Grid Integration Webinar: Exploring Renewable Energy Integration in California

>> Aaron Bloom: Good afternoon everyone and thanks for joining us for our second webinar on grid integration issues We are really excited to have a panel of three experts that have been looking at a lot of the evolving and upcoming issues that are being addressed by the California Energy markets right now We are just really excited to have such a nice group with us today If you guys have questions, we want to try to make this an interactive series, so please use the question feature under GoToMeeting or GoTo Webinar so that you can submit them into us I will do our very best to either interrupt the team while they are talking and make them feel awkward or just ask them one of your really good questions Don’t be worried though, sometimes people have a little overlap and I might mush them together some Please keep your questions coming With that, I am going to hand it over to Greg Brinkman >> Greg Brinkman: Great, thanks a lot, Aaron So today, I am going to be talking about the California 2030 low carbon grid study It’s a study that was sponsored by a wide and diverse group of funders in California and beyond A number of the funders came from industry and groups and foundation money This is a list here of the sponsors and the people who are on the steering committee Then we also had a technical review committee This is an independent committee that was not part of the funders of the study that looked at a lot of our assumptions We met a number of times throughout the study They helped sort of validate some of our methodologies and our interpretations and things like that I would like to say that the final report and conclusions are from the study team and may not represent the specific interpretations of any of these organizations on this list So some of the study objectives of this study, what we set out to answer some questions about what the feasibility would be of achieving a highway decarbonized electric sector in the year 2030 and so we are looking at 50 percent carbon emissions reductions below 2012 levels in California’s electric sector We want to look at both an economic assessment and what we are going to focus on today which is the analysis of integration issues that can happen from high penetrations of renewables that will need to achieve these carbon reductions We look a lot at some of these integration issues and potential curtailment This portion of the study that I will be presenting on today, also we did a lot of sensitivity analyses and different scenario analyses to understand the key drivers behind these challenges Today, we are only going to have time to present some of the bookend kind of sensitivities We did a number of different sensitivities that we will talk about later to understand sort of a detailed guide on some of these potential drivers of integration issues This slide looks at some of the renewable portfolios in each of our different scenarios We have three different renewable portfolios that we looked at On the left, you have the baseline scenario This is basically California’s LTPPP, long-term procurement planning process of 2014, 33 percent case When we started this study, we were before California had the 50 percent RPS We went with this baseline scenario as a sort of counterfactual case It is still an interesting counterfactual case, even though the legislation is in place for 50 plus percent RPS So you can see in this chart, which types of renewable generation make up the renewable portfolio This chart shows only the renewable generation I would also like to note that there is about 20 terawatt hours of energy efficiency of difference between the baseline and the two target portfolios We have two target portfolios that achieve approximately 55 to 57 percent renewable penetration These are the two cases that achieve the 50 percent carbon reduction So we have got two different portfolios to look at and compare The main difference between these portfolios is that the target portfolio is more diverse and the target high solar swaps some of that diversity for more PV This is more indicative of the procurement trends that exist today in California The difference between the two cases is some of the photovoltaics in the high solar case

are swapped for out of state wind and mostly in state geothermal and a little bit of biomass in the target portfolio that is more diverse All of the portfolios have the same assumed rooftop PV penetration One thing that we wanted to study and that we wanted to analyze in this study is the impact of some of our grid flexibility assumptions There are institutional and physical conditions on the grid that really have a significant impact on some of our integration challenges on the system We wanted to kind of look at some of the bookend cases and play with some of the assumptions to understand some of these key drivers What some of the bookend cases I will be presenting today are looking at what we have called the conventional flexibility and enhanced flexibility cases So on the left you have some suite of assumptions that makes up the conventional flexibility, on the right we have a suite of assumptions that make the enhanced flexibility I will mention here that the conventional flexibility represents some increase in flexibility beyond what we see today, and I’ll talk about that in the next slide The four things that make up the conventional flexibility assumption suite is we have a 70 percent requirement for out of state, California titled renewable generation, which also includes nuclear at Palo Verde and hydro at Hoover, and 70 percent of that generation must be imported at every hour of the year This is similar to some of the assumptions that California ISO is making It represents potentially sort of a proxy for the bucket rules in California In the enhanced flexibility case, we only have the physical limitations on imports and exports, no requirement that it is imported The second one is a 25 percent minimum generation rule in California balancing authorities that much come from local fossil fuels and hydro sources This is also a constraint that California ISO was modeling in their LTPP 2014 study process, but the relaxed this constraint very recently, or I should say they are planning on relaxing this constraint