The future of the power grid: The challenge of DERs

REBECCA AHRENS

When you turn on your TV, plug your phone in to charge, or pull the chain on the light in your closet, somewhere out there the amount of power being put onto the grid actually increases just a teeny tiny, incy-wincy little bit.

Every four seconds, grid operators send out a signal to power plants to make sure that the total amount of power being fed onto the grid matches the amount of power consumers are drawing from it. 1 Grid operators have to do this supply and demand comparison so frequently because the power grid exists in a fragile state of equilibrium. If supply and demand fall out of balance, the entire system can come crashing down, causing cascading system failures and massive blackouts that can take hours or days to reverse. 

Natural gas turbines can be ramped up in a matter of minutes to meet fluctuations in demand. 2 But what happens in the future when renewable sources of energy penetrate deeper and deeper into our energy mix? Imagine: in the middle of this perilous, second-to-second balancing act, clouds suddenly pass over the region’s largest solar farm at the wrong moment.

Today on the show: how renewable sources can quickly tip the scales on this fragile balancing act, and what technology and data are doing to help.

From AVEVA studios this is Our Industrial Life, the podcast that brings you stories from the essential industries and investigates how data and technology are shaping the future of the connected industrial economy.

What are DERs?

 

REBECCA AHRENS

I’m your host Rebecca Ahrens, and welcome to part three of our future of the power grid series. In part one we covered some of the main challenges we face as we transition to a decarbonized power grid. Today we’ll dig a little deeper into a particularly fiddly challenge—one that causes grid operators all over the world a lot of headache. We’re talking about: DERs.

Pat, can you tell us what “DER” stands for?

PATRICK KENNEDY

Certainly: distributed energy resources. The word “distributed” means that they’re outside of the viewing of the utility. That means they’re behind the meter.

REBECCA

Like we did in our last show, I just want to mention here that Pat passed away recently. So I want to again take a moment to acknowledge his immense contribution to industrial software. Over 40 years ago, Pat and his team launched a plant information software program known as AVEVA PI System. today, the PI System is still an essential tool for over 1200 power utilities and transmission and distribution grid operators. 3 Like the last episode, this episode contains clips from our interview from a couple years ago.

So Pat, what is a distributed energy resource?

PAT

The most common example today is, of course, rooftop solar. As they are installed more and more, there’s more generation coming from the actual end point rather than from across the wires.

But there are others that actually have a more dramatic effect—battery storage, electrical vehicles—and these are good examples of distributed energy resources.

REBECCA

Individually, they might be of a far smaller scale than the big, centralized plants we’re familiar with today, but DERs will very likely prove to be an essential tool in achieving the energy transition. Experts who study the power grid expect annual investment in DERs to rise by 75% by 2030. And they expect DERs will account for about 10% of the world’s global installed power generation capacity. 4 But while networks of distributed power assets will likely play a critical role in facilitating our energy transition, they also cause some major complications.

PAT

There’s two issues here. The distributed energy resource part of it is, in general terms, considered to be smaller resources, and behind the meter. But the other factor is that when users are putting their distributed resources behind the meter, then a utility trying to predict what power they need tomorrow becomes less certain because they don't really have any data—the data are not telemetered in to the utility from the other side of the meter.

But the other issue is that they are intermittent. And even the commercial solar farms and the commercial wind farms are intermittent.

REBECCA

By intermittent, Pat means that the energy source cannot reliably be turned on and off to meet the electricity needs of the moment. In other words, intermittent energy sources, like solar and wind, are what we’d call “non-dispatchable.” We don’t have any say in when the sun shines, or when the wind blows. Sun and wind can’t be switched on on demand the way that dispatchable energy sources, like coal or gas-powered plants, can be.

As you might imagine, intermittent, non-dispatchable energy sources pose a big challenge to the grid’s management and stability. The intermittent and non-dispatchable nature of these energy sources is the root of why they’re such a big problem for grid management and stability.

So, before we get in to DERs and what they mean as far as grid stability is concerned, can you tell us what we really mean when we talk about managing the grid and keeping it stable?

