 My name is Yana Aranda. I'm the president of engineering for change. And I am so pleased to welcome you all to today's webinar with our incredible panel comprised of Henry Louie, Derek Terry, and Mohamed Baume, who will be giving us some insights on their work, find data-driven design for awkward systems and specifically looking at the case study of delivering electricity on the Navajo Nation here in the United States. So for those of you who are joining us for the very first time today, I wanna tell you a little bit about our organization, Engineering for Change, or you first see for short, is a non-profit organization founded jointly by the American Society of Mechanical Engineers, or ASME, as well as other engineering associations to prepare, educate, and activate the international technical workforce to advance quality of life and ensure that we benefit people and planet. At E4C, we provide upskilling and professional development opportunities at the intersection of engineering and sustainable development for our global community and particularly early career technical professionals through programs such as our Engineering for Change fellowship. And we support mission aligned organizations to achieve their sustainability objectives through our impact projects and services. There are more than 54,000 Engineering for Change members worldwide and a global audience of over one million people that believe engineering can change the world. I'm sure that you all count yourselves among that community. Of course, for more information about E4C, our programs and every opportunity that is available to you as members, I encourage you to visit our website. The link should be in the chat shortly. And of course, invite you to visit us, follow us on our social media channels. So I know that many of you since COVID are probably experts in Zoom and likely don't need this, but for those of you who are maybe joining us for the first time, I would like to make sure that we are getting familiar with the really critical functions, including chat. So at this time, I would like to invite you to please into the chat window, enter your location. Where are you joining us from today? I'm joining you today from Brooklyn, New York. I would love to see where you're joining us from today. If you don't see your chat window, just go to the bottom of your screen and type on the little icon. So we see folks from Ecuador and Somalia, Pakistan and Uganda, Seattle and Manchester, Nairobi, Jakarta. I don't know that flag. Oh, who put the flag and help me out here. Malawi and Amsterdam, welcome, welcome everyone. It's such a pleasure to have you from Yemen to Canada, thrilled to welcome you today to our webinar. Again, if the chat is not open on your screen, look for the chat icon on the bottom in the middle of the slides. And second, if you have any questions during the webinar, we encourage you to please use the Q&A button in order to enter your questions so we can keep track of them for our presenters. Do not just enter them into the chat. However, you are of course welcome to converse with your fellow webinar participants in that chat window or share any tips or any reflections. So if you're following us on Twitter today, please do join the conversation with our dedicated hashtag, hashtag E4C webinar series. So again, really pleased to see the global representation today on here. I see Slovenia and New Mexico and Nigeria and Sweden. And I have not seen people enter their flag icons and it's really testing my skills here. So thank you for that. I'm going to have some fun with flags later. I really appreciate you all being creative in your responses. So with this introduction, I would like to now turn it over to our fearless panelists to introduce themselves starting with Dr. Henry Louie. All right, good. Well, where I'm at, it's morning time. So I'll say good morning to everyone. The three of us are so excited to be here to be presenting to such an international audience on a topic that we find very interesting and it's always nice to see that so many people across the globe are also interested in our topic. We'll start by doing some brief introductions of the folks that will be speaking to you today and then we'll get right into the meat of it. So let me go ahead and go first. My name is Henry Louie. I'm a professor of electrical and computer engineering at Seattle University. We are a school of private Jesuit University in Seattle. I also am the president and co-founder of a nonprofit organization called Kilowatts for Humanity and we do electricity access primarily in and around sub-Saharan Africa. I've also been on various IEEE committees and other leadership positions. So it's really exciting for me to talk to you this morning and joining me today, we have Derek Terry. If you wanna unmute and do you a quick intro that would be delightful. Okay, good morning everybody. My name is Derek Terry, renewable energy specialist here for the Navajo Tributality Authority. And I'm glad that all the wonderful people and literally across the world, I see all the chat rooms and I'm, oh my gosh, I thought that we're just limited to Arizona but I guess it's all over. Welcome and we can dive into it here a little bit more when the discussion starts, but thank you and welcome. Great. And then last but certainly, last but certainly not least, go ahead, Mohamed. Hello everyone, my name is Mohamed Baoum. I'm a PhD candidate in industry and system engineering at Virginia Tech and I joined E4C as a fellow last summer and now I'm serving as a senior