New Standards in Analyzing Stability

More and more decentralized power generating systems are feeding electricity into the distribution grid. Based on knowledge from the automotive sector, IAV has analyzed the stability of Q(U) control – a voltage control concept that has proven highly efficient for many years in conjunction with high voltage and extra-high voltage. The stability analysis carried out in cooperation with grid operator EWE NETZ GmbH on the basis of analytical models and methods sets new standards.

If ever more decentralized and volatile generators, such as photovoltaic systems, feed electricity into the power grid at low voltage in future, grid stability will be jeopardized. This calls
for system services that are provided by the power generators themselves. As a function of current mains voltage U, they must feed in active power as well as reactive power Q, expressed as the Q(U) functional relationship.

Solar energy systems currently do not have any intelligence of their own and are operated only on a controlled basis. Their inverters do not perceive the state of the environment they work in (state of the grid). “In future, decentralized power generators are to go smart and come with Q(U) control”, says Nils Koop, development engineer at IAV. “However, it will then also be necessary to ensure that the power grids maintain a high level of reliability and supply quality. For this, the stability and robustness of the open-loop and closed-loop control methods will be of paramount importance.”
Together with grid operator EWE NETZ from Oldenburg, IAV has examined the stability of Q(U) control proposed in standard VDE AR-N 4105 which is well-known from high and extra-high voltage. It has been the subject of intense debate among experts for some time now. Among other aspects, their misgivings are founded on the very high number and diversity of generators working side by side, on the adverse ratio of installed generating capacity and available grid connection capacity, particularly in rural grids, as well as on the level of measurement accuracy which, for cost reasons, is only limited.

Stability analyses of Q(U) control for distribution grids have so far been based on time series simulations, laboratory tests and field trials. Control stability has been assessed on the basis of a limited number of individual scenarios. Laboratory and field tests as well as simulation have been arriving at differing results, and there has been no generally applicable definition of stability limits. For this reason, IAV and EWE NETZ have analyzed the stability limits for the first time, using frequency response analyses common in control engineering.

Analysis focused on three different grid topologies of the type occurring in EWE NETZ’s grid territory. Topology I is a parallel feeder grid supplied from one side with a long transmission line at the end of which are installed a focal node in the form of several PV systems with high infeed capacity and an aggregated household load. This represents a rural distribution grid in which electric power is transported over long distances. Topology II is also a parallel feeder grid but, compared with topology I, characterized by a more even distribution of loads and feed-in points. It represents a rural grid with high in-feed from decentralized generators. Topology III is a meshed grid in which several PV systems with high power output and high loads are installed on two interconnected busbars. It represents a heavily loaded, closely meshed urban grid with a high demand for electricity. The three models comprise transformers, lines, PV systems and loads as their individual components.

“We have modeled all components of the overall system with differential equations as individual linear, dynamic systems”, Koop reports. “The individual dynamic systems are synthesized into an overall system by means of compositional modeling.” The modeled PV system is made up of the solar module, DC/DC converter (boost converter) and the inverter which feeds the electric power into the grid. On top of this, a maximum power point tracker (MPPT) is used to select the optimum operating point for the solar cell during system operation.

The system’s inverter measures the root mean square of the voltage and the phase angle at the grid connection point. In the Q(U) controller, a target reactive power level is generated from the values measured. This is governed by the characteristic curve of the prevailing mains voltage proposed in standard VDE AR-N 4105.
To examine the stability of Q(U) control, IAV and EWE NETZ analyzed the frequency responses. They illustrate the stability of a system on the basis of the Nyquist criterion. A control system is stable if the transmission function of the open control loop – in the form of the Nyquist plot in the complex plane – does not revolve around the critical point (–1, 0).

All topologies show results that are comparable in terms of quality. Given it mesh structure, topology III reveals a far greater stability reserve. As expected, topology I is the most sensitive. “With Q(U) control active, PV output is shown to be the dominant influence”, Koop reports. “Although line length also has a major impact on stability, the effects are less.”

Altogether, it was shown that although the low-voltage system could be destabilized in theory, it is far removed from this in practice. As a result, the study was able to reproduce and confirm the results of other investigations. It also shows that methods from automotive engineering can be applied successfully to power grids. Hence, IAV is already planning further projects on planning and operating medium-voltage power grids.