A Conversation with Chris Clack: Why the Electric Grid Needs to Get Bigger
This article was written as part of a graduate science writing course at the University of Colorado Boulder.
Chris Clack is a mathematician who builds models to study our electrical system. In January 2016, while working as NOAA researcher, he co-authored a paper that used various models to show that the United States could reduce carbon emissions from electricity generation by 80 percent by 2030, using only existing technologies (and at slightly lower cost). Clack, who lives in Boulder, Colorado, has since founded his own company, which builds models for various clients in the energy sector, including governments, utilities and grid operators.
This is an edited version of a conversation with Clack on Oct. 25, 2016.
Harrison Dreves: Could you summarize the findings the paper you co-authored in Nature Climate Change?
Chris Clack: In really simple terms, the lowest cost electricity option for the U.S., with today's technology, is 80 percent carbon free generation. A lot of people that read the paper assume we wanted 80 percent renewables. We actually asked 'what is the lowest cost system?' and then what falls out from that is a national transmission system with lower cost electricity.
Dreves: Some have criticized your model's solution for its dependence on long-distance, high-voltage direct current (HVDC) transmission lines.
Clack: That's a criticism we get a lot. 'Well of course it will work if you build this grid.' But we never said to the model 'you have to build this.' We said, 'you can build this if it's cost-effective.' And if it built the transmission lines, it would have to tell us how the power was flowing and make sure all the constraints of transmission were met. And the model kept building the lines every time because it was the most cost-effective thing to do.
Dreves: How did you use a mathematical model help you come to these conclusions?
Clack: Because the model can compute these things way faster than the human brain can, it can see patterns that we would never be able to see.
You can put storage in and all these different technologies and you let them fight it out. And then the model will pick a solution. Then you have to go in as a human and figure out 'why did it build this transmission line? Why did it tell us to do demand response? Why did it say that we needed this extra capacity over here?'
The model is doing things that supersede our brains in terms of understanding correlation. When you've got two things, it's easy for our brain to recognize the pattern. But when you start getting thousands of these things all influencing each other, our brains can't compute. These models are designed specifically to do that.
Dreves: How much computational power did it take to run the model?
Clack: We had hardware specifically built to run it. It's a significant amount of computational power. I've been trying to build simpler models that do similar things that you could run on a PC. But this model, you'd need a server farm to run it.
Dreves: The nationwide model you ran optimized to 79 percent of power from carbon free energy. What did your model find on a smaller scale, for example the size of a municipality?
Clack: It depends how big. We went down to as small as 256 separate regions across the U.S. and found a maximum of 60 percent carbon free for the lowest cost solution but it varies depending on the location. Some regions have better resources than others. The regions with the best resource tend to have the least a load. And that's why bigger is better. If you go to the middle of Wyoming you have only five hundred thousand people, but some of the best wind resources in the country. So if you didn't connect Wyoming to somewhere like Colorado or California it wouldn't get built because you wouldn't need it.
The output of one of the model's created by Clack.
Dreves: The City of Boulder is finalizing a commitment to generate 100 percent of the city’s electricity through renewable sources, mostly local, by 2030. Having worked on this model, what would your reaction be to the feasibility of a plan like this?
Clack: This is just my opinion; it's not based on fact. I haven't worked on Boulder specifically.
From all the research I've done it means the prices are going to go up, because, at the moment, Boulder is part of Xcel, which is very large and they're bringing in a lot of wind and prices are very cheap. If you suddenly shrink yourself down to the size of just a city or even just Boulder County, you immediately restrict what you can build. And I don't see a future where Boulder's going to put wind turbines on top of the Flatirons. I just don't ever see that happening and I wouldn't want to see that happen. So there's a lot of space that you automatically can't use. So it then must hinge on solar and storage.
Dreves: In your model we see that the larger the geographical area the more renewable generation in the low cost option. If Boulder shrinks the area from which it’s getting power from the entire Xcel grid down to just Boulder County, would electricity costs likely increase in Boulder?
Clack: Yes. Unless there's some technology that breaks through in the next short period of time. If we're working just with wind and solar, then, just from the variability of weather, if you shrink the size you're going to end up needing more backup generation. That could be either storage or natural gas. Therefore you're going to drive up the cost, because you're not able to diversify your portfolio like you would otherwise.
So, for example, you could cover every property in Boulder with solar and you get something like 500 megawatts of solar energy, which is plenty of capacity for Boulder. But if it's cloudy, it's cloudy over pretty much all of Boulder. You could build wind but there are times when it's when it's not windy as well. The example I give for Boulder: it snows a lot in Boulder in the middle of winter. The panels will be covered in snow and there will also be no wind. What do you do?
Dreves: And people will be turning on their heating.
Clack: And people will want heat. If you want electric vehicles you suddenly have extra demands for that. If you're converting heating into heat pumps and water heaters into heat pumps, all these things are adding extra demand. You want more electric demand because it's more efficient, but you suddenly have to find power from somewhere. If you're not connected to a much larger grid you have to find that power either from storing it, which becomes incredibly expensive because you need so much energy, or other sources.
The problem with weather is it's not random. You normally get what you call consecutive day events. It will snow and you'll have three or four days of light or no wind and snow on the panels. So you've got this long duration where you're going to be using extra energy heating homes but you have no energy production. So if you store it in batteries can you imagine how many batteries you would need just to do that? And if you do it with thermal storage you get huge problems of efficiency and drilling huge holes underground.
The 100 percent goal is great, but in my opinion they need to frame it as: 'We need to be part of a bigger grid. We want to be 100 percent net or zero emissions, but we can't just disconnect ourselves from the rest of the world.' That just doesn't work to be cost-effective.
Dreves: So the true benefit of building a bigger grid is that you very rarely have to store the energy?
Clack: The energy is already stored in the atmosphere. We're just borrowing it. The wind that we use is just stored solar. And what's even better: the atmosphere moves it for you too.
Everyone who has lived in Colorado has seen thunderstorms roll off the mountains and go off into the plains. Just imagine following that. There's lots of energy moving along with that and it's moving the energy for you. It's annoying because it doesn't go exactly where you want it. But when you go bigger you can select sites where, on average, you're always generating something. You will sometimes need some backup, but you reduce the amount you need dramatically. I did a preliminary study for Africa and if you went to the size of the African continent you get the 93% carbon free generation.
And that last little bit, the 20 percent or seven percent that’s not carbon free, can be done with other technologies such as natural gas with carbon capture or batteries. If you make that final wedge smaller, these problems with storage get smaller. Now you're not asking for everyone's house to be full of batteries. You're asking them to have one battery pack and all aggregate together.
I'm a firm believer that if we don't do it low cost we're not going to solve the problem globally. If we have outliers like New York and Boulder and Ft. Collins--who are very rich cities--go 100 percent renewable and pay 50 percent more for electricity, then how do people in African countries, India or Brazil--countries that already have a lot of people in energy poverty--say to their people, 'now 50 percent more of you are going to be in energy poverty?'