fellow. I'm always interested in finding intersection between engineering and global development and I'm really excited to be here with you. Okay, thank you, Mohamed and Derek. So to kind of center our discussion this morning is what we're really talking about is electricity access and it is maybe surprising to some of you but certainly not all of you based upon the locations in the chat that many places in the world struggle with access to electricity and unfortunately the Navajo Nation on parts of it is one of those locations. To paint the broad picture, we would describe access to electricity or not having access to electricity as a form of energy poverty. So energy poverty, a very simple definition is it's the lack of access to modern fuels. Now globally over two billion people rely on solid biomass or cooking and heating, for cooking and heating things like crop residue, dung, charcoal or wood that's how they heat their homes and cook their food. A smaller amount, although certainly substantial is 733 million people don't have access to the electricity grid. So that's approximately one out of every 10 people on the planet. And I should note that a far greater number than this might have electricity going to their house but it is unreliable or unaffordable. So 733 million is sort of the accepted lower bound of that number. So where do people live that don't have access to electricity? Well, primarily it's in Sub-Saharan Africa and villages like this, this is a village in Zambia. There's no power lines that make it to this village. And so this is endemic across the continent, Sub-Saharan Africa. So 70% of the people without access to electricity live in Sub-Saharan Africa. Another 20% live in South Asia and 10% live in the rest of the world. And in my work history, typically you find that a lot of organizations, a lot of attention is focused on Sub-Saharan Africa and South Asia and not that rest of the world slice. So it's important to not forget that even in places like the United States and a lot of our travel communities there are homes that don't have access to electricity. Some estimates that I've seen in North America put the number of people without access to the grid to be somewhere North of 100,000 people, which is a significant amount, maybe even up to 200,000, depending on what you count as access and what you don't. So our webinar today is primarily gonna focus on this electricity access on the Navajo Nation and the challenges with that, but also the opportunities and what we've learned from off-grid electrification on the Navajo Nation. So let me just give you just a very brief background of the electricity access situation. In the United States, rural electrification really took off in the 1930s with the intervention by the federal government that provided huge subsidies and loans to electrify rural areas. However, Navajo Nation, like many travel communities, travel lands in the United States were overlooked by that government intervention. So they didn't receive the benefits. Now, a reality of the situation is that to extend a power line, usually costs somewhere between $20,000 to $40,000 per mile. It's quite expensive. On the Navajo Nation, some of the homes are still remote that there may be 40 miles from the nearest power line. So you can think about how much it would cost to extend the power line to that one home that is 40 miles away. It would be over a million dollars. And it makes sense then that it's not the highest priority or it's not really economically justifiable to spend a million dollars to connect one home. In addition, Navajo Nation, historically, they've been raised sheep and goats and horses and so forth. And so it's a pastoral community where they are used to having lots of land surrounding their homes with neighbors quite far away. So there's low population density. So you combine the remoteness, the sparse population, you put those together and it's a perfect recipe for low electrification. It requires substantial subsidies to connect all the homes. So as a result, there's about 10,000, maybe even 20,000 homes that aren't connected to the grid on the Navajo Nation. Progress is being made, but it's still a substantial gap. And the Navajo Nation represents about the least electrified of all the US tribal reservations. So that's kind of a background. That's maybe why the situation is the way it is. And thankfully, the Navajo Tribal Utility Authority has been making a lot of progress in changing the situation. So let me turn it over to Derek here to talk a little bit more about NTUA. Let's see, Derek, you are muted. There you go. Okay, there you go. Good morning again. My name is Derek Terry. We're a renewable energy specialist here for Navajo Tribal Utility Authority. We have a company that's located here in Fort Defiance, Arizona. We are part of seven different districts throughout the whole Navajo Nation. So within our company, I'll just do a quick rundown of what we do. And so within our company, what we cover is the electric distribution transmission throughout the whole Navajo Nation, as well as communication, whether it be broadband, fiber, or mobile wireless, natural gas, water, wastewater, and photovoltaics. So that would be on-grid as well as off-grid. And then the last is our off-grid systems. So one of the stats I want to add, and then the first slide here is estimated that 31% of homes lack complete plumbing, 28% lack kitchen facilities, 38% lack water services, 32% lack electricity, 86% natural gas services, and finally, 60% lack regular landline