EVI-Pro Lite Updates Webinar (Text Version)

This is a text version of the video for EVI-Pro Lite Updates Webinar presented on Dec. 9, 2020.

Mollie: Hello, everyone, and thank you for joining today's EVI-Pro Lite updates webinar. My name is Mollie Putzig. I work in communications at the National Renewable Energy Lab, and I'll be the moderator for today's webinar. Before we get started, I would like to go over a few housekeeping items so you know how you can participate in today's event. During the webinar, attendees will be in listen-only mode. You can select audio connection options to listen through your computer audio or dial in through your phone. We will have a Q&A session at the end of today's webinar. You may submit questions for our speakers during the Q&A panel at any time during the webinar. If you are in the full-screen view, there is a question mark icon located on the floating toolbar at the lower right of your screen to open the Q&A panel. If you are in split-screen mode, the Q&A panel is already open and is located at the lower-right side of your screen. You may send in your questions at any time during the presentations. We'll collect those at the end and answer them during the Q&A session. If you have technical difficulties or need help during today's session, you can use the chat section to get support. The chat section appears as a comment bubble your control panel.

Now I'll introduce you to today's speakers. First, you'll hear from Matt Rahill. Matt is a project leader and web developer at the Alternative Fuels Data Center and the Clean Cities website. He works with a team of developers and data scientists to build web tools and share transportation data. Next, we'll have Eric Wood. Eric is a senior engineer at NREL leading EV infrastructure projects in California, New York, and in support of the U.S. Department of Energy. He had a background in mechanical engineering and enjoys working at the intersection of automotive, transportation, and power sectors. Finally, we have Lauren Spath-Luhring. Lauren is a project leader and software engineer in NREL's Center for Integrated Mobility Sciences. She works with researchers and developers to build web applications that share information across a variety of transportation and mobility topics. As you can see, these are the presenters for today's webinar. Matt will start with some brief background about the Alternative Fuels Data Center for anyone who isn't familiar with the site, and then Eric is going to talk about the EVI-Pro model and the latest updates to the online tool. After that, Lauren will provide a demo of the new features in the tool and the API. We'll wrap up with a quick note about working with the Clean Cities coalitions, and we'll have plenty of time at the end for questions. Matt, I'll turn it over to you.

Matt: Thanks, Mollie. I know we have a number of Clean Cities coordinators and others on the webinar who are familiar with the Alternative Fuels Data Center, so for those who aren't, I want to share briefly about this website that hosts the EVI-Pro Lite tool. The AFDC provides a wealth of information and data on alternative and renewable fuels, advanced vehicles, fuel-saving strategies, and emerging transportation technologies. The site also features a number of interactive tools, calculators, and mapping applications that aid in implementing these fuels, vehicles, and strategies. The AFDC is a resource provided by the U.S. Department of Energy, and it is administered by NREL.

We launched the site in 1991 as a repository for performance data on alternative fuel vehicles. Since then, it has evolved to become an invaluable resource for fleet managers, fuel providers, policymakers, Clean Cities coalitions, and others who are working to make a transportation system more efficient, cut costs, and reduce emissions. The site has more than 3 million users annually, and we get more than 25 million hits on our station's web service each year, so that gives you an idea of the scope and reach of the AFDC. In addition to the audiences I mentioned, utilities can find a number of resources on the AFDC, and the new version of the EVI-Pro Lite tool is a really great example of that. This tool was first launched on the AFDC in May 2018. The tool was originally focused on helping local transportation planners estimate how much electric vehicle charging might be needed in their area to support a growing number of EVs. With the recent updates to the tool, the tool now has an expanded scope to help utilities gauge how EV charging affects their electric load. Eric is going to dive into that more, but I first want to point out how you can find the tool on the AFDC.

From the homepage, you can find the EVI-Pro Lite tool linked from the large feature space in the center, so that's a quick way to get to the tool for now. In the long term, you can find it on the main tools page, which is available from the navigation bar at the top. From there, you'll see EVI-Pro Lite at the bottom of the calculator section on the left side. Of course, you can always just search for EVI-Pro Lite to find the tool directly. Okay, let's go to the next slide and I'll turn it over to Eric.

