Building a test-bed for Frequency Control in the Lab

State of the Art

Frequency control is currently an issue managed by TSOs in collaboration with large power plant operators. The first change in this scheme has been triggered by Virtual Power Plants which led to the concept of aggregation to also bring smaller players into the overall energy automation picture. In future, assuming a 100% RES scenario, frequency will have to be managed in a completely different way and this is one of the main topics of investigation of RE-SERVE. The main issue however is that frequency, as considered today, is a global quantity and hence a small trial cannot give enough information. However, new frequency measurement and assessment technology will be demonstrated in a laboratory testing environment. Using the infrastructure developed within the project, it will be possible to create a complex scenario linking computational capability across 4 countries: Italy, Germany, Ireland and Romania. Nevertheles, the availability of an advanced monitoring system from Romania will allow the incorporation of real data in the study performing a continuous validation process in support of the modeling effort, while taking into account all privacy and market operations constraints.

Envisioned Scenario for the test

A network model of suitable complexity will be defined using the RE-SERVE network of laboratories. The model will be close to realistic scenarios given the real network data that is available at the RE-SERVE research institutes (RWTH, UCD, POLITO, UPB). The model will be able to present a good validation of the frequency control concepts as defined in the theoretical research of RE-SERVE. It will be enriched by measurements collected in the Romanian grid21. Transelectrica will provide relevant and selected historical real data; collected under real-time conditions from the selected tie-lines (import-export with neighborhoods) and from selected renewables connected on Transelectrica network (at least two renewables connection points), and from at least one group of a large hydro-plant in Romania. These data will come from the SCADA of the TSO (with a refresh rate of 2 seconds or more) and from smart meters in the concept of Unbundled Smart Meter – USM, developed in Nobel Grid project – with which Transelectrica has a scientific collaboration, which allows 1 second based real-time data + measurement of ancillary services in a meterology based environment. Moreover, UPB will provide real data from its 5 mobile PMUs installed in selected nodes of distribution network. Transelectrica will use also some of its network points in the network where they installed high performance PMUs (a set of its 11 points) in the study, using recorded data regarding angles and ROCOF information reported with the highest rate (50 frames/s). Daily data from both Transelectrica and UPB will be input for the simulation tool of the project, to refine different new methods for stability based on frequency, ROCOF and also potentially on active power [aggregated] information. TransE will also participate from the user perspective in designing the new methods of frequency control, based on flexible / adaptable synthetic inertia, as core novelty promoted by the project. In addition, measurement of ancillary services using new smart meters as well as associated Quality of Service will be part of this activity, to be used in combination with WP2 activity.

Approach 

The four laboratories will be managed as a single grid.  The generation units will only be based on renewable sources with a suitable combination of PV and Wind.The sources will be equipped of appropriate control as a result of the theoretical work on Linear Swing Dynamics. The control will be implemented in real HW so to keep the realism.  Communications will be implemented emulating performance and behaviour of a futuristic 5G installation thanks to EDD so that also communication requirements can be verified.  Different types of small and large perturbations will be introduced to analyse the dynamic behaviour of frequency under different conditions in relation to the original system level requirements.