telephone services. So in a composite, we're the majority utility company that oversees all these utilities throughout the whole Navajo Nation. So if you jump to the next slide, Henry. So the Navajo Nation, it covers four states. So we go into Colorado, Utah, Arizona, as well as New Mexico, with a complete area of almost 18,000 square miles. So it's a significant amount of area and a lot of it is still very, very remote. The distances that our crews travel for outages for installations of different utilities is substantial. For you to go from one end of the Navajo Nation all the way to the next, it could easily take you six to eight hours depending on the route that you take. And so hence the districts throughout the various areas of the Navajo Nation. And right in the middle there, if you can see the little area, that is the Hopi Resolation. So Hopi is lucky that Navajo is surrounded by Hopi. So, but they're a good neighbor to us and we're a good neighbor to them. So we help out on a lot of different projects. But in reality, that's the size of the Navajo Nation as it is currently next life. And so, with the inception of NTA, one of the programs that was thought about way early on back in the, I wanna say late 1980s was on the use of solar panels and renewable energy. So if you guys don't remember, solar is relatively new. There wasn't, schools weren't doing it, community colleges weren't teaching it. But now, if you go out and you just walk a few feet, everybody's talking about renewable energy, everybody's talking about sustainability and that's the new go-to word. But I'm glad that the council here with the Navajo Nation as well as the leaders here at NTA, they had a little bit of forthcoming. So back in the late 1980s, I believe in 1989, we had our first, I believe it was a 200-watt panel that we had, I think it was like four of them. So back in the day, they didn't have a 200-watt panel. I think they had like 60-watt or 40-watt panels. And then you just wired them all in a series with maybe a couple of this regular offline car batteries. So back then, one of the things that we wanted to be early on, we wanna be an example for other native communities that are in our same type situation. And so why is that? We gotta stress the importance of energy efficiency and teach that to everybody within the household. Why? Because of the property level, low income, unemployment rates, all-time highs here on Navajo Nation. So I did write down the couple of notes here. So for reference, the unemployment rate right now is 48.5% here on Navajo Nation. So that is high, super, super high. And then you have the average household income is right above $8,000, not per month, not by annually, but per year. Just imagine you trying to live on $8,000 for one year with trying to support a family. Very difficult to do, much less trying to get electricity to your house. So it is a struggle here on Navajo. So to help ease some of the struggles and difficulties of these families, we wanna stress energy efficiency, energy sustainability. So why? So you can limit the amount of resources that come out of your pocket. So you buy the better appliance. You buy something that's gonna last you for a long time, although the upfront cost might be high, you make it last a little bit longer and let it ease your wallet just a little bit more. So next slide. So one of the solutions that we came up with was for the remote families that won't ever have grid connection. And those are different reasons that could happen, one of it is terrain. And secondly, is this the feasibility of it economically, it doesn't make sense to run 20 miles of hard line just for one family. If you're putting in $10,000, $20,000, $30,000, maybe even a hundred depending on the terrain, $1,000 just to run for that one customer, that's just not feasible. So what we end up doing is we give the family an option to run their home off of solar with the energy storage bank and give them power. When it was first inception of the program, the main direction of the solar unit was for safety because the amount of distances that you have to travel here on Navajo just to get to a health care facility could be hours, easily hours, especially if you don't have a vehicle. A lot of them you have to leapfrog to call one family that might have a vehicle and then they take you and then their family ride might not be very reliable. So they take you to point B and then you have to go from point B to C all the way until you finally get to the health care facility that you need. And a lot of these families that I have in the remote areas are elderly folks. So just imagine if you have no lights, if you're living off of a kerosene lamp or a propane lamp that you, I don't know if you guys ever had the opportunity to pump that kerosene or that propane and you pump that little thing and you get light. That's a very real situation right now. The kerosene is still used if you go to a gas station you still have kerosene in the gas stations because people still use that. So the main importance of solar is light. We want to get light for these families. So at night, my elderly folks, they won't trip, they won't fall, they won't hurt themselves. And that's the reason that's the main push of their original solar units. As years went on, the solar got more efficient, the batteries got more reliable, the knowledge base grew. So like I was saying, you wouldn't hear anything about solar years down the road, but now you go to community college, now you go to university, now you have degrees and this