Eric: Thanks, Matt. My name is Eric Wood, and I want to provide an overview of both the EVI-Pro model and the Lite version of EVI-Pro that Matt described as being hosted on the AFDC. As we move to the next slide, I just want to provide a little bit of context for EVI-Pro in general, so kind of the heavier version of the model. EVI-Pro is a simulation model that's designed to try and estimate the charging demand from electric vehicles, and design and supply charging infrastructure that's capable of meeting that simulated demand. The model relies a lot on real-world data. It also is heavily integrated with other models that describe vehicle adoption and sales, travel and mobility, station economics, and the electric grid. The model was originally developed through a collaboration between the California Energy Commission and has since been applied to studies across the U.S., including at the city, state, and national levels.

This slide shows a couple examples of the types of outputs that EVI-Pro provides. In the top right, we see a spatial distribution of charging demand and simulated infrastructure from a San Francisco case study, and then in the bottom right, we're looking at the statewide charging load profile from a recent analysis conducted for the state of California. Moving to the next slide, we can show what I would describe as kind of a semi-exhaustive diagram of how the model works, and so the two steps in the model are really highlighted in this green box, by the Charging Behavior and Network Design modules. These are what we're using to estimate demand and design and supply of infrastructure, but the model also has feedback loops to things that describe vehicle sales, driving behavior, station economics, ride-hailing, grid capacity constraints, so it's well connected to lots of other tools. But, what's really at the heart of the model is going to be described on the next few slides, so we'll advance here and show an example of an individual simulation within EVI-Pro.

This is an example of a travel day consisting of four trips, and, as we advance the slide here, what we do is we take this real-world travel data, typically from gasoline vehicles, and we try to simulate it as though it were attempted to be driven in an electric vehicle. Then, as we advance the slide, we can see that the model will resolve the charging behavior for that individual travel day. In this case, we're still showing a simulation for a 250-mile nominal battery-electric vehicle that has access to 50-kilowatt fast charging and a 7-kilowatt A/C charging at the home location. In this example, we're showing a scenario where the vehicle charges every night and has access to a charger at their home location. You can see kind of what the state-of-charge profile and the driving and charging behavior look like in that plot. We can advance the slide and bring in an additional scenario, and so this is kind of the power of the model, is that it allows us to experiment with what charging behavior could look like under different scenarios. In the scenario illustrated in blue, we're looking at simulation where the individual may elect not to charge their vehicle every night at home, even though they have access, and occasionally make use of infrastructure away from their home location.

In this case, we see a simulated public charging event over the noon hour, and then, advancing one more slide, we'll pull in a third example here where we simulate an individual that does not have access to home charging and is completely reliant on public infrastructure, and particularly DC fast charging infrastructure to make it through their travel day with a reasonable state of charge. I'll go ahead and ask for one more slide advance here, and then we'll just highlight all of the different charging solutions that the model is able to identify for this example travel day. What we try to do with the model is run the simulation over thousands of example travel days from a region to understand in aggregate what demand looks like, both in time and space, and what the electrical load from that charging may be on the power grid. If we advance one more slide here, we'll just take a brief moment to touch on EVI-Pro Lite. As Matt described, back in 2018 we originally launched EVI-Pro Lite on the AFDC, and the goal was really to try and expose the analytic capabilities of EVI-Pro to a much larger audience.

We've been fortunate enough to conduct a handful of studies regionally and nationally with EVI-Pro, but we know that the model is valuable and has insights that can be provided to a broad group of stakeholders that are making decisions around investment in charging infrastructure at lots of different levels. We're trying to expose those analytic capabilities of the model in an intuitive way, in a user-friendly way that really kind of removes this barrier of having to have a strong programming background, or wielding this complicated model to make these estimates, and so far I think we've had pretty good response in terms of use of EVI-Pro Lite. As we advance the slide here, we can look at some anecdotal examples that we've heard back from users of the tool, both in the public and private sectors, so we know that the tool is being used by Federal Highway in some of their alternative fuel corridors conversations. The tool has been used in Broward County and by the Hawaiian Electric Company to look at scenarios for infrastructure build-out and increased transportation electrification in their regions.

