Millen: Hello everyone, and welcome to this conversation with Natalie Mims Frick and Tom Arnold. For the third time ever, and second episode in a row, we’re joined today by two people.
Natalie is an Energy Efficiency Program Manager in the Electricity Markets and Policy Group at Lawrence Berkeley National Laboratory, while Tom is CEO of the tech company, Gridium. I’ve invited Tom on given his over-10 years of experience applying peak demand and energy efficiency management strategies in buildings. This should be fun!
My name is Millen, and I work for Tom at Gridium. Buildings use our technology to develop energy project revenues and streamline operations, boosting building value and sustainability.
Today, Tom and I will be discussing with Natalie some of her team’s latest research, summarized in the new report “Peak Demand Impacts From Electricity Efficiency Programs.”
Natalie, we spend a good bit of time thinking about peak demand, but before jumping in there, I want to ask about your journey into the electricity grid, which included some time at the Rocky Mountain Institute. What was your path to LBNL?
Natalie: Sure. First I want to thank you for having me on the show and also thank the Department of Energy for sponsoring our research.
And to answer your question, I lived and worked all over the country before I landed at LBNL. And as you mentioned, I was at RMI actually working in Hawaii and then in Colorado after I completed my masters in Vermont and I headed to Tennessee after that and worked for a local non-profit called the Southern Alliance for Clean Energy. And then went back out west to start my own consulting and I was an expert witness in energy proficiency proceedings for a while. And I met my current supervisor, Lisa Schwartz, at that time and I must have done something right because she recruited me to work at the lab after that.
Millen: You were on the Electricity Markets and Policy Group at the lab, can you tell us a little bit about this team?
Natalie: The Electricity Markets and Policy Group, or EMP, has about thirty people in it and we all work on a variety of energy issues.
We seek to make an impact through rigorous analysis of policy, economic and technical issues that support a successful transition to a clean, efficient, reliable and affordable electricity sector. And we usually group our research into seven different categories: demand response and SMART grid; electric system planning; electricity reliability and resilience; efficiency; renewable energy; technical assistance to states; and utility regulation and business models.
So, we cover a lot of ground.
Tom: That’s impressive. This is Tom. Love the paper and am very excited to see someone tackle something so comprehensively.
One thing I’m excited to ask you is how do different people define peak demand?
It seems like in one area of the country, everyone has this shared understanding of what peak demand is, but it’s really different all over the country.
Natalie: Yeah. There’s lots of different definitions as you’ll see if you take a look at the paper. It seems like in regions of the country that have—for example, ISO in New England, they have a forward capacity market that allows energy efficiency to participate and so they have a much more standardized definition that the Energy Efficiency Program Administrators kind of stick to; whereas, in other parts of the country, the definitions very significantly… it also depends on what the purpose of talking about peak demand is.
One of the simpler definitions that I saw when I was doing this research was from the Energy Information Administration and they define peak demand as “maximum load during a specified period of time”, which is probably the most basic definition that you can get to.
The challenge ultimately comes in understanding where the maximum load is—whether it’s on the transmission system, the distribution system or if it’s off of the customer’s meter. And then what the specified period of time is: you know, is it one hour in the year? Is it one hundred hours in the year? Is it seasonal? So, even a simple definition becomes pretty complex really quickly.
Millen: In the past, costs and benefits of efficiency programs have centered on ROIs from annual energy reductions—that’s changing?
Natalie: Yeah, I think though cost-benefit analysis has always considered the value of capacity savings from efficiency—from a planning perspective; but I think that the policy perspective is really starting to change.
And historically, energy efficiency goals for regulated utilities have been focused on energy savings—you know, like 1% of retail sales a year, or something along those lines.
But I think there’s a shift in the focus of how efficiency and other distributed energy resources can be used to effectively integrate renewables onto the grid and address challenges of electrification and keep the electric system affordable and resilient. So, I think there is a shift that as program administrators are thinking about energy efficiency, you really have to consider all of the characteristics of it, not just the energy savings.
Tom: Interesting. And can you say more—your research touches a little bit of this, but—can you say more about what the potential linkages are between a given Grid’s peak demand or system peak and costs might be? What are the relationships there that your research is trying to tease out?
Natalie: With regard to peak demand and cost, the electricity system is being built to serve that last kilowatt hour of energy; and subsequently, peak demand translates directly to the capital cost requirement of installed generation capacity that’s being required.