stuff. So with that new knowledge becomes an advancement in the technology of the systems. So as a result, if you look at the screen now, these are our new systems, our 3000 watt systems. They have the ability to run a water pump to run their sister and sister in their separate heat. They have the ability to run an 18 cubic foot refrigerator and not only that, but you don't have, the reliability of the batteries a little bit better so that the system has three days of autonomy. So they're in our long storms. So even if you don't believe it, Arizona, we experienced all four seasons. Henry can attest to that. He just came to a three day storm out as the adult he'd never knew. And then this was this, but we can have two weeks ago. So we experienced the full gamut. We've experienced a hundred degree weather, then we experienced a minus 10 degree weather. So the systems are definitely tested here in Arizona. So we have our trials and tribulations with them. We're almost like a test bed in certain times. We would definitely run these systems all the way through and we run them to the ground sometimes. And so with the whole point of having to try to get these family reliable electricity. So right now currently within our program, it has 500 on here, but we just added another hundred. So we're about 600 systems that we have out in the field right now for these families. And so with the seven districts across the Navajo Nation with three to four trained personnel, they have the ability to go out and touch these systems if needed in the event of an outage. So they go out and they service them, they maintain them and they keep them running. So three days of autonomy, we treat them like a regular outage. So if we have an issue with the systems, it's not up to the customer to service these things. It's up to us into a, if the battery goes out, if the panel gets damaged, if the inverter goes out, anything with the system into a takes care of that cost. So that's a benefit to the customer. But for the customer, again, when you go back to the average monthly income of $8,000, there's no way, even if they say they're going to replace a battery bank at $7,000. It's going to take them 15, 20 years to even try to get up to that point. So to relieve the stress of that family into a, you know, picks up the cost for that. So it can be substantial. So there's many things that go in place. You got education to the family, you got education to our workers that go online to try to keep the systems up and running and reliable. So it is a big endeavor and a big, job that we do here for MTA for these families as well as the solar units to keep them running. So next slide. And so the off-grid systems themselves, you know, we got 300 of the solar, so these are the different inverter types that we're calling out. So based at 300 watt system, 300 watt or 3,000 watt, I'm sorry, 3,000 watt system, 300 are the solar inverter main system, and then you have the 150 outback radian series systems. So two different systems and they're both providing power to our off-grid systems. And so with the help of Henry, all these systems are tied to a, that's SEMSERA monitoring system. Henry can probably talk about a little bit more of the specifics on the SEMSERA, but all of them, you know, they all company back to one database and Henry has access to that, right Henry? Yeah, that's right. So as Derek said, these are the systems that we're gonna be considering in our case study. In fact, we're looking specifically at 150 of the 300 solar, the 150 that we're looking at were installed a couple of years ago. The additional ones have been installed more recently. Here's just some more pictures. It's a 3.8 KW system. It's bifacial. So the array actually has cells on the back of it to catch some reflected irradiance. And here's the batteries there, 16 of them, they're gel lead acid batteries. Each battery is about 200 amp hours of pop. The system schematic looks something like this. There's two, the array is broken into two sub arrays. We have our battery bank here, a KW inverter. Now, as Derek said, we are collecting lots and lots of data from this gigabytes worth at this point. But the data that we're really gonna focus on for most of this talk is here. It's the data going to the homes. It's, excuse me, it's the power going to the homes. So one of the research questions that we're asking is, well, how much power do people use from these systems? Because really it's important to understand that, to understand should the systems be larger or smaller? How reliable are they, et cetera? So we're really gonna focus on how people use the electricity from these systems. We have an opportunity to look at 150 of them for multiple years at this point. And it's really a study like no other. The amount of data that we're getting is truly amazing. And NTUA had some good force thought into equipping them with these data acquisition systems. So it's an amazing opportunity to learn what's going on. The data is sampled sometimes multiple times a minute, but basically what we do is we sample the power output and we convert it to 10 minute averages and then we do our analysis. And the analyses that I'm gonna talk about basically could consider January 1st to the end of December of 2022. So two year period, some of the analyses go all the way up to March, however. So let's jump right into it then. Let's look at the energy consumption characteristics. Well, one of the first things that you'll note when you start looking through the data is that there's a diversity in the