Tesla and NYSERDA are both using the tool in conversations around planning at the city and state level in different parts of the U.S. In addition to those anecdotal examples, we're also able to track use of the tool in a more analytic way, and so since its launch a little over two years ago, we've seen over 10,000 unique users viewing the tool, a total of 24,000 page visits, and almost three and a half minutes per page visit. That three and a half minutes really stands out to us as something that reflects engagement with the tool and people experimenting with different scenarios. We've really been happy, I think, with the level of engagement so far. In advancing a slide here, we can see kind of another example of some feedback of the tool that has kind of been reinforcing for us, and so I'll just read this off. This was from an annual merit review at the Department of Energy's annual review in 2019. The reviewer said, "Municipal and regional governments typically do not have the resources to understand their charging infrastructure needs. Having a quick online tool that gives a ballpark estimate of charging needs is a deeply helpful service. I have witnessed firsthand the amazement when city-level sustainability staff first use EVI-Pro Lite."

This kind of feedback motivated us to find additional ways to provide value through EVI-Pro Lite. One of the big areas that we thought was a growth area for the tool was thinking about the charging load, the electrical load that would be needed to fuel EVs under all of these different scenarios for vehicle technology and charging behavior, and so we'll go ahead and advance the slide here and just kind of highlight that with the recent NREL study. This is a study looking at electricity demand in the U.S. across all sectors, including transportation, but also commercial, residential, and industrial. At the time of the study back in 2018, transportation made up a very small share of electricity consumption in the U.S. But, as we look towards the future and consider scenarios for adoption of electric vehicles, we find that there is significant potential for EVs to add electrical load to the grid or to the U.S. power system, potentially getting as high as a quarter of all energy consumption by the year 2050 for most electrification scenarios. This kind of figure tends to prompt a very familiar response from all the stakeholders in the industry, and that's shown on the next slide.

I'll just kind of summarize it in a simple question, "Are EVs going to break the grid?" and maybe just advance one slide here. Yeah, so this is kind of the typical question that comes up in this space, and like many simple questions, it doesn't necessarily have a simple answer, but there is a lot of research going on in this space. If we advance the slide, we can take a look at some of those studies that have been happening the last several years. What I'm really showing here is that kind of behind the scenes in some way, EVI-Pro has been used in individual projects through the Department of Energy and NREL for informing EV–grid impacts analysis at the national, regional, and city level, including in Los Angeles, as well as studies of the Western Interconnection, and looking at the U.S. power system at large. With this update to EVI-Pro Lite, we're really trying to expose those capabilities to a broader audience, so advancing a slide here, the idea was that we really wanted to expose this capability of the tool, so looking at charging load profiles specifically, and using those to inform demand side models of what energy consumption on the electric grid could look like.

The work that we've done to update EVI-Pro Lite has been supported by the U.S. Department of Energy. We're also very fortunate to collaborate with the talented staff at Lawrence Berkeley National Laboratory, as well as at the Schatz Energy Research Center at Humboldt State University. This collaboration involved trying to define scenarios that would be useful for stakeholders, developing analytic capabilities to analyze load profiles, and eventually developing what we think is a really flexible tool and an intuitive tool for looking at different scenarios. This project did have a lot of industry engagement and feedback, so we reviewed the tool and approaches with various stakeholders, including electric utilities, automotive manufacturers, charging network companies, as well as city and state government, and various research institutes across the U.S. And so upon all of this review and discussion, really I think the two primary goals of the updates are to try to expose users to these electrical load projections in two ways. First, we want to be able to provide a simplified interface that makes these profiles accessible in a really intuitive way that lowers the bar of entry to try to understand what kind of scenarios the tool can provide.

But, at the same time, we also recognize that there is another user that is looking for a programmatic interface and trying to run thousands of scenarios and simulating those demand scenarios in models of the electric power system, and so we're also trying to set up the tool to allow API access and people that access the data in a way that is much more automated than pointing and clicking through the website. I'll go ahead and advance through the next couple slides here. This first one, what I want to do is just highlight all of the different aspects and uncertainties around the size and shape of electrical loads from EV charging that we've observed in previous studies and have really tried to design EVI-Pro Lite around. The first of those sensitivities is around vehicle technology. This is an example from a study we ran a few years ago in Massachusetts, and we're highlighting here the different load shapes that you can observe from simulating a plug-in hybrid versus a battery-electric vehicle in EVI-Pro and EVI-Pro Lite.

