Impact of bidirectional charging stations on the network: a Power Hardware in the Loop implementation

A characteristic common to every power system is the need to balance the energy demand and the electricity supplied. In order to guarantee this balance, a dedicated market, the Ancillary Services Market, is run to ensure the continuity and security of the system. Today, these regulation services are mainly supplied by large fossil-fuel power plants, but with the diffusion of renewable resources and the decarbonisation of the energy production, new regulation service resources will be required. Hence, in the last years an opening process of the ancillary services market to distributed resources has been launched. In agreement with the Italian Regulatory Authority for Energy, Networks and Environment (ARERA), some pilot projects, aimed at measuring the performance of these resources, have been defined. Terna, has identified the following pilot projects as being particularly innovative: - Virtually Aggregated Consumption Units (UVAC) - Virtually Aggregated Production Units (UVAP) - Virtually Aggregated Mixed Units (UVAM) - Relevant Production Units (UPR) not subject to mandatory participation This thesis carries out some tests on a system under the UVAM statements. This case study proposes a model of a car park with 20 bidirectional charging stations, divided in 5 groups of 4 wallboxes. Each charger is capable to inject $10kW$ under V2G operations for a total power equal to $200kW$, which is the minimum requested power for an UVAM based on electric vehicle charging stations. In order to evaluate the impact and performances on a distribution network, the Real time simulation, with power hardware in the loop configuration, is used. In this simulation layout, the medium and low voltage distribution network is simulated on the RTS, while a real charger is connected to the model through a power amplifier interface and three current probes to fed back the currents into the simulation. The injected currents are then multiplied to evaluate an higher number of wallboxes. The first part of the work is based on how the simulation should be configured and how the model, laboratory set-up and time delays influence the stability of the system. Through this studies, a more straightforward method to configure a stable model is proposed, starting from the interface algorithm till the input/output signal processing. Secondly, with a stable laboratory set-up, some simulations are run to evaluate voltage levels, disturbances on the grid side and power losses through the low voltage lines. These tests are aimed at evaluating the feasibility of this kind of regulation technology, based on an high number of charging stations. As expected, the voltages on the low voltage side are lower than the nominal value during G2V operations, while they are higher in V2G. Some analysis are also done on the power losses on the low voltage side, following which some hypothesis on the possible microgrid topologies could be formulated.