Development of distributed control strategies for power electronics systems to provide network services

A major topic in the Asimute project is future efficient and low-carbon energy systems. Previous fossil energy sources are replaced by renewable sources, primarily solar and wind, which are exclusively converted through power electronics. On the consumer side, slow conventional devices and previous rather inefficient power supplies with large-inductance transformers or mechanical elements, such as grid-connected motors, make also way for highly efficient power electronic circuits. However, power electronics are actively feedback controlled, have zero inertia but respond immediately. Conventional power electronics cannot stabilize the grid but expects a near-perfectly sinusoidal voltage with stable and solid amplitude and frequency. The voltage should further not vary with load. Given such a stable voltage, power electronics draws or injects a current but does not directly control the voltage. However, the elements that provided a stable voltage were large massive generators from fossil power plants, which are retired. The effect will be unstable networks with drifting frequency globally in the entire network, and locally unstable voltage, which can lead to local black-outs and damage devices. The latter will particularly affect more rural areas.

In this project stream, we intend to develop converters and corresponding control methods that can stabilize future grids, both on the higher levels of the grid as well as the lower ones that are important for the supply of households and most small and medium-size businesses.

Innovative distributed control strategies are required for the coordination of power electronic systems for the provision of grid services such as frequency and voltage stabilization and distortion filtering. In this measure, the theoretical foundations for such control strategies are developed and tested in simulation and in the demonstrator. In particular, topologies for the informational linking of the power electronic systems will be developed and evaluated with regard to the communication effort, reliability and cyber security. Methods for exchanging information on system dynamics, which are commonly used today for frequency control, are also considered with regard to their applicability for other system services. Based on the topologies, distributed control algorithms and associated communication protocols are created and investigated.

Especially the medium voltage distribution grid and the local low voltage grid in several ways (voltage tolerance violation due to high power solar feed-in, loss of short circuit power, increased grid impedances, unwanted and unknown power flows due to together with future high-power vehicle chargers). While the control of conventional power electronics typically becomes unstable under these conditions and further exacerbates the other problems mentioned, adaptive control methods (grid impedance estimation without violating IEC61000, adaptive controller adaptation, and damping of parasitic feedback with other electronics on the grid) and compact modular electronic circuit topologies (modular partial power converters and power flow controllers) are being developed which, on the one hand, allow renewable energy feed-in, vehicle chargers and the like to be operated stably even under weak grid conditions, while still increasing their penetration of the grid and, if possible, even partially eliminating the problems.

The grid services mentioned above currently appear to be relevant above all for a smaller number of more powerful distributed units such as large PV systems and above all grid storage systems and focus in particular on global grid problems (e.g. frequency stability). The large number of smaller power electronics, especially in households and rural areas, allows typical problems to be improved. This measure develops grid impedance-adaptive dynamic voltage stabilization for appropriate phase angle of the feed-in and voltage filtering based on the grid impedance estimation from the aforementioned measure and tests it in a typical low-voltage grid environment.