Essentially, what we're seeing here is the shorter range for the plug-in hybrid results, more charging away from home in order to try to replenish the battery's state of charge during the day for high-mileage days. But most of the residential charging is happening at a low, level-one power rate, whereas the battery-electric vehicle, with 200 miles of range, we would simulate with very low utilization of charging away from home but a higher power level per vehicle at the home location, and likely higher peaks, as well, in those critical late-afternoon hours. In addition to vehicle technology, we also want the tool to reflect different levels of residential access to charging across the U.S. Here, we highlight a couple simulation results from a study in Atlanta, Georgia, where we looked at scenarios where the vehicle fleet either had or did not have access to charging at their home or their residential location. This results in very different load profile scenarios, both in terms of the time of the day that the charging happens, but also the location.

As you would expect, the home-dominant charging has charging happening at lots of residential locations, and, in an uncontrolled environment, this charging would result in people arriving at home in the late-afternoon hours and plugging in to charge immediately upon arrival and kind of create a peak at a very difficult time of day for lots of utilities and certain seasons of their generation season. The no-home charging scenario, on the other hand, really shows a peak when people get out and about at the beginning of their day in the morning hours, start driving around, arriving at work, plugging in at work to charge, and making use of public charging around town, so very different scenarios there. Residential access is something we wanted to highlight. In the near term, I think it's fair to assume that residential access is probably going to be very high for most EV owners, but there is some emerging research that suggests in the long term, or in deep electrification scenarios, residential access may decrease significantly as we look to electrify vehicles owned by residents of multi-unit dwellings and apartment buildings.

Another dimension that we wanted to highlight was charging behavior, and so in this scenario, we were looking at a simulation from the Denver, Colorado, area that tried to explore what the impact of access to free workplace charging might be across the Denver Metro. Not surprisingly, making infrastructure and cheap electricity available to individuals can really incentivize people to charge earlier in the day and at their workplace location, which is a very different scenario in terms of the power sector, and, in some cases, may be better aligned with solar generation during the day, so kind of a synergy with renewable energy that is highlighted by differences in charging behavior that are enabled through infrastructure access.

The last scenario that I want to highlight is one around load flexibility. It's fairly well documented, as we move to the next slide, that light-duty vehicles are parked for the majority of their lifetime, around 95 percent of a vehicle's life may be parked, and so that provides a lot of opportunity for charging to happen at different times of day. In this example, we're showing some results from a Los Angeles simulation where we simulated charging happening as soon as possible at home and workplace locations, as well as as late as possible at home and workplace locations, essentially trying to complete the charge just before departure, either for the morning commute or returning home at the end of the day. Not surprisingly, the significantly long dwell times at home and work locations, on the order of ten hours in many cases, enables a lot of flexibility in the time of day that this load shows up as, and that's something that is very critical to a supply-side analysis of the power sector.

We'll just advance one more slide here and highlight all of the different scenarios that we tried to embed within EVI-Pro Lite, so this is an exhaustive list of the parameters that are available now to users of the Lite tool. This list actually represents approximately 150,000 different charging scenarios or load profile scenarios that can be examined, and so that presents a really challenging task for visualization and for accessing these data, which is where I'll invite my colleague, Lauren Spath-Luhring, to come in and describe how she and her team developed EVI-Pro Lite to host all of this data and make it accessible in a really intuitive way.

Mollie: As we're moving over to Lauren here, I just wanted to remind everyone you can submit questions at any time. In case you didn't hear it with our audio blip at the beginning, you may submit questions for our speakers using the Q&A panel. If you're in the full-screen view, click the question mark icon located on the floating toolbar at the lower right of your screen to open the Q&A panel. If you're in split-screen mode, the Q&A panel is already open and is located in the lower right side of your screen. Thanks, and you can take it away, Lauren. Lauren, it sounds like you're muted.

Lauren: Sorry. Can you hear me now?

Mollie: Yes.

Lauren: All right. I rejoined on the computer. Let me try this again. Can you all see and hear me now?

Mollie: Yes.

