For conducting harmonic analyses and working in the field of power quality assurance, impedance frequency characteristics represent important information. Typically, we are especially interested in their profiles at the busbars of the power system, where new devices are connected. The course of these characteristics depends on the network elements. These can be modeled in various ways, which is why a part of this thesis includes a comparison of results obtained using different models and line parameters.
Initially, to recognize the impact of certain simplifications, we developed a theoretical model for calculating the impedance characteristic. This was followed by modeling a simple network and its elements in the software tools DIgSILENT PowerFactory and PSCAD. Simulation results were obtained for a geometrically and linearly modeled overhead line of 80 km and a cable line of 10.5 km, where, in each case, both lumped and distributed line parameters were considered. We then compared the models and drew conclusions based on the differences in results. The results are shown up to the 50th harmonic, in accordance with the IEEE 519-2014 standard.
We found that for the most accurate results, it is necessary to consider the installation characteristics, conductor structure of a line, and the distribution of parameters along the line. If geometric data is not available during line modeling, using the line-type model is still acceptable. Its accuracy is mainly influenced by the length and the method used to determine the parameters R0', R1', X0', X1', C0' and C1'. The analysis and comparisons also revealed threshold lengths up to which the use of lumped parameters remains reliable. It turned out that using the script from the PSCAD library to calculate impedance characteristics is unreliable and leads to incorrect results.
|