in future studies to be more of a California wide constraint This is out of concern for dynamic issues and controllability and things like that So as we get more studies to understand these issues in more detail, I think that we’ll be able to use a more precise assumption and requirement to deal with these types of issues In the enhanced flexibility case, we do not have any local generation requirements In the flexibility case, we have just the approximately 1.5 gigawatts battery storage to meet the CPUC requirement In the enhanced flexibility case, in addition to that we have a gigawatt of new pump hydro in California, and 1.2 gigawatts of new out of state compressed air energy storage at the end of the inter-mountain power project mine near Delta, Utah The fourth difference between the two cases deals with the ancillary service ability, the ability to provide ancillary services from hydro and pumped hydro resources For the conventional flexibility case, we basically tuned the ability of the hydro, how much capacity was allowed to fit into these markets so that the results matched the approximate ancillary service provision from the year 2013 I should mention that if you allow the model to constrain only the physical limits of hydro and pump storage, the model likes to use it to serve a very large percentage of the ancillary service provision, which is why we put these limits on here In the enhanced flexibility case, we have less strict limits on how much the hydro and pump storage can provide ancillary services So I should note here that the suite of these four assumptions, the sum total of all of them, can be much more significant than the sum of the individual impacts So combining these different constraints on flexibility, you start seeing these impacts more and more Now in the next slide I wanted to mention the key differences between today and the conventional flexibility assumption I just want to go through these relatively quickly In the model we assume that there is an optimal west wide dispatch We do have hurdle rates to serve as constraints for how much power can flow between regions,

and how optimally the power flows between regions, but it is a system wide optimization We assume that Diablo Canyon Nuclear Generating Station retires, it’s a zero carbon resource, so it would make hitting a carbon target easier, but it would potentially increase integration challenges We also have three million electric vehicles assumed to be on the road, which is in line with some of the governor’s announcements there That adds 13 terawatt/hours of load and gives the possibility of up to 3,000 megawatts of additional load, particularly during times of curtailment The nonrenewable generation fleet changes include some coal retirements outside of California In addition to the announced fleet retirements that we also have in the TESI 2024 common case, we also have the IPP power plant, which is assumed to retire in 2025 based on the LEDWPEIRP We also added some transmissions in the target portfolio cases including a north/south line from Idaho to Southern Nevada, approximately imitating the swift north line, and also assumed transmission from New Mexico to get it onto the Western River Pact Again, I mentioned this already, the rooftop we assume that it’s the same 24 terawatt hours in all of the scenarios As I said, I will be presenting mostly the bookend cases today, but we did model 23 different scenarios, and the different things that were varied within the scenarios were the resource investments, like I mentioned the diverse target portfolio compared to the high solar portfolio, both with and without additional storage We also looked at operational or institutional changes, which include those import requirements I talked about, the local generation requirements, the ancillary service provision, and also the ability to adjust and have flexibility in your import schedules Aaron, did you have a question? >> Aaron Bloom: Hey Greg, I’m getting some questions from people that are asking about the nuclear retirement assumption in Diablo Canyon Can you talk a little bit more about that please and what the impact is? >> Greg Brinkman: Yeah sure, we were really a little bit uncertain about what to do with Diablo Canyon So we wanted to do sort of a conservative assumption, which is to go ahead and retire it and have to replace it with additional renewable generation, which will make it more challenging for the system to operate So we weren’t really sure what to do with it because of some of the studies that have come out on how much it might cost for Diablo Canyon to get relicensed So we basically took the more challenging of the two assumptions and went ahead and retired it I think probably some of the people on this phone are familiar with some of the cost estimates that have come out for relicensing Diablo Canyon, in addition to it being on the fault line, it’s also got some cooling issues that I believe could cost billions or tens of billions of dollars to address >> Aaron Bloom: So I saw that you mentioned that we have geothermal resources, some CPV, how about offshore wind was that considered at all? >> Greg Brinkman: We did not include any additional offshore wind in this study Yep I believe in California it would mostly be floating platforms, which is cost prohibitive for the near term Potentially in a longer term as California goes beyond 50 percent, that could be on the table Thanks for the questions, keep them coming The next thing that we varied for our sensitivities is demand side flexibility We looked at cases with higher levels and lower levels of demand response We also modeled some other sort of conditions outside the model that have an impact on the ability to reach these levels of penetration which include higher west wide levels or renewable penetration, lower gas prices, high CO2 prices, and different hydro resource levels so we did model a wet year and a dry year If you want to see all those, you’ll have to read the report because I’m going to focus on some of the bookend scenarios that show what the impacts could be First, I just wanted to mention in the slides that California can achieve this 50 percent reduction in CO2 levels by 2030 under a wide variety of scenarios, all of our scenarios did meet the 50 percent carbon reduction target, except for the dry hydro assumptions So during a dry year with the build out that we have made, and with the conventional grid flexibility assumption, we don’t reach the target You will notice hear that there is some difference in emissions from the conventional and the enhanced flexibility cases This is primarily due to curtailment in the conventional flexibility cases, and so you have to operate more of your fossil fuel resources to make up for that curtailment So that’s why we see lower emissions in the enhanced flexibility cases compared to