Lauren: Great. Thanks, Mollie, and thanks for your patience, everyone, as we deal with some of these issues today. Welcome. I'm really excited to walk through the new tool that Eric just described to you. What you see here is EVI-Pro Lite, which can be found on the AFDC. Highlighted in green is the Charging Need tab, which is here. This is what the tool used to look like, and today I'll be demonstrating the capabilities of this new Load Profile tab. If you click here, the first thing that you'll see is this form that asks you for your location and for the size of the fleet that you would like to use in your scenario. I'll go ahead and click Colorado since that's where I live, and I'll select Boulder. We've built in some fleet sizes for you to choose from, so 1,000, 10,000, 30,000, or more. You can select more and add a fleet size that's larger than 30,000, but you will notice that if you try to add a number that's too great, you'll be stopped. In this example, I put in 200,000 and this is telling me that in the area I selected, there are only about 101,000 light-duty vehicles in total on the road, so we couldn't have more EVs than that. I'll go ahead and select 10,000 for this example and then hit the Calculate button.

This takes us to our results screen. Here, you can see some of the inputs that I've provided at the top that lets me know where I'm looking, and it also lets me know that we're looking at a fleet size of 10,000 plug-in vehicles. There are two charts here on the left-hand side, one that shows the weekday electric load and one that shows the weekend electric load. In both of these, there are the various charging types broken out, and these can be toggled on and off. You can see how the chart adjusts as you do this. Hovering over the chart shows you the time of day, and the power and megawatts that is at the particular point. On the right-hand side are the additional assumptions that were used to create this shape. Everything here, from your initial screen, is able to be edited, and I'll walk through some of these inputs. You'll notice that these top three questions here have an orange question mark icon. Clicking that provides context to help make an informed decision. In this example, I clicked this question mark and it lets me know that, based on this location, a typical day's temperature.

Starting at the top here, I'll just go through these inputs. Fleet we've talked about, and there is some reference information. Average daily vehicle miles traveled ‒ 25, 35, or 45 ‒ and ambient air temperature, or ambient temperature. Moving down, there are more questions about charging strategy and the fleet make-up. Here, you'll notice that the icons have an i, and these tool tips provide information about the filters themselves, so it's clarifying and lets you know that for this question, "Here is what we mean by sedan. Here is what we mean by SUV." As you go down, there are more questions about the charging strategies, including mix of workplace charging, access to home charging, like Eric mentioned, and this includes sort of a sub-question here about the mix. If I click the info button for this one, it lets me know, "Here's what we mean by 50, 75, or 100 percent with regard to home access charging." Then, underneath that, you can be even more specific and choose the preferred mix of home charging.

By default, we have filled in these ‒ as I mentioned, all of them are editable. I'm going to keep going down since I started going through the questions. When I get to the bottom here, these last two questions are meant to represent flexibility in the load shape. You'll notice there are four scenarios here for a home charging strategy and three for workplace charging. It's because with the home strategy, we have a delayed start-at-midnight option. We don't assume that people will go to their workplace and start charging at midnight, so that's why that's not included there. With all of these, when you change an input, you'll notice that the charts on the left-hand side are grayed out and this recalculate button appears. Now that I've changed something, the chart no longer reflects the selections over here on the right-hand side. When I click Recalculate, you'll see that they adjust themselves and shift.

Another thing that we have built in is the ability to compare scenarios. Right now, this is showing what we call our base scenario. At the top, you'll see this orange button that lets you add a comparison. You can add up to five. Clicking on that lets you choose the best and worst-case scenarios. These represent the minimum peak load and the maximum peak load, and also the ability to add another custom scenario. I'll click Best, and you can see now a couple of things have changed. On the right-hand side, this panel now shows what we've called the best case for this location, and you'll notice now that these inputs cannot be changed as this is sort of pre-built and this is a predefined scenario. Scrolling down, though, your base scenario is still there. If you click on that, you can see your initial inputs for your base scenario. You'll notice that the charts on the left-hand side also changed, so we see now just a single line representing each of the scenarios rather than the stacked plot chart that showed charging at various levels for a scenario. This keeps things kind of more readable.

Like I mentioned, you can add up to five, so even if we add another custom scenario, you'll see another line. Right now, there are just two because, by default, the second custom scenario replicates all the inputs from the initial. Let's say I really want to focus on August or something, and Colorado is pretty hot, so I'm going to change the average temperature for my second scenario and recalculate, and now again the line has shifted. At any time, if you want to delete a scenario, you can click this trash can icon, and that will remove it. When you get that down to a single scenario, you'll see that the stacked plot chart reappears showing you the various charging levels. Speaking of the chart, if you click here on the top right-hand corner of the chart, there is the ability to download chart images or to export them. At the bottom of the chart, you can download the profile data that is behind these charts. This will be a CSV.