JOSHUA RHODES

Yeah, no, that's a great question.

REBECCA

That’s Joshua Rhodes. If you listened to the first two episodes in our series, you already know Joshua. But just as a reminder, Joshua is a research scientist at the Webber Energy Group at the University of Texas at Austin where he focuses on issues related to decarbonizing the power grid. 5

Keeping the grid in balance

 

JOSHUA

So what we're really talking about is the system operating in harmony or in balance. Because electricity is different from other commodities or other products in that for every bit of electricity that's being consumed, it is having to be generated at that exact same moment. We have to keep that supply and demand in balance, you know—and on a second-by-second level.

And so, grid stability—in kind of the sense in which we're talking about it—is really kind of keeping the frequency of the system at 60 hertz, which is what it's really designed to stay at. And so if supply or demand get out of—too far out of balance, the frequency can drop or the frequency can go up. And that can wreak havoc on the system. And if it goes far enough, and the grid gets unstable, then the whole thing can actually come crashing to the ground or, you know, end up in a blackout.

REBECCA

And I guess we should clarify, that when we talk about the grid here in the US, we don’t actually mean a singular grid, right?

JOSHUA

You can really think of the grid as a huge machine. In the US, we've got three grids: the eastern interconnect, the western interconnect, and Texas. But these are massive machines that span over, you know, hundreds of 1000s of square miles. And we have to keep them in balance.

REBECCA

So at first glance, keeping the grid in balance might look like it’s just a matter of making sure we have enough power to keep up with demand. But of course, it’s a little more complicated than that. Can you say more about maintaining the frequency of the system?

JOSHUA

Yeah, so at the same time that the grid is operating at a 60 hertz frequency, most power plants—the traditional types of power plants, be they hydro or coal or nuclear or natural gas—they're actually what we call synchronously connected to the system.

REBECCA

Synchronously connected to the system. OK what does that mean?

JOSHUA

So the actual shafts inside of these power plants are spinning at 60 hertz, they're spinning at the frequency of the grid, because they're generating that electricity at 60 hertz. And these massive steel shafts have mass. They're heavy. They're large, they're made of metal, they have a lot of mass.

And so they have some amount of, kind of, storage of frequency in them such that if the frequency of the grid starts to fall, that spinning mass will actually push back against the decline in frequency. And so that's this rotational inertia—that’s the actual, like, physical spinning mass on the grid that would try to push back against a fall in inertia. We essentially try to keep enough of this inertia on the grid, such that if we do lose a power plant, or we have a massive spike in demand, that we're able to overcome that temporary mismatch in supply and demand so that we can bring on other resources to otherwise, you know, balance the grid.

If a particular power plant goes down and it disconnects from the grid, it’s not going to be able to provide any service to the grid. It’s all the spinning mass in the other power plants that are still connected to the grid that are able to kind of make up for that lost power plant for a little bit of time before we can bring on other resources or dial up power plants that are already online, or dispatch batteries, or whatever it is that we’re looking to do to inject power. It’s something that just kind of buys that time to keep us going.

REBECCA

So you kind of hinted at this already when you were talking about these rotors, but just sort of break it down. Why are renewables—in particular, solar and wind—considered a problem for grid stability?

JOSHUA

They’re a different type of power plant because they don't have that spinning shaft that is synchronously connected to the system. Now, wind does have spinning parts. I mean, you look at a wind turbine, and you can see it spinning.

REBECCA

But I guess it’s not spinning at 60 hertz like the shafts in traditional power plants. The wind turbine will spin faster or slower depending on how windy it is.  So, then, does something happen to the electricity after it’s created by the wind turbine to bring it in line with 60 hertz?

JOSHUA

that electricity is usually put through an inverter. And so the inverter is what brings it up to 60 hertz, and then that’s what pushes it out to the rest of the grid. And usually, that inverter relies on the frequency of the grid to be able to transition that power from direct current—which is what’s being generated by the wind turbine—into alternating current to be sent out to the grid.

REBECCA

I see. And, of course, there’s nothing spinning on a solar panel either—much less at 60 hertz.