So, if you’re able to reduce that demand then you can reduce the amount that needs to be on the system and also, you can use those other generating resources for serving other purposes like reserve margins or other grid requirements, and that should bring down cost—and often does bring down costs.
And so any time you can bring the peak demand down—either through efficiency or demand response or other DERs, you’re most likely going to bring down your total system cost as well.
Tom: And just to be clear, when you’re studying the energy efficiency, you’re talking about—I think in the paper you say “non-dispatchable”. That’s different from demand response, in terms of what you’ve come across.
Natalie: Absolutely.
And this is something that I think is confusing to a lot of people who have looked at this paper because any time you start talking about peak demand reduction, people generally kind of get into this demand response mindset; but all of the research for this paper is around demand reduction from energy efficiency, and so it seems like it was hard for some people to wrap their head around that–that efficiency really can reduce peak demand.
That’s been interesting to see and try and help people understand that a little bit better.
Tom: And is one of the differences that energy efficiency—of course, except for the weather variability—is more or less permanent? Versus demand response is “we’ve got to work” it to dispatch it.
Natalie: Yeah, I mean once your efficiency is installed, it’s there for the rest of its lifetime, right? So, you get those reductions forever and—not forever—for as long as the measure is viable and as long as it’s installed correctly, then you’re guaranteed those reductions to occur. Whereas, with dispatchful resources you have to have some kind of participation and response.
Tom: So, what does it add up to? I mean, if you did a ton of research obviously. I’m really impressed in the scope of what you guys studied and it must’ve been countless nights digging through filings and reports. But at the end of the day, how does energy efficiency stack up in terms of its effects on peak demand?
Natalie: This was the beginning of our research on this, so we didn’t have a really strong statement at the end or a table comparing the overnight costs from EIA or Lazard to what energy efficiency peak demand reductions are. But generally speaking, we found that it’s a low-cost way for utilities to meet peak demand.
And as we continue this research, we hope that we will be able to make stronger statements about how energy efficiency compares to other resources in terms of capital costs and what you have to put in to get out that kind of kW reduction.
Millen: To measure this, your team built off of the program administrative cost of saving electricity. So, first, what is that? And secondly, what is the new metric you created?
Natalie: Yeah, this work builds on some of LBNL’s foundational research on the cost of saving electricity.
The program administrator—which might be a utility or a third-party implementor–the program administrator cost of saved energy is expressed in dollars per kW/hr of electricity savings, and the costs include all costs incurred by program administrators, but not participant costs.
So, for example, it includes administration, incentives paid to customers, evaluation, measurement and verification, marketing, things like that. And then the benefits or the kWh savings used are the lifetime savings for that program.
And we built on that metric for this study and we calculated the cost of saving peak demand, which is also from a program administrator perspective. And the metric’s a little bit different: it’s expressed in dollars per kW saved, and the costs are pretty much the same and the benefits are just the kW reduction.
But we didn’t look at the lifetime savings, we only looked at the first year savings.
Both of these metrics are useful for comparing relative costs of types of efficiency programs and comparing efficiency options to other supply and demand choices. Probably the most important thing to recognize with this study with the cost of saving peak demand is that we didn’t divide the costs between kWh and kW benefits; so it’s one program cost and it has kWh energy savings and kW demand savings. And we just took that cost and applied it to both the kWh savings and the kW savings.
So you can look at either one of those numbers for a program and realize that you’re getting a lot of other benefits that aren’t getting recognized in there. So for example, if you look at the dollars per kW, it means that all of the energy savings that are occurring to that program are basically happening for free because you’re wrapping the full cost of the program just to the kW savings.
And we did that because we weren’t certain of the best way to divide the cost between the kWh and the kW savings, and so we just took the total cost and applied it. So that’s a pretty important thing to know when you start looking at these numbers.
Tom: That’s very interesting. Can you talk a little bit about the geographic spread here? There was quite a diversity—and I guess this follows theory on how demand reductions are mapped to climate zones. Is it easier in a dry climate? A humid climate? A warm climate? Etcetera, etcetera. And that might be interesting as well, in terms of how these grids are changing across those areas as well.