Going all the way down to the bottom, I wanted to point out these links. The first is to assumptions. Clicking that will take you to the underlying assumptions and methodologies for the EVI-Pro model and for the EVI-Pro Lite tool that you're looking at. Some of this content is old, or not old, but it's been here before for the existing tool. The part that we've added here, Section 3, describes the updates for this new load profile tool. Clicking on that, the page jumps down and you can see some of the methodology, the assumptions that were used, the various inputs, and some reference material down at the bottom. Going back to the tool, we also have a link to our developer API. I'll open that in a new tab. This takes you to NREL Developer Network, and here is where we have information about the underlying APIs that power the EVI-Pro Lite tool. We have two APIs, and I should back up and say that an API, or an application programming interface, is a way for one machine to access data from the database directly from another, so it's machine way to transfer and to share data.

We have two API endpoints for the daily load profile, one which is broken down by EV type. Clicking on one of them opens it up to show some of the documentation here, the various inputs that are required and the options, and some example output down here at the bottom. In order to use an API at NREL Developer Network, you need to have an API key. You can sign up for one here by clicking this. It's just a simple form that asks you just a little bit of information about you and how you might use the API. The last thing I wanted to point out here is that while these APIs show 24-hour load profiles, we have heard from some people that it would be useful to concatenate these in order to make an annual load profile. In order to facilitate this process, we have linked to a GitHub repository that has a Python script for people that are interested in that. Going back to the tool here, that was it for the demo. I think I'm the last speaker.

Matt: Thank you, Lauren. I actually have one more slide. Mollie, if you wanted to switch back to that.

Mollie: Yeah.

Matt: For anyone who isn't familiar with the Clean Cities coalitions, I want to draw your attention to this national network. The coalitions serve as local transportation experts and ambassadors, and they're really a great resource for anyone working on local transportation projects. You can use the Clean Cities website to find the coalition in your area and get in touch with your local coordinator. Specific to EVI-Pro Lite, I know that some coalitions have used the tool to help city transportation planners set goals for charging infrastructure needed in their area to support an anticipated number of EVs. With the new features in the tool, we're hoping it will help utilities and transportation planners discuss the impact of charging on electric loads and plan for the future of EVs. For those Clean Cities coordinators who are on the webinar today, I want to encourage you to use the new load profile scenarios to spark discussions with utilities in your area. That brings us to the end of the presentation. Mollie, could you start the Q&A portion?

Mollie: Yes. Thank you to all of our speakers. As a reminder, you can go ahead and submit text questions at any time. If you didn't hear at the beginning, you can submit text questions through your control panel using the question mark icon located on the floating toolbar at the lower right, or, if you're in split-screen mode, the Q&A panel is already open and is located at the lower right side of your screen. Our first question here is a fairly quick one, "Is EVI-Pro Lite focused on light-duty EVs only?"

Eric: Yeah, so I'll get my camera going here again for some of the questions. EVI-Pro Lite is currently limited to light-duty electric vehicles, and particularly in a traditional, personal ownership context, so we're not looking at taxicabs or public vehicle fleets for light-duty vehicles. It's really focused on personal light-duty. However, EVI-Pro Lite I think is really valuable in terms of trying to migrate and expose some of the research that's being done using the full version of EVI-Pro into the public realm, and so we do have work on medium- the heavy-duty electrification, and would be eager to migrate some of that work into EVI-Pro Lite or a similar tool in the future. But, at the moment, it is light-duty focused.

Mollie: Okay, so the next question is sort of a follow-on to that. Someone wants to know, saying, "These results are for light-duty vehicles. What about trucks, motorcycles, and other types of vehicles?"

Eric: Yeah, it's very similar there. The tool is light-duty focused currently, but we're eager and looking for opportunities to migrate more use cases into this or a similar tool, including for commercial vehicles, heavy-duty, motorcycles, and taxi caps and transportation network companies. There are lots of exciting and emerging transportation electrification modes out there, and we're kind of starting here with light-duty but eager to look at additional modes.

