By Jeffrey Taft and Ronald Melton, PNNL, and Dave Hardin
(Note: this is Part IV in a series. The previous articles can be found at this link.)
Grid coordination is the operational alignment of assets to provide electricity delivery. Although the grid has always relied upon effective coordination, it has become a higher priority issue for electricity distribution because of the rise of distribution-connected distributed energy resources (DER) such as solar, batteries and EVs (electric vehicles), that are not owned or directly controlled by a utility. Some theories for resolving these DER coordination issues are being discussed but few rigorous models or formulas exist. Grid architecture addresses grid coordination from a structural perspective using a methodology built upon a firm mathematical foundation.
The previous article focused on industry structure and how distributed energy resources can cause hidden coupling and result in undesirable side effects when not coordinated between transmission and distribution operations. These side effects were addressed by analyzing three structural alternatives which directly impact how grid entities interact with each other and coordinate operational activities. These alternative models include; Total TSO (Transmission System Operator), Hybrid DSO (Distribution System Operator) and Total DSO. Let’s take a closer look at grid coordination structure as it relates to these models.
Definitions
To understand coordination in relation to the grid, it is useful to start with some definitions:
- Centralized system – multiple separate entities that operate under the maximum supervision and control of a central coordinator that is responsible for solving a common problem (see top of diagram 1). An example is an aggregator directly controlling the loads within its domain.
- Decentralized system – multiple separate entities operating independently with minimal supervision and coordination between components (see top of diagram 1). Note that decentralized systems are similar in structure to centralized systems but differ in the level of coordination between the central coordinator and the separate entities involved.
- Distributed system – a decentralized system in which the parts cooperate to solve a common problem (see bottom of diagram 1).
- Coordination – the act of different people, organizations or things working together under an agreement to produce a goal or effect. This includes the means by which a set of decentralized elements cooperate to solve a common problem, thus becoming a distributed system. As a result, coordination structure is a key aspect of distributed systems and distributed control.
- Grid coordination – the systematic operational alignment of utility and non-utility assets to deliver electricity.
Diagram 1 – Decentralized (top) and Distributed (bottom) Systems
Coordination models
Two structural coordination models used to compose and coordinate services (e.g. web services) into complete workflows are centralized and peer-to-peer (see diagram 2). The centralized model uses some type of orchestrator (coordinator) to invoke and synchronize pre-determined services. The peer-to-peer model uses choreography to accomplish the same thing but requires the services to directly participate in service synchronization which is real-time and dynamic. Distributed systems typically use some form of peer-to-peer interaction and communication.
Diagram 2 – Coordination using Orchestration and Choreography
Coordination and control
Coordination and control are related, but different, concepts. Coordination is the act of different people, organizations or things working together under an agreement to produce a goal or effect while control, a subset of coordination, is direct influence or authority over people or things through the power of regulation and restraint. The level of autonomy a system allows a component to exercise, or self-govern, determines whether control, or coordination, is most appropriate. Control systems are typically implemented within a common governance framework (e.g. distribution SCADA within a utility) whereas coordination is typically synchronization between self-governing entities (e.g. energy/power/ancillary markets).
A structural basis for system coordination
Coordination schemes in operation today have evolved over time to meet a variety of needs such as dispatching generation, ancillary services and load, complying with regulation, and enabling markets. These networks are critical for grid operations, but they may not be adequate for a future of high penetration DER.
An electrical grid is a real-time system with tightly coupled constraints. Supply and demand must always be in equilibrium. A Transmission System Operator maintains this balance using centralized system optimization techniques that achieve objectives such as cost minimization while maximizing reliability. This includes vertically integrated and other utility structures as well as ISO/RTOs (Independent System Operators/Regional Transmission Organizations). These techniques work well today – but how can they be adapted for a distributed future?
Distributed coordination network implementations should be developed from architectural structures that are contained within a conceptual framework that is founded upon a rigorous and formal mathematical basis. This ensures that the network has the foundation needed for robustness (see diagram 3). The network implementation consists of actual communication, sensing, computation along with decision/control elements such as message flows, interfaces and coordination signals. These elements need to reside within a coordination structure that is derived from a conceptual framework that enables a class of solutions, not a specific solution.
Diagram 3 – Developing a Coordination Network
A conceptual framework is needed that enables a primary optimization problem to be decomposed into secondary problems and those secondary problems to be decomposed further into sub-problems as needed. A well-known mathematical theory called “layered decomposition” has been developed that enables this type of optimization decomposition, “Layering as Optimization Decomposition, A Mathematical Theory of Network Architecture.” (see diagram 4)
Diagram 4 – Mathematical Basis for Layered Decomposition
Layered decomposition and grid coordination structure
Applying the concept of layered decomposition to the grid induces a multi-layer structure, called “Laminar Coordination Structure” (see diagram 5), that has a chain of “coordination nodes” that enable scalable message flow between core repeating building blocks called “coordination domains.” The flow of information through this decomposed structure is smooth and follows regular paths.
