In the discipline of Grid Architecture, we recognize that an architecture is composed of three types of things: structure, back box components, and externally visible characteristics. Components are black box in the sense that we are not concerned with how they work or are implemented internally; we only care what their functions are and what behavioral characteristics they have. So, for instance, a bilateral bulk energy storage device may be a lithium ion or sodium-sulfur battery, a flywheel or some other physical device. What matters at the architectural level is how much energy it can store, how fast it can charge and discharge, how may cycles it can support, etc., not what is inside of it.
When we teach Grid Architecture, we caution the students to be wary of specifying components that we refer to as magic or anti-gravity boxes. These would be components that have no actual basis in reality and no reasonable line of sight to becoming available and may in fact be contrary to actual physics. Architectures containing such anti-gravity boxes may seem good on paper, but the reliance upon something that cannot be obtained is a fatal flaw. We teach the students how to do sanity checks to prevent writing an architecture that has anti-gravity boxes embedded in it.
Unfortunately, many in the electric industry employ anti-gravity box thinking. It shows up constantly in discussions about how all we need to do is provide open access and markets/pricing signals for consumers’ DER to respond to the grid’s needs. An anti-gravity box will make everything work. It’s magic! Not only will the anti-gravity box resolves how to get all the expected tens of millions of grid-connected devices to work together, it will also extract fabulous amounts of value from all these resources and deliver this value from one set of consumers’ DER to all consumers, all the while operating the grid so as to avoid any significant need for new grid investment.
This anti-gravity box has a name: optimization. Optimization will make it all work; optimization will extract loads of latent value through utilization of consumers’ DER.
Well, here is some sanity checking. First, the proponents of the optimization magic box do not spell out how such a thing can exist. Those who understand optimization mathematics and methods know that optimization involves a set of feasible solutions (those that satisfy all the constraints of the problem at hand) and that among the feasible solutions there may be one that maximizes or minimizes some objective and is called the optimal solution. However:
- In a heavily constrained problem, it may not be possible to find even a feasible solution, let alone an optimal one; with too many constraints no feasible solution may exist and even if it does, it may be difficult or impossible to find.
- An optimal solution, if it exists and can be found, may well be brittle, meaning that a slight change in underlying conditions can make the formerly optimal solution invalid. An example of this was when earlier in 2022, flooding of coal mines and railroads in Australia caused a shift in the large signal component of grid supply, making the market optimization fail; hence AEMO shut down the electricity market for a period of time and went to direct dispatch of generators.
- The presumed latent value to be unlocked by optimization does not necessarily exist, and even if it does, may not be accessible via grid control optimization due to human behavioral factors that are not accounted for in the formulation of the optimization problem (a common shortcoming of attempts to use consumers’ DER to solve large scale grid management issues).
Relying upon magic anti-gravity boxes is poor thinking if we are to address the growing climate mitigation and adaptation needs in the electric system.