the original Also, the portfolio does make some difference, especially in the conventional case, as in the high solar portfolio we do see a little bit more curtailment compared to the more diverse conventional portfolio So this slide shows the value of the renewable energy and energy efficiency, and that it depends on the rest of the system and the institutional framework In the target portfolio, we see $4.3 to $4.8 billion reduction in production costs This is compared to, I’ll give you a teaser for the capital cost assessment, which was done by JBS Energy This is slightly lower than the capital costs of these scenarios, but I’ll mention that in a few minutes So we also see here that in the baseline case, the enhanced flexibility assumptions are worth about $65,000,000 more per year in operation costs compared to the conventional flexibility assumption So 65 out of the 10 plus billion dollars of operational costs is a pretty small number The impact of those flexibility assumptions today in a low renewability penetration case is relatively modest However, in the target portfolio, as we get to renewable penetration levels above 50 percent, we start seeing over $500 billion in benefits So that flexibility that we don’t see a whole lot of benefit for today, we do see a lot of benefit for in a higher penetration scenario This slide shows the curtailment and curtailment, like some of these other outputs I’m talking about, varies significantly between the scenarios The last three bars in this chart are the curtailment percentage for the enhanced flexibility cases, and the right three are the conventional flexibility cases Really in the conventional flexibility cases, it’s the combination of that 25 percent local generation requirement with the 70 percent import requirement that drives the curtailment So in the high solar PV portfolio, we do see in the conventional flexibility case close to ten percent curtailment In a lot of the other cases that are more diverse and/or have better flexibility available on system, we see much lower curtailment levels, less than a percent in some cases except the target conventional flexibility case I don’t want to dwell too much on this slide but this shows the steepest ramp of the year in the target enhanced portfolio The green line there, you might recognize that as sort of a duck style curve, which Kyle is going to talk about a little bit more in a few minutes, as you have that real steep peak in the afternoon/evening hours as the Sun starts setting That’s the net load curve The other traces are the various supply and demand side resources that are contributing towards that ramp The table on the right show how many megawatts those resources are ramping between 3:00 and 4:00 pm, which was the steepest load ramp in this particular scenario These five resources, the physical imports, the storage, the gas dispatch, the demand response, and the hydro combine to provide most of the flexibility available on the system On this case, you can see it’s mostly the physical imports, the storage, and the gas dispatch, which combine to provide that ramping level But that is this particular ramp, if you look at other ramps at different days and different times in the system, you see these five sources of flexibility contributing differently, and sometimes the demand response and hydro do provide significant ramping But in this case, they didn’t >> Aaron Bloom: Greg, I’ve got a couple questions that are coming in that are asking about some of the underlying assumptions Can you remind us, what year are you guys comparing the carbon emissions against? 2016, what year is it? >> Greg Brinkman: It’s 2012 So we’re comparing to 2012 But California carbon emissions were relatively flat over some of those years, so depending on which year you look at, it doesn’t change the overall take home messages of our study much >> Aaron Bloom: Okay And about those gas prices $7.00 seemed a little high >> Greg Brinkman: Yeah, so the gas prices are I believe from EIA projections that WECC has gone through and adjusted slightly for western conditions Those gas prices, we’ve kind of gone back and forth a little bit on them, and it doesn’t change the main results of our study although it would have on impact on the cost effectiveness of the renewables And there are some sensitivities done on the gas price, not on our integration analysis but in the JBF’s energy capital cost analysis where JBF Energy and Bill Marcus compared the capital cost to these projection cost benefits

We’ve done some sensitivities on that It’s not that interesting for the flexibility and the integration questions, but potentially interesting for the capital cost side of things So the quick two bullet version of the overall conclusion of our study is that these results that I’ve shown today indicate that achieving a low carbon grid, 50 percent below 2012 levels, is possible by 2030 with relatively limited curtailment if the institutional frameworks are flexible If there are less flexible institutional frameworks, which is the conventional flexibility assumptions we talked about, and a less diverse generation portfolio which is the high solar portfolio that we talked about, this can cause higher curtailment up to ten percent, higher operational costs up to $800 million higher, and higher carbon emissions up to 14 percent higher So the conclusion kind of shows that if California goes forward and is able to achieve that diversity and flexibility the integration impacts are not that dramatic But if you start looking at less flexible institutional frameworks and less diverse generation portfolio, the impacts are more significant The companion studies that I mentioned a little bit, I’m going to give you the one bullet version and then in the next slide there is a link