Advanced Control of Power Converters for Efficient Use of Distributed Energy Resources in Future Smart Microgrids
Year: 2012
Type: dissertation
Abstract: This work proposes a vision on the growing scenario of Smart Grids, where a pervasive introduction of distributed generators and innovative power management schemes forces a deep review of the power distribution network establishment. Despite a potentially general value of the proposed approach, the analysis is focused on low voltage microgrids as the most flexible and open architectures to demonstrate the feasibility of the network renewal. In recent years, distribution network has experienced a massive introduction of Distributed Generators (DGs). So far, this has been done based on a hierarchical government of the network, where utilities set regulations and keep control of the connections from the top of the structure. The final user has a passive role in the management, even participating to generation, owning a DG unit. This is basically due to historical reasons: the distribution architecture has been built starting from a centralized controller driven paradigm, and all the updates have been done in the same direction. This approach has been valid until the electrical generation was supplied by a limited number of large plants: huge amounts of power with all the related control challenges, but static generation infrastructure. The upcoming scenario is instead a generation that is strongly decentralized toward distributed energy resources. At the same time, development of Information and Communication Technologies (ICT) is constantly rising. The combination of these two phenomena has the potential to change completely the power distribution system, both in the architecture and in the roles of players in the energy market. From an architectural point of view, distributed generation units could become a combination of renewable sources such as solar, wind, small hydro turbines, fuel cells etc. and traditional sources like gas or diesel generators required for supply continuity of sensible loads. Coupled to the primary energy sources, Energy Storage (ES) can be used to introduce a degree of freedom in energy management: large storage units like batteries or flywheels can absorb the excess of generation, to provide energy during peak demand times, or perform more complex optimizations. Small energy sources, like super capacitors, can be used to improve power quality during transients such as voltage sags or frequency variation transients. These generation and storage units shall be equipped 6 6 with an intelligent digital controller able to measure all the required local variables and perform local control of power converters and also capable to communicate in a bidirectional way, for example via Power Line Communication, with other generators. The power topology of these conversion units can be generally represented as a first conversion stage, dependent on the kind of energy source, and a second stage that is a current controlled inverter. Such devices are named “Energy Gateways” (EG) and represent the basic elements of future Smart Grids in the vision presented in this work. To limit the system complexity, a microgrid is considered: the microgrid is connected to the main low voltage distribution network and presents a number of distributed generators, all equipped with EG. In the proposed approach control of the micro grid is distributed, without centralized controllers: the microgrid internal behaviour can be optimized, as well as the microgrid behaviour seen by the main grid, based on EG to EG communication and control architecture. Considered the degrees of freedom of the resulting system, different optimizations could be performed. In more details, the optimization considered in this work is the distribution loss minimization and it is obtained by properly controlling converters active and reactive power references in a distributed way. In particular, the normal overrating of power converters opens the possibility of injecting distributed reactive power, that can be locally delivered to the loads, reducing the absorption from the main grid and the consequent losses and voltage drops. Also the active power can be partially controlled, depending on the availability of energy storage or controllable generation units. Different techniques have been proposed: stand-alone Energy Gateways with power references control based on local measurement only, without need for communication, or distributed solution where each EG communicates with the surrounding EGs, computes its optimum local power reference and leaves the control to another EG with a Token Ring logic: iterating the local optimization, the system converges to a global optimum of the loss. The optimization is a constrained optimization, depending on power availability, limited by converters power ratings and by distributed generators and storages, and the effects of these constraints are taken into account in the distributed optimization analysis. The analysis have been performed first analytically and then by simulation, developing a specific set of Matlab scripts that gives a flexible tool to define and test every microgrid and its generators and loads, over which the distributed control algorithms have been developed. Some of the major limits of these techniques have been addressed, among them the need for transmission of synchrophasor with tight real-time requirements. At the end, a sub-optimum controller is proposed, expecting to overcome the over mentioned limits that will be developed in the continuation of this research activity. Abstract 7 7 In parallel to this first topic, a second more specific problem has been investigated, representing another aspect of the Smart Grid paradigm. As part of a visiting period with the PEMC (Power Electronics Machines and Control) group at the University of Nottingham, the combination of a STATCOM with energy storage and an engine powered synchronous generator has been studied to smooth the frequency variations during sudden load changes. This represents one of the applications of limited energy storage with high dynamic capabilities, like supercapacitors and is of interest both as retrofit of existing synchronous generators, that could experience disconnection problems in a more dynamic environment represented by the Smart Grid, and as a way of ensuring power quality in a microgrid fed by a backup generator. The first application requires a radical increase in the power size of the system, while the second one is more centred on Smart microgrids: independently on the control approach, a microgrid has to be designed to support islanded operation, i.e. to be able to feed loads even when disconnected from the main power network, and one of the options for this backup supply of the microgrid is the use of synchronous generators as emergency voltage sources, typically with diesel engines as prime movers. This second scenario has been considered and studied and an innovative control technique has been proposed able to ride through step changes in the load demand without changes in the prime mover speed and therefore without large changes in the voltage frequency. This has been achieved through fast active power injection from the energy storage as soon as the load transient is detected. From a qualitative point of view, the new load power demand is supported by the storage until the synchronous generator prime mover overcomes the related transient. This technique has been investigated in simulation and validated in a 10kVA experimental setup, representing a simple isolated microgrid with a single generator and resistive loads, confirming the effectiveness of the proposal.
Author: Alessandro Costabeber
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Keywords: smart microgrids, distributed energy resources, distributed generation, converter control, energy storage, power quality, power line communication, two-way communication, inverter control, energy management, low-voltage microgrid, decentralized generation, supercapacitors, voltage sag, frequency transients
Open Access status: green
OpenAlex ID: W2275196313