JOSHUA

Solar doesn't have any moving parts. And so its direct current generation is also being formed by those inverters before it's being pushed to the grid. So if the grid frequency does start to decline, traditionally, the way we have operated, these resources would not have that inertia to push back against a decline in grid frequency.

REBECCA

OK, so this issue with grid frequency and rotational inertia means that creating a modern, decarbonized power grid is definitely not going to be as simple as just installing more DERs and renewable plants.

Keeping the grid efficient

 

PAT KENNEDY

In many cases, we have plenty of power. It’s just not in the right place. It's not at the right level, it is not available, it can't be dispatched—there's all these kinds of issues to go with it.

REBECCA

This is Pat Kennedy again.

PAT

And that's very important for a grid operator because it means they're noise—they're not really useful. When you look at managing the grid, you have to ask yourself, how do you do that? A lot of the resources on a grid are very, very slow-reacting stable resources, whether they be a nuke, or combined cycle power plant or whatever.

REBECCA

Just for any listeners who might be confused, Pat is talking about nuclear power plants here, not nuclear weapons.

PAT

If you now have to compensate—let me say for clouds moving over a photovoltaic farm—now you have to have very fast-moving resources.

REBECCA

And how fast do different resources move? In other words, how quickly can different types of power generating resources get electricity onto the grid at the right frequency?

PAT

In the power generation business, your resources have different abilities to be dispatched. And their efficiency changes.

REBECCA

What do you mean it changes?

PAT

For example, a small—what they call a “peaker unit”—these would be generally gas-fired turbines. Sometimes their efficiencies are barely above twenty: they're about the same as an internal combustion engine. If you look at a supercritical power plant base, it may still be a single turbine. But now your efficiencies are up in the mid-thirties and it takes much longer to line out.

Then you move up from that, and you get into what's called a combined cycle plant. And that means you generate power from gas generally. And then you take the superheated steam that comes off and you generate more power from the steam. These combined cycle plants—it takes hours for them to line out, but they're very efficient. They can be above 60%. And in fact, the most efficient in the world is called the J series and it's 63%.6 So in that small change, going from the small peaker turbines to the combined cycle plant, you went from 20% efficient up to 63% efficient.

REBECCA

Right, that’s a substantial difference.

PAT

Now consider that if you put a lot of intermittent resources on the grid, you have to have faster-acting turbines and faster-acting resources. And that's how you can lower the efficiency of the grid.

REBECCA

Right, ok, because you have to use these peaker plants to compensate for how intermittent renewable power sources are. The peaker plants are quick enough to respond to sudden changes, but they’re super inefficient. Which, you’re saying, makes the whole system less efficient.

PAT

And that's one thing we have to be careful of, because as we move to higher and higher percentages of intermittent resources, the grid gets more difficult to control.

And so when we mandate for example, 50% renewable power, then what's happening is we're taking the fraction of intermittent power, and we're raising it to a level we haven't seen before. And that's where we'll see the effect of everything from storms, earthquakes, anything that affects these, weather-wise, will cause disturbances in the grid.

REBECCA

So the grid is more susceptible to disturbances. How big of a disturbance, potentially, are we talking about here?

PAT

Ask me about the Northeast blackout in the U.S.

REBECCA

Sure. What’s the Northeast blackout?7

PAT

The Northeast blackout was a phenomenon that happened about close to 20 years ago. But what happened is that the power generation was at a very high rate, because it was hot and humid. And we had just deregulated, and so there were a lot of companies that were driving their lines fairly heavily, because the price was high.

REBECCA

In the 1990s, many U.S. States began deregulating their electricity systems to introduce more competition with the aim of lowering prices. It’s a transition that we call “restructuring” today.8 Prior to restructuring, most customers in the US bought electricity from big, monopolistic utilities. Deregulation led to the advent of independent energy producers, each of which owns its own power generation assets.

PAT

And so when the price is high, you want to deliver more at the high price. They had an incident with a transmission line, tree vegetation, line broke. The actual root cause was at FirstEnergy in Cleveland.