to the study website and all of these studies so that you can read up on them But Bill Marcus and JVS Energy found that annualized capital costs of those incremental renewables, transmission, and storage, for the target enhanced portfolio was about $5.1 billion, which is about $230 million more than the production cost reduction This would be about 0.6 percent of the annual revenue requirement in California and depending on what assumptions we look at, if you look at different gas prices, different discount rates and different economic conditions, Bill Marcus found that this could range from a three percent cost benefit to a six percent cost increase based on the annual revenue requirement GE Energy also looked at some of that dynamic issues in the LCGS scenarios This was a qualitative analysis, not a quantitative analysis We looked at some previous studies and looked at some of the impacts that we showed from this study and the production cost modeling and found that there could be some risks, although there are mitigation options that exist today that could provide reliability for lower costs and emissions compared to additional curtailment to provide that transient stability What we like to say is that we just need to focus on good engineering practices and certainly more followup work on these types of questions With that, that’s the end of my study This slide I believe will eventually be posted It does have links to these studies and a low carbon grid website >> Aaron Bloom: Great, well, thanks a lot, Greg We have gotten a lot of good questions from people I kind of intersperses them along the way This is just a reminder to participants, please send us your questions and we can answer them What do you know, we get one right away So Greg, here is a question for the alliance for you Why is it that geothermal isn’t integrated at the same level of solar and wind? Talk to us a little bit the geothermal assumptions you guys made >> Greg Brinkman: Yeah, sure So when we made the original diverse portfolio, we wanted to make a portfolio that was sort of a blend between what we might be seeing today with procurement trends and a system that we felt comfortable would be reasonable in its operation Then to study the impact of the different portfolios, that’s why we made the high solar portfolio We are not really putting necessarily a stamp of approval on these exact portfolios, but we built one to be diverse and one to be less diverse so that we could understand the differences between those You shouldn’t take this to be that this portfolio is exactly what ENREL is recommending for California in 2030 >> Aaron Bloom: Okay, I’ve got one more question that has kind of come up in a couple of different ways from users It’s about imports Can you talk to us a little bit about import capability and how you guys did some sensitivities around it? >> Greg Brinkman: Yeah, so for the import capability, California has a lot of import capability, clearly We did assume some coal retirement throughout the West Some of the renewables that we put in were out of state generation So we included in the target diverse portfolio, it does include about 12 terawatt hours of wind in Wyoming I believe that with nine or ten terawatt hours of wind in New Mexico As I mentioned before, in the conventional assumptions, we required that 70 percent of that energy must be imported by California at all times In the enhanced assumptions, we relaxed that and did not require that That is where a lot of the sort of curtailment started coming in is when we forced all that energy to come in

Other than that energy and the Palo Verde nuclear and the Hoover hydro, no energy was forced into California In the conventional assumption, because of that 70 percent rule, there is basically a minimum import level that California never imports less than 2,000 megawatts In the enhanced cases, you actually see some exporting from California We kind of explored the bookend to that and certainly allowing some flexibility around those imports as opposed to importing at the levels that California was importing today It makes a significant difference In the extra slides that I think we are going to post and in the report, there are some detailed analysis of the imports I will talk now, but if you have more questions, please email me or check those reports and the extra slides >> Aaron Bloom: Great, well thanks a lot Right now, what we are going to do is transition to talking with Paul Denholm who has got the work that builds on a lot of this analysis from the low carbon goods study to understand everybody’s favorite bird in the power sector, the duck >> Paul Denholm: Thanks Aaron Again, my name is Paul Denholm I have been looking at solar integration issues for a while here at ENREL along with others You can see that Greg Brinkman and Jenny Jorgensen, two of the authors of the LCGS study were also co-authors of this study We actually used a lot of the LCGS assumptions and modeming framework You have probably all seen this This is the duck chart This was published by the KAISO in 2013 They raised the big issue of the over generation risks So we wanted to kind of do a deep dive on this chart understanding what the implications of the duck chart mean for over generation and how much over generation might happen and all of the issues around over generation So over generation is simply when we have got more energy than we know what to do with Over generation can result from a lot of different factors, but a big part of it is flexibility on the grid A lot of this flexibility is both technical, as well as institutional That’s one of the big themes of flexibility work in general is understanding the differences between kind of technical flexibility on the grid, as well as kind of economic and institutional flexibility One of the things about over generation is that over generation is an easy problem to solve technically You simply turn off the wind or solar generator All wind generators can be curtailed All large-scale solar installations can be curtailed, simply It gets a little bit more complicated if you start having to curtail residential rooftops, PV, and very high penetration We may have to consider how we are able to better control rooftop installations, but certainly controlling large scale solar is easy The problem, of