consumption. Not all the homes, although they all have the same system, all are located on the Navajo Nation, they don't all have the same daily pattern. What you see here is just four examples of daily electrical energy use. Some of the homes exhibit wide variety. So they might have days where they're fairly low, some days where they're fairly high consumption, some are fairly consistent, same day-to-day energy use. Some have strong seasonal patterns, some have something happen to the home and their consumption either decreased or maybe it increased. But the overall takeaway here is that there's just not one profile, there's just not one typical consumption characteristic, which means that we have to be a little more nuanced in our approach, right? There's a variety of consumption patterns, not unlike what you would find really anywhere else in the world. So one of the things that we're really interested in is how has consumption changed on the longer term? So what we did is we looked at the total consumption and took the average of it of all the homes per month from inception. So this goes all the way up to about March of this year. And we do notice a couple of attributes here, right? We notice that there is a seasonal profile here that tends to peak in the summer. This is interesting. There may be a couple of reasons why for this. One is that there could be some electricity that's used for cooling. We also think that it could be attributed to work and school patterns, which have a seasonality to it. People, children might be or adults might be home during the summer, more people in the home usually means more electricity uses. It also could be related to the availability of the energy. Sunny are in the summer, so maybe people know that and they can consume more energy. One of the other things that we know is that there's actually from 2021 to 2022, a decrease in energy consumption. This is a little atypical from maybe your experience in sub-Saharan Africa where generally we see growth. There could be some reasons for that. One is that in 2021, the Navajo Nation very much had COVID restrictions applied to it. So people were working from home. They stayed at home a lot more in 2021 than 2022. So as more people left the home in 2022, maybe that had to do with the decrease, maybe that was why the consumption decreased. It's also possible that the components degraded from 2021 to 2022, making less energy available. So we're doing some surveys and we're looking into this in more detail, but right now that's what the data is showing is that there is a decrease. Now, if we drill down instead of looking at overall averages of all the homes, if we look at how much energy each home consumed on a particular day and we plot the histogram of it, we get something that looks like this. So on the AC side, the average consumption was 3.58 kilowatt hours per day. But you can see from this distribution, there's a wide variety of that. Not every home consumed 3.58, right? There was a wide distribution. Some consumed multiple times that, some consumed far less than that. By the way, 3.5 kilowatt hours a day is actually quite low. If you look at, say, New Mexico, their average for grid connected homes, it's more like 20 to 24 kilowatt hours per day. So this is still providing access to electricity, but it's not replicating the grid. And I think that's an important aspect. Now, if we look on the DC side of the system, the DC side of the inverter, the consumption is probably about five kilowatt hours a day. This is based on estimated inverter losses. So that jump from 3.58 to five has to do with the inverters, it's internal fans turning on, it's keeping its own lights on, et cetera. And that is actually more energy, that difference that about 1.5 kilowatt hours is about 15 to 20% of the homes actually consume less than that on the AC side. So it can be significant. And I know as a result of that understanding, NTUA is now installing systems with much smaller inverters to try to reduce those standby losses because eight kilowatts of consumption very rarely ever occurred. So this is what it looks like in the form of a histogram. If we wanted to look at it in a slightly different way, this is looking at the, this is an empirical inverse cumulative distribution function where we can see the percentiles or quantiles and how much they consume. So as an example, if we look at the 50th percentile, that's the median. So half the homes essentially consumed more than 3.14 kilowatt hours a day, half consumed less. On either of the extremes, if we jump up to the 95th percentile, then we get to 6.77 kilowatt hours a day. So 95% of the homes consume no more than this. Understanding this curve is important in terminating what's an appropriate size for your system. Generally speaking, we don't size around the maximum. We would just design around maybe the 95th or 97th percentile because we'll see to meet the needs of everyone, you end up with a very, very large system. Yeah, so here, if you wanna, as a little guide, this is how you would interpret that. The 50th percentile, you draw a line straight up and you would go to the left to see the consumption. Now homes consumed generally had a wide variety in their daily consumption. So even within a home, there was a variety in consumption. Some homes were a little more consistent than others. Some exhibited a wide variety. So if you look at their average value, which is this dash bread line, some homes stayed within a multiple of it. Some, it was three or four or five times their average