (PECs) perform electrical energy conversions by utilising Silicon power semiconductor devices. PECs also enable integration of renewables, energy storage for grid support and e-mobility loads with the decentralized electric grid. Replacing Silicon with Silicon Carbide (SiC) semiconductors will advance PECs’ efficiency, thus lowering operating cost. It is foreseen that millions of PECs will be deployed in the decentralized electricity grid, making SiC the ultimate technology for a cost-effective grid operation. In addition, grid digitalization sets the ground for developing digital twins not only of the grid, but also of key grid components. Deploying millions of PECs in the grid, creates a unique opportunity to transform PECs from only performing electrical energy conversions to become intelligent, data-processing systems. Having such roles, PECs will be transformed to cognitive components, empowering real-time condition monitoring and actuations enabling performance enhancement. PELCom aims to deliver SiC-based PECs with embedded intelligence empowering real-time data processing for decision-making services using high-fidelity digital twins. The scientific approach of PELCom consists of three steps. Firstly, the transformation of driving circuits to real-time data processing units based on embedded electrothermal and reliability models; secondly, the device-agnostic tuning of the embedded models eliminating costly and impractical characterizations of millions of PECs; and thirdly, the development of an edge-computing system within the PEC that will serve as a decision-making scheme enabling actuations for PEC’s performance enhancement. PELCom will deliver the advanced technology required to address society’s expectations for sustainable, reliable and cost-effective electricity grids.