PV integrated to distribution grids with high renewables integration
The main idea is to define and develop technics of DG clustering.<br />These technics aim to create a set of microgrids with a minimum of assets able to communicate , to provide reserves and interacting with the main grid.<br />The various steps are :<br />- Definition of the multi cluster architecture<br />- Identifying the various renewable, electric vehhicles whose can participate to frequency control and voltage control.<br />To propose strategies of control of the various assets and the energy fluxes. The goal is to control the frequency and the various voltages depending on considered nodes.<br />- Studying the dynamic behaviour during operation changes between islanded and connected modes<br />- Testing the proposed solutions in laboratory and in situ<br />- To assess technicaly and economicaly the various developed startegies<br />- Recomendations and specifications
Modeling the basic elements of teh microgrid
Data colation from the operator SOREA
Control implementation on the Socomec's VSC's
Build the energy management algorithms
Tests on laboratory and in situ
After the strucuration of the project, the first action was a state of the art on the criticity in microgrids : renewables integration, the classical storage and the problematic of the Vehicle to Grid (V2G), the protections and finally the centralized and decentralized controls.
The algorithms regarding the components control have been designed and tested in simulation :flywheel, batteries, electric vehicle ...
A part of these controls have been adapted and implemented on real devices : SOCOMEC's 33 kW VSC's.
Concerning the energy management, a first linear approach has been applied  to calculate the voltages at the different nodes of a residential grid with a high degree of EV's penetration. These results are exploited t o develop algorithms based on a mixte dynamic strategy . Based on a decentralized stategy the charge of the various EV's is performed using non linear programing methods . The economic aspects are investigated as well .
On other hand, a real time optimization of the whole energy fluxes was done with a multi-agents approach. To have a wide sight on the optimal power flows, a centralized approach was performed taking into account the random characteristics of the wind generators and solar PV productions.
The actual investigations are focusing on building simple and accurate models; these models are necessary for the OPF based on multi-agents
The architectures are fixed and the next steps will be focused on the developed algorithms validation on real time simulator and in situ
3 international journal (IEEE Transactions) another one at second lecture.
10 International conferences
The distributed generation is growing hugely due to kingly decisions at national levels. These distribution means in particular those based on renewable sources produce their energy without coordination with the consumption.
This production was marginal in the last decade and not neglictible nowadays. The distributed generation reaches 30% in some countries (Netherlands), even more than 50% in some particular cases. Some American studies plan an upper penetration rate in the middle of this century, other studies in Europe plan from their side a 100 % renewable energy based grid (EREC RE-thinking on 2050).
The grid operation integrating a high rate of renewable energies (almost non-controllable sources, difficult to predict, which could have negative impacts and dysfunctions) requires to develop an intelligent network (Smart Grid) which will be characterized by a specific algorithmic and physical architectures; the energy management will play a key role as well. It is the objective of PARADISE project.
PARADISE is a long term anticipating approach regarding the grid as it will be seen with high rate of renewable, intelligent loads, Plug-in Hybrid Electric Vehicles (PHEV) . The storage can take the shape of embedded batteries inside the PHEV’s or distributed stationary ones. By emphasizing the system approach, PARADISE will allow a global overview of a new architecture and a distributed management approach.
Monsieur Bertrand RAISON (Laboratoire de Génie Electrique de Grenoble)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
AER AER ALCEN
G2Elab Laboratoire de Génie Electrique de Grenoble
CEA-INES Commissariat à l’Energie Atomique et aux Energies Alternatives / DRT/ LITEN/DTS-INES
GIPSA Grenoble Images Parole Signal Automatique
LBMS Laboratoire Brestois de Mécanique et des Systèmes
SOCOMEC Société SOCOMEC
Help of the ANR 758,480 euros
Beginning and duration of the scientific project: November 2013 - 42 Months