course, of curtailment is you are throwing away free energy You are throwing away zero cost, zero carbon energy and that makes it harder to meet your OPS goals and it makes the economics of wind and solar more challenging So the duck chart is nothing new We have been looking at the duck chart for ten years here at ENREL There are lots of other studies that have been done by various institutions E3 has done some studies Some folks with the Union of Concerned Scientists has done studies All of them have identified this issue around the duck shape and what can be done with it These studies have also discussed medication measures Here is one that talks about how to teach the duck to be a more streamlined shape by changing various factors on the grid We wanted to do a deep dive again and look at different penetrations of solar and understand how the duck chart shape change and how much curtailment might result What we did is we took the low carbon grid study framework, the modeling database and methods, but we modified it to make it more kind of today’s grid situation We did things like we unretired Diablo Canyon We uninstalled from the geothermal to maybe kind of look like today’s grid Then we started putting more solar onto it and see what would happen Here is a bunch of assumptions These are largely out of the LCGS study You can kind of read this on the slide deck or in the study itself We kind added Diablo Canyon back in and made sure that we had this system to be kind of more like today’s It looks like Aaron’s got a question >> Aaron Bloom: Yeah, so I was wondering when you are talking about these assumptions, what do you guys assume about the energy imbalance market? >> Paul Denholm: So as with the LCGS study, we did assume that California can kind of play with its neighbors kind of like it does right now with hurdle rates to add a little bit of friction, but we didn’t explicitly model the energy imbalance market in terms of stability to kind of true up short term imbalances Okay, so here is a duck chart for you This is March 29 This is an 11 percent annual solar situation You can see what the problem is

You can see that the net load drops down to well below 10,000 megawatts This is the kind of situation where you might see some curtailment It turns out that indeed, you can’t get the net load down to that level because of the minimum generation constraints on thermal, hydro, nuclear, and other types of generators So instead of ramping down to say 7,000 megawatts, the minimum generation level based on the assumptions we made, again trying to model that approximately a grid of today or a grid that might be under business as usual scenarios , we could only ramp the thermal and hydro fleet down to around 12,600 megawatts That results in curtailment On this particular day, about five percent of the potential of generation from wind and solar was curtailed Now, it’s important to note that while people talk about storage, there is already significant amounts of storage in California We weren’t able to effectively use that storage to avoid some of that curtailment By changing the normal patterns of pumping and discharge on the existing pump hydro fleet in California, in this case pumping during the day which in today’s grid would kind of be unheard of, but now we have got the situation where you have got a low cost energy available in the middle of the day We were about to avoid a couple thousand megawatts of curtailment during the middle of the day That’s a good thing, but even with the existing storage, we weren’t able to avoid all the curtailment Now, during the summer, things look fine because we don’t have that low demand situation We have got lots of demand in the middle of the day from air conditioning We don’t have a whole lot of curtailment or any curtailment during the summer Going back to our duck chart, if we try and force more PD energy on the system, we are just going to get more curtailment Here is the net low going from 11 percent scenario to 15 percent scenario You can see that net low going lower and lower and in this case, the net low would theoretically get down to 5,000 megawatts Again, we just can’t get that low All we can do is we can add a little bit of solar around the shoulder periods, but you can’t violate that minimum generation constraints on your hydrothermal fleet So something has to happen there Again, you are just going to get more and more curtailment if you try to force more solar off the grid The result of this is is you just get more and more curtailment That makes the economics of solar look worse and worse These curves show the average curtailment or curtailment in all of the PD energy, as well as the marginal curtailment or the curtailment of the incremental amount of solar You can see that by the time you get to say 20 percent solar, any additional amount of solar in the grid becomes less and less valuable In this case, we are only able to use between say 60 and 70 percent of any incremental energy at 20 percent solar So that is that marginal curtailment between 30 and 40 percent We are basically throwing away 30 to 40 percent of the incremental amount of solar energy by the time we get to 20 percent solar That corresponds to an increased cost or decreased benefit So the impact of curtailment has been expressed in a couple of different ways If you see these decreased values or energy mills at Lawrence Berkley National Lab has expressed this as a decreased value curve, sometimes here at ENREL, we express this is as an increased cost curve based on the fact that you are selling less energy so your cost of energy for the energy that you do sell goes up You can see that if you start with $0.06 per kilowatt-hour, which is where we are approaching in California, you could see that the marginal cost of solar is getting to $0.08, $0.09, $0.10 per kilowatt-hour Compared to other low carbon resources, solar becomes less and less competitive The question becomes what are we going to do? W We framed this in terms of either fattening the duck or flattening the duck A lot of people talk about flattening the duck, which is trying to avoid that deep below But we also like to talk about fattening the duck, which is how to do we accommodate the normal shape of solar? How do we accommodate that large solar production? Fattening the duck is simply letting that belly shape happen and seeing what we can do to accommodate a natural resource of the solar resource by decreasing the system minimum generation constraints If you can decrease the constraints of thermal and hydro generators, we can accommodate more of that midday solar Flattening the duck, of course, is by shifting or adding demand in the middle of the day and using that demand later through demand response or energy storage Those are the real two kind of general categories of accommodating solar by either flattening or fattening Here is the case where new storage will help flatten the duck In addition to the existing storage, California has the storage mandate that will add 1,325 megawatts of storage We can see that by adding that storage, we can shift demand from the middle of the day to later in the day We can reduce the curtailment instantaneous curtailment by a little over 1,000 megawatts

