Introduction
The distributed energy generation shows a very rapid growth and reveals an increasing complexity for grids managers due mainly to prosumer sites, i.e. producer and consumer sites.
The intermittent nature of renewable energy sources, e.g. photovoltaic (PV) generator, remains an issue for their integration into the public grid resulting in: fluctuations of voltage and/or frequency, harmonic pollution, difficulty for load management. This leads to new methods for power balancing between production and consumption [].
Urban areas have great potential for intensive development of PV sources. To increase their integration level and obtain a robust power grid, the smart grid could solve problems of peak consumption, optimal energy management, and demand response. The smart grid is being designed primarily to exchange information on grid needs and availability, help balancing power, avoid undesirable injection, and perform peak shaving [].
In this context, at urban scale, the proposed system is a building-integrated DC microgrid, which provides a solution for the self-supply of buildings, electric vehicles, and grid-interaction control []. The DC bus can supply directly many building appliances (lighting, ventilation, electronic office equipment) as well as an electric vehicle.
The main scientific issue is the difficulty of global optimization due to the risk of mismatch between production/consumption predictions and real time operating conditions, on the one hand, and the need to take into account the constraints imposed by the public grid, on the other hand.
This paper presents the urban DC microgrid in Sect..
Urban DC Microgrid
The concept of smart grid appears and leads to microgrids for prosumer sites in order to reduce losses and peak energy demand, and also to play a role in local regulation, through the data communication. In urban areas, at the local level, the microgrid may be integrated to the building prosumer and connected to public grid by an adapted controller. At urban scale there are several building-integrated microgrids and parts of traditional public grid, all connected to the grid by a point of common coupling []. Intelligent switches are used to allow connection and islanding. Furthermore, a communication network is added, i.e. communication bus, whose routers are dedicated to direct messages following energy management priorities or special areas. Some dedicated controller interfaces generate and receive messages. The urban DC microgrid developed below is building-integrated and connected to the smart grid as described above.
Power Management and Optimization
The microgrid controller must provide the interface between the public grid and the loads (e.g. buildings, electric vehicles), aiming an optimal power management.
Figure (forecasting, smart metering, monitoring) a specific interface associated with the urban microgrid was designed as proposed thereafter.
Fig. 1
Power management interface principle
The developed microgrid controller presented in Fig. ]. The obtained results are the optimal power evolution of each source for which the total cost is the minimum for the considered time duration. These powers cannot easily be implemented in real-time control. The solution is to translate the power flows into a single interface parameter for power balancing control, which is the predictive control parameter, one of the outputs of this layer. The second output concerns the predictions to be transmitted to smart grid (injection and supply). The predictive control parameter is applied in the operational layer, which algorithm controls the power balancing in the microgrid system.
Fig. 2
Microgrid controller
The algorithm provides real-time references of the system powers and the coefficient of possible load shedding.
For urban microgrids several operating strategies are developed based on sources that make up the microgrid (PV sources and wind turbine, storage, public grid connection, micro-turbine or bio-diesel generator) and loads (buildings electric loads and electric vehicles charging stations). Figure presents the main possible strategies. Renewable energies supply the building and charge the electric vehicles. The renewable excess energy could be stored and/or injected into the grid. The grid, if available, is used only as back-up for the building and the electric vehicles. The micro-turbine operates only if the grid is not available. The electric vehicles, if required for stringent situations, can supply the building and/or provide energy to the grid. The messages received from the smart grid command the microgrid operating mode aiming compliance the actual availability of the grid.
Fig. 3
Energy management strategies for urban DC microgrid
Building-Integrated DC Microgrid for Grid-Connected Mode
The DC microgrid given in Fig. consists of physical power system and its controller as presented above. The power system includes a DC load and sources which are connected on the DC bus through their dedicated converters, while the DC load demands power directly from the DC bus.
Fig. 4
DC microgrid building-integrated for grid-connected mode
The power balancing control principle and constraints are presented following the power flow schema shown in Fig. ]. The proposed strategy is to operate with the minimum energy cost for the considered period. Within the given limits, the public grid can supply or absorb energy to or from the microgrid. The same applies to the storage, charge or discharge operating mode. The DC load can demand power up to its maximum power, but limited power can be a load shedding result. For the PV array (PVA), two controls are implemented: a maximum power point tracking (MPPT) control to extract the maximum power and a limited control to extract a limited power to meet power balancing for some stringent situations. The power balancing shows that the adjustment variables are the public grid and storage, within their physical and functional limitations. The predictive control parameter decides the contribution of these two sources, grid and storage. This control parameter must be the image of the power flow optimized in energy costs. The energy cost optimization take into account the day-ahead forecasting of the PVA production as well as the load power demand. Combined with this robust power balance strategy, the energy cost optimization is formulated as minimization of the total energy cost with respect to system physical constraints and imposed limits. To minimize the energy cost, energy tariffs are imposed as follow. The storage can be used as often as possible; an arbitrary but lowest tariff is given. In order to avoid the two operations, very penalizing energy tariff is proposed for PVA power limiting and load shedding. Public grid tariff is suggested to be lower than the PVA or load shedding tariff. There are two grid tariffs: peak hours and normal hours. The end-user can accept certain amount of load shedding.