In the course of the last decades, key investments in power systems have improved remarkably the overall quality of the associated information and communication infrastructure. Emerging wireless connectivity in conjunction with expanded computational resources and the proliferation of effective techniques for knowledge extraction drive a paradigm shift in power system analysis and provide a fertile ground for enhanced monitoring and control capabilities.
This special session targets fundamental/applied research challenges pertaining to the interaction of data-driven machine learning tools with wireless connectivity schemes in smart grid systems. In this context, contributions are expected to introduce novel cross-disciplinary approaches towards high fidelity estimation of state variables and informed system control.
Threads of interest include:
- Robust smart grid connectivity with incomplete data streams, focusing on resilient learning methods able to mitigate phenomena inherent to wireless communication, such as transmission failures, channel distortion, asynchronization and malicious data injection;
- Novel edge computing solutions tailored to smart grid architectures, emphasizing challenges related to scalability, latency, availability, privacy considerations and heterogeneous computation/storage capabilities;
- Physics-informed algorithmic methods and learning frameworks for efficient and scalable computation of smart grid tasks, by extracting meaningful representations of limited amounts of measurement streams to guide short-term operational decisions.