Changing the dispatch equation

Using transmission intelligence + flexibility

Published In: Intelligent Utility Magazine March/April 2011


"I THINK WITH TRANSMISSION, WHEN WE LOOK AT WHAT'S going to be happening over the next 10 to 15 years, we have to go back to the basics."

With these words, Phil Harris, CEO of Tres Amigas LLC and former head of PJM Interconnection, opened a provocative discussion about the future of transmission in North America. Speaking as a panelist in a recent EnergyBiz Leadership Series webcast, "Game Changers: What you thought you knew about transmission," Harris talked about the opportunities available in a more efficient transmission grid. Here are some of his comments, edited for style and length.

In transmission, the changes that are taking place are on the security-constrained dispatch of transmission. If we look at the dispatch equation for 100 years - and this is important to understand - we've dispatched our system (this way). As you know, an electric grid, as long as it is in balance, it is reliable, as long as generation equals load.

Let's look at the parameters behind that dispatch equation. We know the generation. The ties (import or export), we know. The load is not measured in real time, but it's forecast, and our transmission grid is set with predetermined limits. These limits are often the manufacturing limits set on the various equipment.

And, as we now know, a lot of that equipment is 50-plus years old, so our transmission system is modeled in the real-time grid today based on old empirical data. Now what's happening as we move forward into the future is, with the smart grid, the load is being more and more measured in real time.

As a matter of fact, the Department of Energy had a study that said presently less than 20 percent of the information from a generator to the use of electricity is under the use of power electronics - i.e., where we know everything about it from the source to where it's actually used - and in 20 years, 80 percent of that electricity that is generated, to the point where it's being used, will all be controlled under power electronics. That is a massive game-changer, and this is where smart grid is going, and this is why elephant companies such as Cisco and IBM, GE's ecomagination and so forth, are all involved.

So the future's leading us into the actual dispatch, which keeps the grid reliable, where we know all the generators, we know the ties, and we're leading to the future in which we know, real-time, what's happening on the load. And now we have to look at what's happening with transmission. It's the last unknown in being able to optimize the grid dispatch for the future.

To look at that, we have to look at the security-constrained economic dispatch. If we look at the way that works, we have to account for all the transmission limitations because we know that not all generation can reach load. This is the security-constrained economic dispatch.

So in that security-constrained economic dispatch, we have the known, and then the estimated, inputs into that model. The real-time generation is known, the real-time flows are known. The load forecast is changing in our world because of smart grid, and we will know all the elements that lead to that.

We have to get beyond the engineering limits with assumed conditions and understand all the known conditions that are taking place on the transmission grid. These are things like transformers, real-time status, the sag of the line, whether there's ice on the line or not, whether the lightning arresters are working - the entire status of everything on the transmission system. When we know that, and we tie that with the knowns of the use of electricity so it's not forecast, then we can have a perfect and optimum dispatch of the electric grid.

When that happens, we have the capability of making huge potential improvements in transmission. If you look at just the congestion cost alone - and a lot of that is driven because we're using estimated numbers on what the status of the transmission grid is - congestion costs alone, for example, in PJM were published to be $1.3 billion last year and over $12 billion since 1998. Now a large part of that, of course, is physical constraints, but a lot of it is due to modeling because we simply don't know the status of the grid.

So putting the smarts in the transmission system, so that the transmission system is a known in the security-constrained economic dispatch equation, along with the elements that are taking place in the smart grid, that allows us to know real-time what the use is, and will enable us to get a huge efficiency in this machine called electricity, and the ability to be able to move it forward.

The interesting thing about this is why aren't we doing more of it right now? You're really talking about trillions of bits of information that are out there from all the different elements that are taking place on the transmission grid, and to (be able to) do things with that data, transform it into real-time, and make it available for information to be used for the dispatcher and others during the buying, selling or trading of electricity. This gets in everything from the demand side all the way to the generation side. And again, this is why these elephant companies - the Ciscos, the IBMs, the GEs, the Microsofts - are still so interested in electricity today.

I think this is one of the biggest challenges that we have in operating the electric grid of the future: the dispatchers are going to be faced with a large volume, a hundredfold increase over what they're seeing today. I also think that this is something that NERC should really take quite seriously in its strategy for the future: how are the dispatchers going to deal with all this data and information that is coming in from all these real-time measurements on the demand side, the use of equipment, and on the transmission side, and transform that into capabilities?

This is certainly doable.

If, when we look at the aerospace industry, and look at the vast amount of information that is needed to fly in space today, and maintain these relationships, there are a lot of models for how you can do that so you can fly this ship and maintain it. And, as you know, in the electric grid, we fly an airplane that can't ever land. So as we make these changes and this data comes in, understanding how that works in the dispatch equation, how that transfers to the appropriate signals for the dispatchers, and how that transfers into public information for trading and use is a big challenge.

Related Topics


The Solution to Overwhelming Complexity is a Paradigm Shift

We continue to look at our mounting problems through the lens that we have used since the grid was created. "We need to bring more data into the equation and find a way for dispatchers to manage it," the gist of this article, is based in a top-down, center-based control paradigm. We need much more data, that is certainly true, but more data will overwhelm our current system. Consider how the internet manages complexity and massive amounts of data. The communication protocols of the internet (TCP/IP, etc) enable massive amounts of data to self-route. They provide an appropriate network architecture to manage complexity and in so doing, unleash tremendous potential for innovation, and importantly for the grid, this approach would enable exception-based management by dispatchers. With a shift to a bottom-up, edge-based control paradigm, potential solutions become more feasible, because now they better match the problem. The key to managing an incredibly data rich grid will be opening up to new players and new possibilities, letting go of some elements of control and changing the paradigm. Managing data as it comes into the system, out at the edge, for example, reduces the data that must be carried over the communication system. Intelligent edge devices as well are needed to substitute for better equipped dispatchers. John Cooper, Ecomergence The Advanced Smart Grid

Evolution of the Electric Grid

I agree with Dr. Harris that more grid-level data needs to be made available. I also agree with him that knowing the state of the grid is more valuable than estimates of the state of the grid (hence the term "state estimator". However, I'm not as sanguine about the prospects for "perfection" and "optimality" in a smarter grid with more actors. Dispatchers, and dispatch algorithms face limitations on computing power and the availability of information. Consumers are not going to sit quietly while a faceless individual in Valley Forge decides when their air conditioner can run, and when dinner needs to be put off for the sake of economic efficiency. If we're really serious about engaging consumers in this grid modernization effort we call the Smart Grid, then consumers have to be in charge, not the dispatchers. Rather than exercising control over the grid, dispatchers monitor the grid, signal when the grid is out of balance, and only step in when automation on the edge of the grid can't or won't keep the grid from collapsing. I think some policymakers are very serious about engaging consumers, but they have no idea what's required and what the ramifications are. Jack Ellis, Tahoe City, CA