Paul Verrax, PhD student at L2S and SuperGrid Institute.
Tuesday 8th December at 11:00 via Microsoft Teams at
Abstract: The emergence of meshed HVDC grids is seen as a promising option to interconnect large amount of renewable energies over long distances. The protection of such grids requires the fast identification of faults such as short circuits affecting the transmission lines. Communication-based methods are thus not suited due to the transmission delays.
Many approaches involving a model of the transient behavior of the faulty line have recently been proposed. Nevertheless, an accurate description of the traveling wave phenomenon complex computations ill-suited for fast fault identification.
This talk presents a single-ended fault identification algorithm using a closed-form parametric model of the fault transient behavior. The model combines physical and behavioral parts and depends explicitly on the parameters that characterize the fault, namely the fault distance and impedance. When a fault is suspected, the fault parameters are estimated so that the model fits best the received measurements. The confidence region of the estimated fault parameters is used to decide whether the protected line is actually faulty or not. The proposed algorithm is tested on a 4 station grid simulated with an electro-magnetic transient software.
The method is able to identify the faulty line using a measurement window of less than 0.5 ms. This allows ultra-fast fault clearing and can hence improve the overall reliability of future HVDC grids.
Bio: Paul Verrax graduated in electrical engineering from Supelec and got a MSc in Control, Systems and Robotics from KTH, Stockholm in 2018. Since 2018 he is doing a PhD at University Paris Saclay and in collaboration with the Supegrid Institute based in Lyon, France. His research interests are mostly towards the protection of HVDC grids, the modeling of transients along transmission lines and system identification techniques applied to power systems.