The interesting thing is that storage can also help fatten the duck by making the system more flexible I am not going to read this, but you can read this either here or in the report how that adding storage especially local storage, you can help decrease some of these requirements for operational flexibility Instead of using partially loaded thermal generators for things like providing operating reserves, frequency stability, and voltage stability, some of the things can be provided by storage As a result of fattening the duck, we can lower that minimum generation constraint Here, we have a situational where we would have been able to reduce the minimum generation level from that 12,600 megawatts to about 10,000 megawatts and accommodate an increased amount of solar Here is the reduction in curtailment that occurs by changing some of the constraints, enabling local storage, and local PD to provide some of the services We have reduced the instantaneous curtailment on this day from 7,000 megawatts to under 4,000 megawatts >> Aaron Bloom: Yeah, so one of the questions that I’m getting again is about some of the underlying assumptions with respect to hydro and nuclear plans How flexible are those in your model? >> Paul Denholm: For nuclear, that is simple We established Diablo Canyon as a base load plant where you do not change output For hydro, we had to use the hydro limits that were in the LCGS model Obviously, it’s tough when you talk about hydro because you have got to do detailed modeling, but we have kind of the standard assumptions that catalyst that is used in their LTPPP models for how much you can move the hydro plants So what we see when we can add this additional flexibility is a lot more solar on the grid with less curtailment Instead of having a situation where by the time you get to 20 percent solar, you get say 30 percent marginal curtailment, we have been able to shift these curtailment curves off to the right Now, we can start talking about say 20 to 25 percent or potentially even more penetration of solar with relatively low curtailment Then if we talk about a few other flexibility options, such as demand response, we are able to add additional great flexibility shifting additional load We can start talking about PD penetrations of again somewhere in the order of 25 to 30 percent By accommodating the natural duck shape and that is really what we think is going to be critically important here is both accommodating the natural shape of the duck curve, as well as doing some shifting, we can get to levels of PD penetration in the next decade that get kind of interesting from a standpoint of a diverse portfolio This doesn’t include a lot of other options Obviously, more electricity storage could help CST with thermal storage can also help, as well as greater interchange between regions We didn’t do any change in terms of how we let California play with its neighborhoods, but that is an option as well There are a lot of different options This kind of is a first shot at looking at some of these We need to do more research to understand how we can get even higher PD penetration and accommodate the duck shapes You can read these conclusions If you have any additional questions, I will be happy to answer those before we transition to Josh >> Aaron Bloom: So Paul, help me figure this out It looks like you can get to about 25 percent solar without adding storage is beyond the mandate Is that kind of what you are seeing >> Paul Denholm: Yeah, so looking at about 25 percent is where we started, 25, maybe even 30 percent is where we started seeing that kind of increase in curtailment At that point, we consider a lot of the flexibility, the kind of inherent flexibility in the grid Beyond that 25 to 30 percent, we do start to need to be a lot more aggressive and really kind of start looking maybe at additional storage for some of these other technologies That’s one of the reasons why we began to look more aggressively at the kind of the cost benefit of storage, especially because storage right now is expensive We really need to understand the potential benefits of storage, which is one of the reasons why Josh Eichmann undertook the study of the California storage mandate so we can start understanding these cost benefits of new storage and understand how storage can better play into the grid >> Aaron Bloom: Okay, great, well, actually a couple of more questions that are flying in here So I think we have got a little bit of extra time Paul, are you ready for a couple more? >> Paul Denholm: I can answer a couple more, but let’s make sure we have got plenty of time for Josh >> Aaron Bloom: Okay, have you guys looked at the cost benefit of thermal versus electric storage? >> Paul Denholm: Yeah, we have compared CSP thermal storage We haven’t done a whole lot of work with thermal storage in buildings I think thermal storage in buildings is one of the kinds of interesting opportunities for shipping air conditioning demand That I think is going to be really important Most of our work has been more on CSP thermal storage, which can do a lot of load shipping as well, but I think really thermal storage needs to be better understood, especially