Natalie: Yeah, we decided to get some good geographic representation and also get representation from different ISOs and RTOs as well. And so we had nine states that we looked at for this particular study: it was Arizona, Arkansas, California, Colorado, Texas, Maryland, New York, Massachusetts, and Illinois—I think that was nine. And so, kind of all over the map and we did also try and get different climate zones when we were picking these states and the data that we would gather.
Unfortunately, none of this work is forward-looking, so there’s no forecast involved with it. The data that we gather is all from evaluation, measurement and verification reports or other kinds of regulatory filings that confirm the energy savings and program costs and demand savings that the utilities are reporting to their regulators, so it doesn’t take into account how things might look in the future; but one of the sensitivities that we did look at was climate zones and we wanted to test this hypothesis that program with dependent measures would have a lower cost of saving peak demand in more extreme climates.
And we found that the cost of saving peak demand does tend to be a little bit lower in states that are hot and humid, like Texas, or that are hot and dry, like Arizona; but we also found that in Illinois, for example, the utilities had a pretty low cost of saving peak demand, and they’re in a cool and humid climate. So, we didn’t have any very strong conclusive results about the climate zone: it seems like the program cost was really what was driving the cost of saving peak demand, but this is definitely an area for future research and we’re continuing to work on this project and build it out, and gather more data from more states.
We’re hoping, with a more robust sample size, we’ll be able to dig a little bit deeper into what the factors are that might be driving the cost of saving peak demand.
Millen: And what did you find from program to program about which ones are most cost effective?
Natalie: Yeah, it tracked pretty closely to the cost of saving energy findings; so residential lighting is the lowest cost in terms of peak demand savings, and that was about—I think—$730-some dollars/kW; and then prescriptive rebates for medium and large commercial and industrial customers was next, and then small commercial/industrial rebate programs, residential HVAC and whole-home retrofits kind of came in next—and they were all pretty close together in terms of cost: like around the $2000 to $2500/kW value.
And then our low-income programs came in at the highest cost, which is also similar in what you find to the costs of saving energy research that we’ve done.
Tom: Yeah, it’s interesting. You know, Gridium serves the commercial market, and often we get lumped in with commercial and industrial. I found those results intriguing. Could you say more about what you think is going on there, in terms of the data on cost effectiveness in that sort of bucket?
Natalie: Yeah, you know, the way that commercial and industrial programs are categorized you get a lot of different measures in the programs. It’s not the same as say, a residential lighting program, which is going to be driven by—at this point—LEDs, right?
A commercial industrial custom program is going to have anything from agriculture to industrial process, and so that gives you a very broad range of cost and subsequent benefits. So, I guess just continuing on that C&I custom program example, we had a lot of programs that we were able to look at for this project. I think we had about 240 program years, which means that each program, if it’s implemented for one year, then it counts as one program year. And we looked at five years of data for the nine different states. So, they’re a pretty big part of the efficiency portfolio in the states that we looked at and they cover a broad range: it’s industrial, agricultural, commercial.
You could be looking at retrocommissioning, it could be a custom rebate. But we found that about a third of the programs had a cost of saving peak demand of less than $1,000/kW and then another third had between $1,000 and $2,000/kW. So two-thirds of the programs were under $2,000/kW, which is relatively low.
We also found that for the custom programs in warm and humid climates, the costs were a little bit lower than in some of the other parts of the country. And we found that the custom C&I programs that focus on retro commissioning, for example in California, in Maryland, in Illinois and Texas were much more variable. So that was some of our higher level findings, just kind of digging into that custom section.
I can talk more about prescriptive or the small business segment, if we want, but I don’t know how much time we want to spend on that particular topic.
Tom: No, I think it’s really interesting and mirrors some of our commercial experience as well. Back to the cost question—I mean, this is kind of the million-dollar question. You’ve been super conservative. You’ve basically said what are the total costs, not including any division of energy efficiency, and you come up sort of in the pretty reliably in the $1,000-$2,000/kW. To me, that’s an encouraging result and that would seem to be the conclusion from your study as well.
For those that aren’t sort of familiar with capital planning and utilities, why do you think that and why did you draw that conclusion from the data?
Natalie: I think that as we start to get more focused and get better data for peak demand savings, that number is going to come down because it really is the highest number that we could come up with at this point in time.
And it’s useful considering against your other supply and demand side options for the electricity system, you have to be careful when you’re making those comparisons between supply and demand side because a natural gas plant isn’t going to provide the same services that energy efficiency is or vice versa.