REBECCA

An alarm was supposed to alert operators in the control room to the broken line, but the alarm system had failed. Because they didn’t have access to reliable data, operators spent an hour and a half just trying to understand what was happening. And during that time, three other lines failed.

PAT

And then that resulted in then propagation of blackouts that started across the country. And it actually had reversed flows out of Canada, and it brought down a couple of nukes, and it continued over to the east coast. And MISO, which is the ISO that manages the Midwest—had just started up—was not able to contain it. And the blackout went clear over to New Jersey, where PJM, actually, was the ones that stopped it. But there were millions of people that were blacked out at that time.

REBECCA

Wow. And so how long before they could get power running again?

PAT

The actual repair lasts for some time. I mean, they’re able to bring power in, and there’s emergency measures, and you can generally get some power up if the lines aren't damaged. But if the lines are damaged, and you have to repair lines, you can imagine that you're trying to bring up a transmission line. There's a lot of people that are hanging on that line, and you have people with radios, making sure it's safe before you energize and you energize a step at a time. And it takes a long time to get full service restored. But that's an interesting study, because of the cross-border nature between Canada to the US.

REBECCA

And I think that's a really good illustration of how expansive the interconnection really is, right? Like, it's not just localized necessarily to one city. If the power goes down in one city, maybe not everybody would assume that that could ripple over state lines and even country borders.

Meeting the challenge with data

So, where does this leave us? How do we organize and integrate all these small-scale, intermittent energy sources into the grid in such a way that we aren’t exposing ourselves to these big risks?

DAVID BARTOLO

Data, data, data.

REBECCA

This is David Bartolo again. You might also recognize him from parts 1 and 2, but just as a reminder, David is the Head of Asset Intelligence at AGL Energy in Australia.9

DAVID

Yeah, the biggest hurdles to make that transition to decentralized, renewable variable generation assets, from the centralized fossil fuel, highly dispatchable and controllable assets, is you've got vast distributed small assets, representing huge amounts of data that need to be collated and centralized, to give you a virtual power plant understanding of it as a whole.

REBECCA

Sorry—sorry to interrupt, but before you say more, can you just explain for those who might not know what a virtual power plant is, exactly?

DAVID

We call it a “VPP.” And our vision of virtual power plants is, at the residential level, to have many customers who have solar systems on their roofs, combined with battery storage technology, and to be able to aggregate all that data, understand how much energy we could supply to the network at any time, and also how much energy we could take from the network at any time to charge those batteries. And dispatch them as what we call a “virtual power plant”—a gigantic battery, to support the network, either at times when there's not enough energy discharging those batteries, or at times when there's too much energy, especially during solar max, to charge the batteries for times of shortage, especially during peak arrival times when everyone gets home from work.

REBECCA

So just to make sure I'm understanding—the vision is that the virtual power plant would not simply encompass, like, the behind the meter assets, or the things that are just on the residential houses, or just you know, your own, say solar farms, or wind farms or hydro assets, it would it would encompass all of those things. All of those things would be monitored in real time.

DAVID

Absolutely.

REBECCA

Is this kind of an effort to minimize—or compensate for, I guess—the fluctuations that you get with renewable energy that you don't tend to get as much with traditional energy sources?

DAVID

Yes. With renewables comes the problem of variability, which can’t be controlled. The sun shines only during the day, and when it wants to. And the level of clouds is variable. And wind is exceptionally variable.

And also, with those forms of generation, the average capacity factor is only from 25 to 35 percent of what you can get out from megawatt-installed of renewable compared to the megawatt-installed of fossil fuel. So you need to install three to four times as much to make up the same energy.

So at times, you're going to be generating four times the amount that you need, and at times, you won't have any energy due to no wind, no solar.

REBECCA

So how do we deal with that? Battery storage?

DAVID

So storage becomes a massive challenge, which is required to make this transition away from carbon-based fossil fuel to carbon free renewables.