Diagram 5 – Laminar Coordination Structure using Layered Decomposition
Each layer of the hierarchical structure (i.e. coordination domain) contains controls and connectivity that solves a specific sub-problem. The constraints on the solution of the sub-problem are communicated between layers through the coordination nodes. This structure helps enable distributed controls by: 1) ensuring that control roles of each layer are well-defined and are aligned with the overall system, 2) improving resiliency by allowing devices to act independently in real time based on immediate operational conditions, and 3) minimizing and structuring message traffic to facilitate scalability and interoperability for large numbers of connected devices with mixed ownership models. The consistent structure helps standardization efforts address integration challenges through the development of open industry standards
A coordination domain is a group of components that reside at a specific layer in the layered decomposition structure. Each domain has a single coordination node and all actionable information between coordination domains occurs through coordinator nodes.
As an example, transmission/distribution coordination could be implemented using laminar coordination and coordination domains as shown in diagram 6. This shows layered decomposition applied to a Total DSO model of the grid. The TSO communicates with each DSO through a coordinator node which then communicates with other coordination nodes that reside within the utilities local service area. The actual mapping of coordination domains to grid areas must be determined by the DSO.
Coordinator nodes (see diagram 7) have common structure and interfaces but adapt to the requirements of a specific coordination domain. The common structure and interfaces are required to enable high levels of cost-effective scalability and interoperability.
The signals communicated between layers determine the type of coupling between the layers. The two kinds of signals are referred to as “primal” and “dual.” Primal coordination signals are essentially resource allocations whereas dual decomposition signals behave like prices, even if money is not involved. These signal types are a result of using math to induce a structure for coordination whose structural properties are well-understood.
An example of a coordination domain communication framework is Open Field Messaging Bus (OpenFMB), which provides a distributed framework for integrating intelligent grid edge devices, points of common coupling for microgrids and other intelligent systems. Coordinator nodes communicate through interoperability interfaces that adhere to existing or new open industry standards. A related initiative within the national labs is the Grid Modernization Laboratory Consortium (GMLC) Interoperability project.
Diagram 6 – Coordination Domains and Laminar Networks
Diagram 7 – Coordinator Node in a Coordination Domain
Mapping layered decomposition to the grid
The previous article discussed three models, “Total TSO,” “Hybrid DSO,” and “Total DSO”, that help address the hidden coupling problem.
The Total TSO model closely resembles the grid of today. The distribution utility follows the dispatch decisions of the TSO but otherwise operates independently from the transmission system operator.
The Hybrid DSO model is a shared model and requires more complex coordination and communication between the transmission system operator and the distribution system operator in addition to 3rd party DER aggregators. Both the Total TSO and Hybrid DSO models do not conform to a layered laminar coordination structure.
The Total DSO model (see diagram 8) does conform to layered coordination structure and represents an optimal approach for high DER integration.
Diagram 8 – The Total DSO Model Mapped to Layered Decomposition
Transmission/Distribution (T/D) Coordination and the Total DSO Model
The Total DSO model requires that transmission and distribution system operators share the common goals and objectives of electric grid balancing optimization and reliability. They need to cooperate and coordinate continuously with each other through mutually understood and clearly demarcated roles and responsibilities. The TSO is solving the primary system optimization problem for many utility service areas while the distribution utility solves the secondary optimization problem within a local service area. These must always be aligned and synchronized.
Applying layered decomposition defines a new set of roles between the TSO and DSO. The TSO is responsible for system wide grid balancing operations while the DSO is responsible for managing all distribution grid assets including DER. These roles meet at the transmission/distribution (T/D) laminar coordination interface where energy and services are exchanged. (see diagram 9)
In the Total DSO model, 3rd party aggregators have an important new role in completing the coordination chain by hosting laminar coordination nodes and communicating back to the utility. In this role, aggregators aid utilities by:
- Increasing support for reliability and resilience
- Enabling DER to support system operational flexibility
By taking an active role in managing and operating the grid, the distribution utility is in the best position to maintain power quality, safety, flexibility and reliability while growing the DER infrastructure within the local service area.
Diagram 9 – TSO/DSO Coordination for DER Integration
Summary
Coordination is the key to making both distributed, decentralized and centralized systems work effectively. It has become a key issue for grid modernization due to the proliferation of non-utility distributed energy resources.
Effective coordination requires a well thought out grid coordination structure for effective grid operations. This structure should not be ad hoc but rather induced from a rigorous mathematical basis such as “layered decomposition.” Layered decomposition provides a technical foundation and results in a structure that is called “Laminar Coordination Structure.” This layered structure consists of coordination domains that communicate through coordinator nodes using interoperability interfaces based on open industry standards. Laminar frameworks can be used to generate many different layered distributed architectures and can also be used for comparative analysis of architectures.
A layered structure for transmission/distribution coordination that removes hidden coupling is the “Total DSO” model. Transitioning to the Total DSO model requires timely, consistent and robust shared coordination responsibilities between the TSO and the DSO. In order to achieve this, the industry needs to adopt, or develop where needed, appropriate coordination domain and coordinator node standards and interfaces that enable reliable and secure grid operations. This will empower utilities to implement the capabilities needed to become an active operational partner in managing the grid.
Next-Gen Grid Architecture Series:
Part I – Next-Gen Grid Architecture: Laying the foundation for a next gen grid architecture
Part II – Next-Gen Grid Architecture: A prerequisite for grid modernization
Part III – Next-Gen Grid Architecture: A peek into industry structure