And so you just have to be aware of what exactly you’re comparing, and it’s not necessarily going to be apples to apples, even if it is just the “capital cost”. So it’s useful to start thinking about how efficiency can provide demand savings and what those costs are, and looking at how it performs compared to other resources. I would not… there’s lots of room for improvement in these numbers and making them more solid and that does not just come from gathering more data, but also from program administrators and utilities starting to be more thoughtful in how they’re determining what those peak demand savings are and coming up with more robust numbers and sharing them more broadly and being more transparent about the assumptions that go into them.
Tom: Yeah, I think it’s right. I think it’s the grid of the future and if you’re going to work on the grid of the future, you’ve got to work on these kinds of questions.
Let me just come back to the cost question again and just reiterate for the audience: I mean, these are program administrator costs. The participant costs are not in here. Again, why are we focusing on the program administrator costs? Why do you sort of not include the participant costs of these programs?
Natalie: Yeah.
Tom: Obviously, they join in program…
Natalie: Yeah, it’s certainly a simplifying assumption. In some of our other work that we’ve done, we have looked at the total cost of saved electricity, and that does include the participant cost.
One of the major barriers to calculating that is that not very many utilities or program administrators provide that data and whenever you start mixing and matching your data sets—you know, taking participant costs from over here and program administrator costs from over here, you end up with less-robust results.
And so we typically only look at the total cost of saving electricity or potentially, the total cost of saving peak demand if we’re able to get enough data from the utilities about what they estimate participant costs were. So, that is something we’re considering as we move forward, but there’s just not that much information that’s being reported on what participants’ costs are; not that many utilities provide it.
Millen: Our studies reflect a trend that we see, which is utilities and grid operators are paying more attention to peak demand. Why is that?
Natalie: Yeah, I think as we have this increasing need for a more flexible and resilient electricity system, with changing costs of generation and all of these different trends that are occurring: from electrification to just needing to build up reliability and resilience, you really have to start thinking about what all of efficiency and other DERs are offering.
What are the characteristics that they’re providing to the grid and what grid services do you need at what time of the day, at what time of the year, in what part of the country? And so, starting to dig into that and really thinking about efficiency as a resource and not just this thing that you install and it doesn’t have any impact—is really starting to become more important.
Millen: Let’s say I am a utility, what’s this study’s advice?
Natalie: Yeah, I can’t reiterate enough the need for good data and the need to track the savings that you are reporting or that you’re finding, and reporting them accurately, and being clear and transparent about how those savings are occurring.
You know, you can’t manage what you don’t measure, and if you’re not measuring anything then we’re never going to be able to start building up this body of research or articulate that efficiency really can be used in these ways that people haven’t historically thought of it as a resource.
Tom: It seems like you’re at the beginning of studying this and yet, you’ve taken it further than we’ve seen in our little research about how energy efficiency as a resource contributes to peak demand reductions. What’s next?
Natalie: Yeah, there’s so many questions that are left. (Laughs)
As I said, we’re still working on this project; we’re continuing to gather more data from additional states and trying to get newer data from the nine states that we started with.
We’re going to be digging in a little bit deeper to try and understand how program administrators or utilities take these demand reduction numbers that go into their efficiency filings and transfer them over to planning or to procurement in some way that reflects that, you know, they don’t need as much capacity or as much energy or they don’t need some other type of grid services as much as they did in the past because of efficiency.
It’ll be interesting to learn about that.
We’re also trying to understand what the different definitions of peak demand are that are being considered; as I’ve been harping on, you know, trying to get some more consistent reporting and better understanding so that we’ll be able to do robust comparisons across programs about what the peak demand costs are for efficiency. It would be great if you could look across a geographic region and say, “Residential air conditioning programs cost this much and save this much on average” and feel like we’re making a pretty solid comparison. But right now, it’s hard to tell if it’s in utility A’s territory you’re talking about peak demand reductions for one hour, and in utility B’s territory you’re talking about peak demand reductions over 100 hours. You know, can you really make that comparison and have it be robust? I don’t think so.
But those are all things that we’re continuing to think about and wrangling.
Tom: Fascinating.
Millen: Yes, thank you Natalie. This has been quite interesting and I really enjoyed it.
Tom: Natalie, I enjoyed the discussion and I hope our audience does as well. And look forward to having you back on the show.
Natalie: Great! Thank you guys so much.