REBECCA

If you’ve listened to the previous two episodes in this series, you’ve heard how important it is that we expand our transmission system. Expanding transmission networks will allow us to move energy around more easily to where it’s needed most. But the other big part of the equation, as David just hinted at, will be massively scaling up grid-scale energy storage around the world. Because if we can store energy on a larger scale, then we can use it to support the hour-to-hour variability of solar and wind.

And while investment in grid-scale battery storage2 is rising quickly around the world, we’ll still have to pick up the pace dramatically if we’re going to reach net-zero any time soon. Experts think we’ll need to add an average of more than 80 gigawatts of storage per year between 2022 and 2030.3

But let’s get back to our discussion.

So, you're collecting, you know, data and information about these installations and about the battery capacity—do you share that with the entity that oversees the grid? Or how does that relationship or data sharing work?

DAVID

Yeah, we absolutely do have to do that. So, the network in Australia is controlled by a company called AEMO, Australian Energy Market Operator, A-E-M-O, and we must dispatch all assets, renewable, storage or conventional fossil into that market, providing real-time data on our capacities, our capabilities, and also bidding into a market with certain complicated services related to those assets. So that data linkage has to be high resolution and real-time. And also, we are inventing now, to gather data from thousands of sources in the field rather than large, centralized assets, to understand what capacity we are able to bid into that market.

REBECCA

So if you need to bid into that market, I imagine you want to forecast how much power capacity you’ll be able to supply. That must be difficult to do with intermittent energy sources like solar and wind.

DAVID

So forecasting on a big wind farm—we’re already produced very complex artificial intelligence-type algorithms to better forecast what we can do in the next five minutes, so we very accurately bid into the market. And there's huge financial advantage in being able to do that.

We need to be able to do that—from, you know, one to seven or eight wind farms and a few solar farms—we need to be able to do that with thousands of household-level, and many hundreds of commercial-level installations to forecast what we can do in the next five minutes as well. So that's another part of the challenge.

I think that describes the challenge and the formula to getting it right. If you're going to use time-series data, use technology that is designed for time-series data—not designed for a SQL server.

REBECCA

Got it. And then apart from the meter data, I assume you're also collecting weather and status of the asset, for example—that sort of data as well?

DAVID

Yes. There’s two vectors there. There's all the external data: satellite data, irradiance data, and it could be weather data. We've already brought in all our bureau of meteorology data into our system so you can do comparisons to that.

But then you go further at the hardware on the site—in the house, or in the factory. There's an inverter, which converts and controls the battery. And the controller for the solar plant has a richness of data within it that really helps diagnostics and helps understand performance and capacity equations as well.

So there's a interfacing with that hardware, maybe through an OEM’s API through their website—you can gather that data that way—quite efficient, because you don't have to install anything. But maybe ultimately, there may be an interface that AGL licenses that plugs in to your site hardware and does it directly. We're still inventing that as we go.

Centralized data and IoT

 

REBECCA

And how important is it that all of this time-series, real-time data, as you're describing, be in one system? Like, why couldn't you just have it, you know—your meteorological data comes from one place, and it's stored in one place and your asset data comes from one place, and it's stored in one place? Why does it all need to be collected in one single source of truth—to use a buzzword that everybody likes to use?

DAVID

That's actually a brilliant question. And we're seeing—there's two schools of thought in the world today: where you try to analyze in separate domains and leave the data where it is, or you centralize. We're finding with time-series data, we're able to do it so efficiently and integrate it so easily into the one source of truth. Correlating data of the same timestamp to what's going on across a portfolio is extremely valuable.

So where we’re bringing in also, another source of data, is internet-of-things sensors on a large site. Being able to correlate that with everything that's going on on the site in the one time-series historian is paying massive dividends for us. And once it is, because you can do it very efficiently, very easily, and without heaps of storage—because not a big data approach—it’s a very time-series-efficient approach.

Then to do the analysis is really easy. We can export blocks of our data to the universities of the world to play with to see if they can come up with improved algorithms. They can tap in through cloud services to our stream of data in real time. It's all in the one place. And then they can try to come up with improved algorithms at an experimental and at an R&D level. But also we can do that with service providers or with our own team. So once the time-series data is in the one place, everything's correlated, the analysis is much, much easier. So that's why we're doing that.

