Abstract / Description of output
As we move to a decarbonised energy supply, the profiles of supply and demand will change significantly, and possibly rapidly. Major aspects on the demand side include the electrification of transportation and of heating, while on the supply side there is the integration of high proportions of renewable generation. On both the supply and demand sides there will be a large number of smaller resources distributed throughout the system1.
At present, system-wide architectures for integrating these changes are focused on the transmission operator, even though the new small resources are mostly at the (local) distribution level. Because coordination with distribution constraints is typically not part of these architectures, the interactions with the transmission operator usually do not consider distribution constraints directly. However, distribution constraints must be properly taken into account when the volume of distributed resources has grown to a significant scale with respect to the part of the system to which they are connected.
Thus if we have flexible distributed resources on the relevant macro scale, a decentralised architecture for system operation will be essential. This includes agreement on the operational rules to ensure that actions taken by network operators and users (on the basis of a common set of data and control/price signals exchanged) result in consistently stable and reliable system operation, and a good overall outcome for the system with respect to a chosen objective. It is implausible that a single central optimization across all levels of the system with possibly tens of millions of flexible resources will be feasible, because of the inherent complexity including the huge data and computation requirements. The natural alternative paradigm is to consider the system at different levels, to have a series of aggregation points at connections between levels of the system, and to have the required control signals passed both ways across these points.
This paradigm implies relatively limited computation at any one level of the system. Indeed, the transmission operator scheduling challenge might become simpler than it is at present if, thanks to a decentralised architecture, it sees a single virtual resource at each grid supply point (GSP)2. The number of entities that it interacts with directly will then be limited by the numbers of transmission-connected resources and GSPs.
Details of such a hierarchical system and of the transition to it may depend on the present combination of market design and scheduling mechanisms. Issues to consider include: whether a system has automatic generation control; detail of design of the real time market; to what extent operation is currently automated; and whether there is a single control room for network and energy balancing.
At present, system-wide architectures for integrating these changes are focused on the transmission operator, even though the new small resources are mostly at the (local) distribution level. Because coordination with distribution constraints is typically not part of these architectures, the interactions with the transmission operator usually do not consider distribution constraints directly. However, distribution constraints must be properly taken into account when the volume of distributed resources has grown to a significant scale with respect to the part of the system to which they are connected.
Thus if we have flexible distributed resources on the relevant macro scale, a decentralised architecture for system operation will be essential. This includes agreement on the operational rules to ensure that actions taken by network operators and users (on the basis of a common set of data and control/price signals exchanged) result in consistently stable and reliable system operation, and a good overall outcome for the system with respect to a chosen objective. It is implausible that a single central optimization across all levels of the system with possibly tens of millions of flexible resources will be feasible, because of the inherent complexity including the huge data and computation requirements. The natural alternative paradigm is to consider the system at different levels, to have a series of aggregation points at connections between levels of the system, and to have the required control signals passed both ways across these points.
This paradigm implies relatively limited computation at any one level of the system. Indeed, the transmission operator scheduling challenge might become simpler than it is at present if, thanks to a decentralised architecture, it sees a single virtual resource at each grid supply point (GSP)2. The number of entities that it interacts with directly will then be limited by the numbers of transmission-connected resources and GSPs.
Details of such a hierarchical system and of the transition to it may depend on the present combination of market design and scheduling mechanisms. Issues to consider include: whether a system has automatic generation control; detail of design of the real time market; to what extent operation is currently automated; and whether there is a single control room for network and energy balancing.
| Original language | English |
|---|---|
| Publisher | Catapult Energy Systems |
| Media of output | Online |
| Publication status | Published - 1 Jul 2023 |