In the eFORT project, Fraunhofer EMI develops dynamic simulations that contribute to increase the resilience of electricity grids. The methods are demonstrated using a distribution grid in the Sarentino Valley (Italy), which is operated by Edyna.
The solutions developed by the project partners during the course to the project are demonstrated within four pilot cases located in Spain, the Netherlands, Italy and Ukraine. The contribution of Fraunhofer EMI focuses on the third pilot case located in South-Tirol (Italy). The specific power grid under consideration is a remote MV/LV distribution grid in the Sarentino Valley, which has a high proportion of hydropower.
Due to the geographical location of the grid under consideration, there is a high vulnerability to being disconnected from the superimposed grid, but on the other hand, due to the high proportion of hydropower, there is a possibility of being operated as a grid island, at least temporarily especially in summer time. In addition to the quasi-dynamic investigation of island operation, grid stability with a particular focus on frequency stability is also examined with the help of dynamic RMS simulations. A particular challenge here arises from the low inertia of the distribution grid, which is made more difficult by the increasing number of prosumers.
Result and Discussion
Based on the close and efficient cooperation with all partners, in particular with EDYNA and SELTA-DP, work was initially carried out on the development of a dynamic model of part of the distribution grid. In addition to the lines/cables, the model contains static generators, loads and dynamic generators, which were modeled as synchronous generators. In a first step, standard models and fictitious controller and machine parameters were assumed for voltage and speed controllers. Yet no protective relays were taken into account. All loads are modeled with a constant impedance. The grid model is now available as a PowerFactory model and enables the simulation of the dynamic behavior of the grid, for example in the event of a disconnection from the superimposed transmission grid or a significant load step. A further version of the model has been created for use with a dynamic RMS, which is also being developed by Fraunhofer EMI as part of the eFORT project and is continuously validated against commercial simulation software tools such as PowerFactory and Neplan.
The following example shows exemplary simulation results based on the dynamic model of part of the distribution grid under consideration. The grid, schematically illustrated in Figure 1, contains a number of switches that enable lines to be switched off and the grid to be split into two parts. In the example, some switches are open so that two separate grid parts are created, each of which is connected to the superimposed grid. Due to the generation from biogas, hydropower and photovoltaics, grid A (green) has a power surplus, while grid part B (blue) must be fed from the superimposed grid.
At time t =1 s, both grids are disconnected from the superimposed grid, resulting in the traces shown in Figure 2 for the respective frequencies and the aggregated turbine outputs for the two sub-grids.
Due to the failure of the external grid, not only the active power flows but also the (negative) reactive power flows go to zero (see Figure 3), so that the voltages initially rise sharply. Assuming loads of constant impedance, this results in an increase in load power. This temporary increase in load power is reflected in an initial drop in frequency and an increase in the aggregated turbine power in grid section A, even if this is statically oversupplied. In the long term, the frequency stabilises above 50 Hz and a reduced aggregated turbine output is obtained.
Outlook
Further steps that will be performed in the following project period are the adjustment of model parameters in cooperation with grid and generation plant operators for a better approximation of the real dynamic behaviour and their validation. The final goal is to evaluate the frequency stability of the distribution grid in island operation with regard to significant load steps and generator failures. The RMS codes developed at Fraunhofer EMI will also be used here. Established measures such as the temporal rate of change of the frequency immediately after the event (ROCOF) and the minimum frequency passed (NADIR) will be used as evaluation criteria.
While the prediction of the frequency mainly depends on the inertia of the grid and the magnitude of the active power imbalance, the evaluation of the angular stability, as another important stability phenomenon, requires an accurate dynamic description of all generators and reactive power generating components. With increasingly accurate modelling of the grid components, e.g., by replacing standard parameters for machines and controllers with real parameters, the prediction accuracy of the rotor angles of the individual generators also increases. The temporal traces of the rotor angles, in particular their temporal and local scatter, then enable a reliable assessment of the angular stability.
Conclusion
Being able to predict the dynamic behaviour of the distribution grid for a list of credible contingency scenarios like islanding, generator failure, line failure and others, and through the subsequent assessment of the evolution of frequency and rotor angles, the operator gains a powerful tool to improve the stability of its grid during the grid planning phase and operation
Contact us
Follow us on:
