Elektromechanische Modellierung aktiver Verteilungsnetze zur Analyse der transienten Systemstabilität

Erlinghagen, Philipp; Schnettler, Armin (Thesis advisor); Bernd, Engel (Thesis advisor)

Aachen (2019)
Dissertation / PhD Thesis

Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2019


Current energy policy goals lead to a redesign of the electrical energy system, which results in an increasing substitution of conventional power plants by distributed energy resources. The distributed systems are usually fed by renewable energies and they are usually installed in the distribution grids. To sustain the overall system stability and to prevent blackouts in the European transmission grid, analyses of transient stability, among other stability aspects, are necessary. The amount of information as well as the modelling and computational efforts for distribution grids must be kept as small as possible, using equivalent models. In this work a method for the dynamic modelling of reference systems that can be parameterized in a stochastic way. Using these reference systems, systemic key data can be identified which is used to parameterize the equivalent systems. For this purpose, laboratory tests are performed to identify realistic stochastic parameter spaces. The stochastic component modelling that is developed within this thesis is successfully tested on pure machine systems, machine systems with controllers, power electronic systems and hybrid systems. The approach allows a much more realistic modelling of reference systems than before, where only single grid sections with fixed component parameters are used. With the help of a more detailed overall equivalent model than in previous works, new features can be included in the equivalent systems. For example, different specifications of the reactive current droop control within one grid, or geographical as well as manufacturer-specific clusters can be modelled. Using a variance-based sensitivity analysis, systemic key data is identified that can be used to parameterize the equivalent model: the primary indicator is the peak power of the infeed and loads. The secondary indicator is the cumulated frequency of the plant sizes, which can be used to increase the quality of the solution and to decrease the safety margins. The dynamic parameterization as a novel approach is validated against synthetic grids and a real grid. The results show that the models themselves and the whole method is valid. Using simple functional dependencies, generic equivalent models can be parameterized using a small amount of systemic information on the grid. The only exceptions are the coupling impedances. They need to be modelled as impedances with an exponentially increasing dependency on the voltage. If enough information on the grid is available, approaches for initial parameterization, which are further developed in this thesis, can be used to increase the quality of the solution.