in the California context where there is just a huge demand in the summer for electric, for air conditioning >> Aaron Bloom: Well, great Paul, and thanks everybody We will keep these questions coming in If we have a little bit of extra time after Josh is done, we can open it back up to the rest of the panel With that, Josh, are you ready to take it away? >> Josh Eichman: Yes, thanks I think it’s a nice transition to move from discussions on the duck curve and looking at the maximum amount of renewable penetration we can get and the sacrifices that we are making that point and then see what opportunities exist for energy storage to come in and help support that I am happy to be here presenting today and about this study to look at the really the value for storage We published it late last year, so it’s still pretty fresh for everyone We can go to the next slide, please What I want to do is just give you a little background on the storage in California, which many of you may be aware of and some of the modeling methodologies that we are using Look at the scenarios Then I am going to briefly look at the results, really kind of want to whet the appetite and then we will provide links afterwards and of course open to questions and any followup messages I am sure we can respond to this as well You can go to the next slide If you are aware, in California, there is a storage mandate for achieving 1.325 gigawatts worth of storage The type of storage isn’t specific, but it’s specific about when those procurements will happen and in which sectors those will happen You can see on the figure here, there is the three investor and utilities There is sort of a roll out of storage through the years They are in different sectors, transmission sited, distribution sited, or customer more like behind the meter storage We are following a lot of these assumptions in the model that we are working with You can go to the next slide To provide a little more background on that, we are also using similar to the LCGS and the duck paper, we are also using this long-term procurement plan information The ISO has developed this database, along with support from the BSE It follows along with the proceeding the LTPPP document We are going to use that model We are looking at 33 and 40 percent renewable generation What this model does is production cost model You are going to look at simulating unit commitment, energy dispatch, and provision of reserves It’s on the hourly time step I won’t go through all the details here But one of the interesting things that this 2014 version of the model, they implemented a value at which they have essentially placed a value on the curtailed energy This creates some kind of interesting things in the model Moving forward, I am not sure that maybe the best way to do that, but this is definitely one of the ways That’s what we are looking there Then in addition to using the production cost model, which is going to get you information about kind of the system costs and prices, emissions, and how the resource mechanisms are changing, we are also using a six price optimization model that uses historical prices and allows you to look at kind of the revenues You can go to the next slide For the scenario that we are looking at, I really separated them into two categories We have the base case scenarios and we have the sensitivities The base case, as you would expect, with the kind of no storage case where we are moving storage from the California system and seeing how it responds Then one where we are putting storage back in as prescribed by this storage mandate, which you see as the base there on the left Then the thermal, we are looking at providing only energy, so not being able to provide ancillary service for some of them Another one, we are looking at kind of a similar breakdown of for the energy, but then providing all reserves with that instead of energy and reserves Then a final one that provides reserves only It’s really allows us to kind of isolate where the value is coming from in terms of the services being provided For the sensitivities, we had thought for a long time about which ones to include

We ended up including a fairly long list of these One focus is on what if you were to provide predominantly regulation with this storage What if it’s longer duration storage? What if you change that a bit eventually to the point at which the value for curtailed renewable energy is placed We also have one that kind of goes along with the LCGS study where we are looking at this no export limitation If you relieved export constraints on the system, than what does that do? Also, looking at different levels of capacity for the storage If you go to the next slide Now, I want to go through just a few of these kinds of high-level results Like I said, hopefully, get you interested in matters of interest We will provide some of the links at the end You can read more By adding the storage, as compared to not having storage in the system, which includes California and the entire West, you see a reduction in the production costs of the system wide operating costs It can range from the 78 to 144 million with different renewable penetrations Then if you then know how much storage, we can install So you can equate that to essentially a dollar per kilowatt year value as well One of the interesting things that we found is that I actually avoided generator startup costs They play a significant part in this production cost reduction You see between 30 and 67 percent That is something that you are currently not going to get repaid for any kind of energy market setting or any other markets right now That was one of the interesting findings moving forward I think we will need to continue to find a way to address that You can go to the next slide So more of the base case results, if you are looking at the value and you are just providing ancillary service, we found you can achieve 90 percent of the value for adding storage Then adding, in addition to being able to provide ancillary service, then you provide energy and ancillary service optimized and then you will increase that ten percent more So really, the ancillary service, as I think many people recognize is one of the highest valued assets for storage Then there are issues then at the market depth because typically ancillary service markets are relatively shallow As soon as that is saturated, then in fact the price [inaudible] when we looked at storage and how that impacted the renewable curtailment, that was under this no exports We had still maintained