Computationally Efficient Modeling of Electrically Short DC Cables with Frequency-Dependent Parameters
Abstract
In recent years, the application of DC power systems has attracted much attention. As an integral part of any DC power system, accurate and computationally efficient modeling of DC transmission lines consisting of cables and overhead lines is essential for the transient analysis. Developing a line model that takes into account the frequency dependency of the Per-Unit-Length (PUL) parameters and is valid over a wide frequency range is typically challenging.
This thesis initially focuses on the representation of the DC transmission line with a lumped model. In this regard, a Passive Lumped Frequency-Dependent Parameter (PLFDP) model is proposed for accurate representation of DC transmission lines consisting of a pair of single-conductor lines. To extend the application of the PLFDP model to DC cable systems with sheaths and/or armours, the PLFDP model is enhanced to take into account the electromagnetic coupling between the conducting layers of the multiconductor DC cable system. Furthermore, when the sheaths and/or armours of the cable system are solidly grounded, a procedure based on the Kron Reduction technique is introduced to develop a more simplified passive model. The proposed PLFDP models (i) include the frequency dependency of the PUL parameters, (ii) are scalable, i.e., once the model is developed for the unity length of the cable system, it can be simply scaled to represent other lengths, (iii) and have guaranteed passivity as they consist of passive circuit elements. Consequently, this thesis proposes passive lumped models that provide almost the same level of accuracy as compared to the most accurate line model available in PSCAD/EMTDC, while they significantly decrease the simulation time for electrically short lines (up to 11.3 times in a specific simulation study conducted).
Also, when the sheaths and/or armours of an electrically short cable system are resistively grounded, a novel approach is proposed to reduce the dimension of the original PUL parameters. The proposed method for the extraction of the reduced-dimension PUL parameters significantly decreases the simulation time (up to about 300 times in a conducted simulation study conducted) when compared to the application of the original PUL parameters, while it provides the same level of accuracy as compared to the application of the original PUL parameters.