Data-driven methods, including graph neural networks, have been studied for accelerating power flow calculations in recent years, but very little attention has been paid to the solution feasibility, which can be obtained by traditional solvers. This paper presents a Variational G…
Reliable monitoring of hydroelectric generators requires descriptors that capture both electrical loading and electromagnetic field behavior. This work investigates operating-regime identification in the Porjus U9 10-MW Kaplan hydrogenerator using synchronized measurements from t…