IEEE Power & Energy Magazine - September/October 2019 - 20

The complexity of the grid goes well beyond the levels
we associate with many sophisticated systems. In fact, we
designate it as ultra-large-scale (ULS) complexity, using a
concept developed at Carnegie Mellon University. This level
of complexity arises from the grid's many already complex
structures, and these structures are interconnected and interact in complicated ways. Such systems have characteristics
that pose special challenges, including
✔ inherently conflicting diverse requirements
✔ decentralized data, control, and development
✔ continuous (or at least long time-scale) evolution and
deployment
✔ heterogeneous, inconsistent, and changing elements
✔ operation involving wide time scales
✔ operation involving wide geographic scales
✔ normal failures (failures are common and frequent,
not exceptional events).
The consequences of these ULS properties are many and
sometimes surprising. One consequence is that it is not possible to define a comprehensive set of use cases for grid modernization; therefore, many architectural methods depending
on use case sets as inputs fail to yield strong results. Instead,
we use a combination of emerging trends, systemic issues, and
legacy constraints as key inputs to developing new architectural views for the grid. To do this, grid architecture replaces
many older architectural paradigms with newer ones developed from applying mathematical methods and ULS principles, in addition to the experience acquired from dealing with
grid modernization problems. Table 1 shows several of these
paradigms shifts.
One of the most important grid architecture paradigm
shifts is the extensive focus on structure because it determines the essential limits or bounds on what a complex system can and cannot do. Most design and implementation processes explicitly or implicitly assume a structure, whereas in

the grid architecture work, we are primarily concerned with
specifying structure. There are two reasons for this.
1) When the structure is correct early on, system pieces
fit neatly into place, downstream decisions are simplified, and investments are future proofed.
2) When the structure is incorrect, integration is costly
and inefficient, investments are stranded, and benefits
realization is limited.
Grid Architecture treats the grid as an interconnected
set of structures in a paradigm called network of structures.
While the system of systems paradigm concentrates on IT
systems as components, grid architecture views the grid as
comprised of seven classes of structures:
1) electric infrastructure, i.e., the physical means of
electricity delivery
2) industry structure, including electricity market structure (not market rules)
3) regulatory structure (relationships, not regulatory rules)
4) digital superstructure (information systems and communications networks)
5) control, treated as a structure rather than an application
6) convergent networks (electric, gas, water, traffic, and so
forth), which may be found in a smart city environment
7) coordination framework, i.e., the often intangible
structure that determines how disparate elements of
the grid cooperate to deliver electricity.
These structure classes have complex relationships whose
details vary by region, industry segment, regulatory regime,
and other factors.
We have inherited a massive amount of structure from the
20th century grid and with it many structural constraints that
inhibit grid modernization. Consequently, we define the core
problem of Grid Architecture as one of determining the appropriate new structures or minimal structural changes to the grid that
✔ relieve crucial constraints on new capabilities

table 1. The Grid Architecture paradigm shifts.

Classic Architecture Paradigm

Grid Architecture Paradigm

Comments

Focus on components

Focus on structure

In simple terms, the boxes of a block diagram are
the components, and the lines are the structure.

Structure is inherited or
improvised

Structure is formalized

Structure should have underlying mathematical
or other rigorous bases.

The grid is a big circuit

Interactions between electricity
markets and grid controls

Where centralized wholesale markets exist, they
are closely related to grid controls.

System of systems

Network of structures

The grid comprises multiple complex structures
that are interconnected in complex ways.

Data tsunami

ULS complexity

ULS complexity is a consequence of the network
of structures.

System integration

Convergence and platforms

"Platform" is a very general architectural concept
that applies broadly to system structure.

Architectural elegance

Architectural quantification

Elegant (i.e., pretty) architectures are not the goal;
alignment with objectives (as determined by
rigorous quantification) is.

ieee power & energy magazine

september/october 2019

IEEE Power & Energy Magazine - September/October 2019

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - September/October 2019