Mollie: Okay. Great. The next question is sort of multi-part. I'll break it up a little bit, starting with, "Is there a way to have the option to select other fleet total options in smaller increments?"

Eric: I can try that one, and maybe I'll invite Lauren to clean up anything that I get wrong. Currently, there is a discreet set of small fleet sizes that are available, and so those were limited to 1,000, 10,000, and 30,000. The tool is really designed to think about load profiles in an aggregate sense, so looking at all of the charging that happens across the region, not necessarily charging that just happens at a given site or a given facility. The smaller fleet sizes that we have generated, as you'll see in the tool, are quite a bit what I would call noisier, so they've got a lot more variability from time step to time step than the larger fleet sizes. This is a result of a smaller sample size of vehicles being simulated resulting in a noisier output. Since the tool relies on a discreet set of simulated outputs, we had to be a little bit discerning in deciding how many different fleet sizes to simulate. You will notice, though, that on the high end, you've got the option to input any fleet size above 30,000 that's within the light-duty fleet size for that region. For that, as you might suspect, we're simply scaling the 30,000 fleet size result up to larger fleets. As we get to about 30,000 vehicles, we find that the load profile shape tends to stabilize and become more smooth, and so scaling becomes an appropriate assumption at that point, which, unfortunately, is not for smaller fleets.

Mollie: Great. Thanks. The next part of that question mentions that the term fleet may be confused with commercial fleets rather than the overall area EV population, so what are your thoughts about using the term fleet? Yeah, it goes on to say, "We are discussing many medium-duty, heavy-duty commercial fleet considerations, so could the term cause a question with users?"

Eric: Yeah, I definitely recognize the confusion there and the tendency for fleet to be used in a commercial context. What we were aiming for was really looking for a descriptive term that described a large group of vehicles, even if they were under individual ownership, and so fleet is what we used to describe all of the light-duty vehicles, or I should say all of the light-duty personal vehicles being simulated in a region. I don't know that as a team we arrived at a better terminology for what to call that group of vehicles than fleet, but would be open to feedback from the audience or ideas from Matt and Lauren on the subject.

Mollie: Another user wants to know, "Is there an easy way I could access the dataset for an entire state rather than downloading one scenario at a time?"

Lauren: I can jump in here. There is no sense of geography in the API, so when you look at the tool, like I mentioned, on the initial form, you're asked to select a state and then an urban area or city. That's probably what you're thinking, you would have to go through each one and download the underlying data for each of the locations that you selected. Using the API, that becomes a little bit simpler. Well, I shouldn't say simpler. The API doesn't have a sense of geography, like I mentioned, but the other variable input that you could use, including temperature and fleet size, might help you get at a state data, but we don't have something that specifically puts out the data for a state because that, like I mentioned, doesn't have a sense of geography. You could manipulate the inputs ‒ fleet size, temperature ‒ and some of those other things to sort of replicate what your state might have, but we don't have something that specifically provides that information.

Mollie: Okay. Thanks, Lauren. We did have one attendee suggest an alternate to the word fleet just to think about. They said maybe call it a population instead of fleet. Something to think about. The next question here says, "Does the pro version have the ability to take marginal electricity costs at different times as inputs, and generate differences in total generation or distribution costs for different charging scenarios?"

Eric: A deep cut with that question. I like it. The model traditionally has been using flat electricity prices by time of day. Considering those different locations, we do have the ability to simulate different prices based on time of day. Actually, one of the scenarios in Lite does take advantage of that, so the delayed start-at-midnight charging for the home charging strategy tries to emulate a time-of-use rate where charging is delayed to start until midnight. Now, as far as the full model goes, marginal electricity prices aren't something that are considered. The tool is really kind of consumer or demand-oriented, and so we're mostly thinking about retail electricity prices currently. But, given the nature of that question, I would encourage whoever asked it maybe to follow up with me offline to continue that conversation.

Mollie: Great. Thanks, Eric. If you want to follow up with any of our speakers today, their contact information is on the screen. The next question here is, "Have any utilities used your modeling to help them set demand charge parameters?"

Eric: To set demand charge parameters? I'm unaware of a use case like that with the Lite tool. I will highlight, as I did in the slides, that HECO in Hawaii used the tool for some of their EV proposals. We know that the tool has been used in Maryland. We actually did a more detailed analysis for Maryland, in the public utilities commission there, so there are some examples of it being used. In New York state, the Department of Public Services has used the tool trying to set targets for infrastructure investment. But, when it comes to demand charges specifically, I'm not familiar with utilities using it in that way.

