Development of an adaptive intentional controlled islanding scheme and post-islanding corrective control actions
C.A. Karagiannis Kaltsikis (TU Delft - Electrical Engineering, Mathematics and Computer Science)
M. Popov – Mentor
I. Tyuryukanov – Mentor
M.A.M.M. van der Meijden – Graduation committee member
M. Ghaffarian Niasar – Graduation committee member
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Power systems have undergone major changes the past few decades, leading to an increase in the frequency of occurrence and severity of blackouts. Intentional controlled islanding constitutes the final resort to rescue the system from a blackout by separating the network into islands in a controlled manner. Various post-separation control action such as generation rejection and load shedding have to be considered in order to restore the active power balance in the created islands.
The transient stability of the constructed islands can be enhanced by identifying and grouping together the coherent generators of the power system. This can be achieved by representing the electrical network as an undirected graph and performing constrained graph partitioning with respect to the generator nodes.
If the demand is higher than the active power produced, the frequency of the formed island starts declining. To deal with this problem and keep the frequency within acceptable limits, it often becomes necessary to curtail a proportion of the load. An advanced underfrequency load shedding scheme can dynamically adjust the shedding steps based on accurate power deficiency estimation and thus minimize the total load shed.
In this thesis, the coherent generator groups are identified for several power systems. Then constrained graph partitioning is used to split these networks into islands, while forcing the coherent generators to remain in the same island. For this step, three different algorithms performing constrained clustering are implemented in Matlab and compared for two different study cases. Moreover, an advanced underfrequency load shedding scheme is applied on various constructed islands by utilizing PowerFactory and Python. Its objective is to minimize the load shedding amount, while keeping the frequency within the predefined limits. The amount of load that needs to be curtailed is estimated based on the frequency gradient of the generators in the island with an excess in power demand.