the no export assumption We thought there as a reduction when the storage was added in the curtailment Actually, I think those numbers might be switched around I think we had greater curtailment reduction when we were in the 40 percent case, but we will fix that in the slides Then we also saw when you are adding storage and we look at carbon emissions, for California, it’s going to be able to reduce the in state emissions, but on the entire system, it has kind of a mixed impact We actually might depending on the resource mixture outside of California, if you are going to use your storage charges, and increase the amount of generation coming from coal, you could actually have a negative impact on emissions You can go to the next slide Those are the base case findings I wanted to do another high-level look at some of the sensitivity cases that we looked at One of the first ones was incremental storage, both on the capacity and the duration We were in both of those cases to see if you had a smaller portfolio, what is the value look like We found that as you add more storage and not surprisingly, as you add more storage to the system, the incremental value that additional unit of storage is going to be less than the previous one Then in terms of the duration, now you are going to take your storage and increase the duration that you provided for In the model, we have two, four, and six hours for the storage We were to increase that by either one hour or increase it by four hours, we can see a

relatively modest increase in the production costs reduction coming from that We can go to the next slide Just a few more of the sensitivity case results In terms of the regulation, if we are providing very specifically just that regulation As I mentioned, these can be pretty shallow so they can cause some of these zero price issues where you actually—and this is actually as a result of the model more than anything else because it may not happen that way in the actual markets A saturation of the regulation market by storage is then signaled by the zero prices in the model We saw those in that case I will do one more of these points When we looked at removing the [inaudible] limitation on California, what it does is it reduces a lot of the curtailment because you can import and export more freely This is a greater cooperation between these regions, but in terms of the storage value, it actually relieves some of the flexibility issues that you are seeing today, which overall, would reduce the value for storage It’s one of these kinds of back and forth issues We can go to the next slide Hopefully, I have provided a high-level overview, like I said A little bit quick, but obviously, there is the entire report Nice, light reading for you, if you are interested in reading up on that I would be happy to take questions now Thank you >> Aaron Bloom: Thanks a lot, Josh We actually have a few questions that are coming in one of the first questions I am going to take is about system inertia under these scenarios Greg, can you help us understand a little bit more about this >> Greg Brinkman: Yeah, sure I would reference you to if you are interested in this question about inertia to look at the GE report that was done as part of the low carbon grid study It is linked to in the study materials If you go to the low carbon grid study website, or look in this presentation, it’s there Now, they didn’t do a full model of these particular scenarios, but we did pull some analysis from the production cost run to look at the [inaudible] nonsynchronous generators and things like that in different regions Now, there is still because we haven’t done a full modeling effort on any of these scenarios, there are still some questions, and really there are questions there really are questions worldwide about how much of your generation can be synchronous versus nonsynchronous in a specific region, particularly can you have a small region in Southern California that has very little synchronous generation when it’s part of a huge grid that has a lot of synchronous generation It’s questions that we really haven’t answered yet, but we are doing some work and initial results have not shown a big problem We really need to do more work to get a more precise answer to that question PD and others are moving forward with this type of work >> Aaron Bloom: Great, well here is a question I have got for Josh Josh, can you comment a little bit about how increased renewable development in the rest of the West like Arizona or other states, might affect the ability to export surplus energy? >> Josh Denholm: Right, you have the idea of being a good neighbor Right now, it’s California has a very high mandate and they are headed towards 40 percent renewables and the rest of the West isn’t or is on a much slower trajectory You can sort of rely on some of the flexibility from their generation Then as they start, then I think you are not able to sort of offload that problem, which as I mentioned before when we are looking at releasing that export constraint, now all of a sudden, you are able to more freely interact with the rest of your regions of California who was unable to reduce their curtailment just by that increased level of flexibility If other regions are starting to add renewables, than I think storage would then play an option as well If every region in the West has similar renewable portfolio and California didn’t allow the exports because no one is going to be able to export We are all getting solar generation at roughly the same time You may get some regional aggregation for wind That may be to a less extent I can see an increasing the value of storage as other parts of the West add renewables >> Aaron Bloom: Thanks a lot, Josh Well, that just about wraps it up for us today We are looking forward to having a lot more communications like this in the future

So please stay tuned to your email for further distributions from myself or the rest of the team to talk about future projects One project that we are going to talk about in the next couple weeks is that Eastern renewable generation integration study In the meantime, if you guys have follow up questions, please write the authors, write me, aaron.bloom@enrel.gov We would love to hear more from you Thanks a lot for participating today We will see you next time