Mollie: Great. Thanks, Eric. I don't have any more questions at this time. For anyone on the line, please feel free to go ahead and submit your questions into the question panel. We'll give you another couple minutes here. Okay, we do have another question. "If, after using EVI-Pro Lite, there is an interest in using the full version, how can the full version of the tool be accessed? Apologies if this was addressed at the beginning. The audio was sort of spotty."

Eric: Yeah, so one of the things that I've really appreciated about the EVI-Pro project has been that we've constantly been in a state of development on the pro side but have found ways to try to make the tool accessible through the AFDC and through the Lite version as development is ongoing. Currently, we're undergoing development on a couple fronts with support from the California Energy Commission and NYSERDA in New York state to add additional features to the tool, and so we haven't gotten to the point yet where we're comfortable turning EVI-Pro into an open-source software project. That's certainly something that's been on our radar in terms of trying to expose it to a larger user base and in the pro format. However, just not to that point yet, and so use of the pro model requires partnership with NREL at this point to familiarize with the model and to deploy it in new instances or new geographies, or for additional scenarios including for commercial and medium- to heavy-duty vehicles. That's something that we're very open to and would be happy to follow up with folks online if there's interest in that topic.

Mollie: Great. The next question here says, "I noticed the results show unmanaged charging. Any considerations for managed charging scenarios?"

Eric: That's another very good question, and I apologize to Matt and Lauren for hogging the podium here. Yeah, so EVI-Pro and EVI-Pro Lite are really what I would call demand-side models, so they're looking at transportation requirements, vehicle attributes, charging behavior. They're not really taking into account anything on the supply side or from the power system explicitly into the model. Like I said, we do have the ability to consider time-of-use rates and different electricity types of scenarios in the model and try to simulate what consumer responses to those scenarios would look like. But, in order to do something with managed charging, we would need to couple EVI-Pro to another model of the electric power system, which we have done in a handful of projects with the support of the Department of Energy and the Vehicle Technologies Office. I would be happy to follow up with some more specific details on examples of those projects that have looked more closely at managed charging and the supply side of this equation.

Mollie: Great. The next question here is an API question, "Can I access this using Python? Where do I find the example of how I might do this?"

Lauren: Yeah, you can absolutely access the API using Python. It's a REST API, so it can be consumed by any programming language that can—you can certainly use Python. Reach out to me or through the developer portal if you need help with that, but certainly it can be accessed via Python.

Eric: I'll just add on there, Lauren, that the GitHub repo you mentioned that is accessible through the developer site has an example of Python script for accessing the API to help folks get started. That script is really kind of oriented towards concatenating these daily profiles into an annual load profile, but should have a lot of resources, if you're familiar with Python, for how to work with the data programmatically.

Lauren: That's a great point, Eric. That resources is there, and that's a great example of how you do it.

Mollie: Okay, I'm not seeing any other questions at this time. I'll just give it one more minute. If you do want to submit a question, go ahead and do that now. Okay, we have our most common question, "Are the slides going to be available?"

Matt: Yes, so we are planning to make the slides and the recording of the webinar available on the Clean Cities website.

Mollie: Another quick question, "What is the name of the link to GitHub repo?"

Eric: Oh, good question. We'll have to pull it up. I want to say that the repo is named EVI-Pro Lite and just contains a single Python script. Maybe Lauren can help me confirm that.

Lauren: That's what it's called, and I'll put a link to that repo here in the chat for whoever might want to access it. There is also a link to it from the API docs.

Eric: Yeah, and it's posted under the NREL organization on GitHub, so that's another way to try to navigate towards it.

Mollie: Okay, if you're looking for that link, it is in your chat box now from Lauren. Okay, that looks like all of our questions for today, so thank you, Matt, Eric, and Lauren, and thank you for everyone joining today's EVI-Pro Lite updates webinar. If you are looking for future updates like this, or news about EVI-Pro Lite, you can subscribe to the NREL transportation newsletter with the link on the last slide up here. Thank you, everyone, and have a great rest of your day.

Lauren: Thanks, Mollie.