REBECCA

Got it. So just to rephrase, it sounds like you're saying—there are kind of two things I’m hearing. One, it's beneficial, because it just makes that correlation of these different data sources much easier for you. But there's also this data sharing element that is much more easily accomplished when you have it all in one central location, already tied together, maybe, you know, organized in such a way so that it's very easy for someone to tell like, oh, this is referring to this particular wind turbine on this farm. I know what this data means.

DAVID

Absolutely. And in verifying certain data incidents. So if we’re tunneling in to a big incident that happened on the network, we can verify what happened on one site with the data we're seeing on the other sites really quickly, really easily. And also, if you want to share that data—as you describe—securely, there's only one security overlay you need to place, rather than doing it on seven vectors of data—and it just gets harder.

So we can then run diagnostics, and be able to maintain those assets in the field with other services, helping our customers have the best possible solutions on their rooftops and with their batteries. So at the moment, we do that with centralized assets. We monitor their health, we identify anomalies, we're able to maintain and repair before ultimate failure. We want to be able to do that in the future with distributed assets, which is a huge data challenge.

And if the history of how we matured our data capabilities on the utility-size assets is anything to use as a guide, if we didn’t make the hard decision—one data infrastructure, this is the way we do it—earlier on, we wouldn’t have been able to achieve any of things I’ve described to you. Like, in 2012, we made a big decision: the real-time data infrastructure will be one technology for all sites. And then we were able to mature and mature and mature our systems—in ways we didn’t even dream of when we made the decision. We did not know about advanced pattern recognition diagnostics. We did not know about internet of things. But we made the big, bold decision about what our data infrastructure would be, we stuck with it, and then we didn’t have to revisit that again. And we could just mature, look for opportunity, innovate, and build opportunity and value again and again and again.

REBECCA

So you were saying that you didn’t necessarily know that IoT and pattern recognition—that these things were coming down the pipeline. But you set yourselves up—it was almost kind of prescient—you set yourselves up to be successful with those because you picked the thing that was going to allow you to take advantage of those innovations.

DAVID

Yeah, absolutely. I think that’s what we’re facing in Australia now with the energy transition: the next ten years will represent the biggest change in this industry since the 1890s, when we first electrified. And if we’re not set up correctly with a data system and revisiting how you solve the data problem, which we’ve explored in detail, I think you’re going to find yourself at square one again and again.

REBECCA

Well, I really appreciate you taking the time to chat with me. I mean, this is one of my favorite subjects, just because, you know, as we were just discussing, it is so—it can be so powerful, and empowering, in both the literal and the metaphorical sense of that term, when it works.

And there's so much at stake when it doesn't work. You know, it just seems like one of the most, kind of, important issues that you could choose to think about. So, I really appreciate you taking the time to, kind of, walk us through this and walk us through what AGL is doing.

DAVID

Thank you, Rebecca, the questions and the discussion is—you can discuss this for days. It's exciting. It's frightening. It's the reason we're getting up and going to work in the morning at the moment. Because the market is getting more difficult to do business in, the commodity is life-and-death, and it has to be there. And we’re trying to transition the whole thing.

We are in a revolution. All the questions you've asked are so fundamental, and it's frightening that we don't have all the answers. But data and technology will play a part like never before.12 And making the right decisions now is going to make a huge difference in the next 2, 3, 5, 10, 12 years. Yeah.

REBECCA

Hopefully the next time you see a solar panel on a neighbor’s roof, or an EV charging port in someone’s garage it will be a reminder that little by little, one of the most important technological revolutions of our lifetimes is gaining speed.

As David said, it’s exciting. It can be frightening. There’s plenty of obstacles ahead, but maybe just beyond them, is a sustainable, energy-secure future.

That’s our show for today, folks. And this also concludes our Future of the Power Grid series. Thanks, again, everyone, for tuning in, and a special thanks to our guests, Joshua Rhodes, David Bartolo, and again a special acknowledgement to the late Dr. Pat Kennedy and his family.

End transcript