that they consumed on Sundays. So this could be days where maybe they had company over, days that were, was extremely hot or extremely cold, whatever the reason. And so this also, this wide variety that some of the homes exhibited makes it challenging to design the battery bank to figure out how many days of autonomy is actually provided. And it's important characteristic to look at on a per home basis, right? It's the variation that happened. Now just to provide a specific examples, here are 25 homes and the box plot of their energy use. So what we see here is that little green horizontal line, that's the median daily consumption for that home. And then within the box, it's the 25th and 75th percentile. So it gives you an idea of the range that typically were exhibited by homes. You can see for most homes, the consumption is less than five kilowatt hours a day, although there are some that is a bit higher and some have small ranges, some have large ranges. One of the most important things to note here, if any of you are researchers that are trying to model off-grid consumption is that the distributions aren't normal. They're not Gaussian distributions. So if you make that assumption that the day-to-day consumption follows a normal distribution, that's really not that accurate. These are, they don't follow any parametric distribution that I'm aware of, although maybe a feature research task can be modeling distribution functions to fit these. So they all exhibit sort of a skewness in the positive direction with outliers on the positive side. Very briefly, let me talk about some load profiles. So this is some work that Scott O'Shea, who I think is in the audience, worked on and it looked at consuming, like when that energy is consumed over the course of the day. And generally we see that in the late evening, consumption tends to be quite low and then it rises in the morning. This is probably when people are getting ready for the day. We kind of plateau throughout the day, maybe even have a little dip. This could be refrigerator load or there might be some people home with some appliances on. But then we have a peak in the evening, maybe when people get back from work or whatever they were doing during the day. So it's kind of an evening peaking load and then it decreases. The evening peaking load makes it challenging for solar. We generally wanna have a daily, a daytime peaking load. So it's more coincident with when the sun is producing energy. So this does pose a little bit of a challenge then to designing the systems. We also looked at how it varied from one season to the next. And you can see that it's fairly consistent with most seasons, sort of in the spring and summer months, you have more load that occurs during the day, but more or less you see that two peaks that generally happen. So what does this all tell us in terms of design? Well, we looked at how the PV arrays are sized based upon the load that was actually recorded. So the first thing that we did in this research was to figure out how much energy we could expect these PV panels to produce. And to do that, we used a simple formulation. I don't think I'm gonna go into it in detail, but essentially you can estimate the energy that can be produced by PV array based upon the average insulation of the area, accounting for the tilt and the latitude and longitude and the losses. And we picked 4.5 kilowatt hours per meter squared per day. That's the January average insulation for the homes that have the solar systems. And then from that, we said, well, we need to apply an array to load ratio about 1.3 because you have to oversize to some degree, make up for times when the consumption is a lot higher or the battery charging profile doesn't let you consume as much energy as could be produced. So when we do that, we could take the two equations and we can put them together. And you come up with a result that for every one kilowatt hour of DC side load requires 385 watts of PV capacity. So based on those assumptions for every kilowatt hour of DC side load, that's how much capacity you would need to serve it. And so what we did is we looked at the distribution then of DC side load and figured out how large of a PV array we would have needed. Now this is retro, I mean, we're looking, we have the benefit of having this data. So we're looking back. We didn't have this data, of course, during the design phase, but what it tells us is that because the consumption is fairly low, we could serve about 50% of the homes with a 1.7 KW array. Now remember that the arrays themselves are somewhere around here, 3.8 KW. So in some cases, in fact, many cases, the PV array could be made a lot smaller and still provide the homes, but it wouldn't be able to provide enough energy for all the homes, right? But it does suggest that an approach maybe in the future would be to offer maybe smaller arrays for maybe half the homes and then the other half could get these larger arrays and that could be a way of saving money if interested. A similar approach would be to look at the battery bank sizing where we look at the days of autonomy and we do some calculations based upon that. And I think I'll go through this a little bit fast, but we make some assumptions on how the batteries are gonna behave over time, how deeply they'll be discharged and what their charging efficiency would be and based upon these reasonable assumptions, for every kilowatt hour DC side load, you need about six kilowatt hours of battery size. So we again can look at the actual DC size load, DC side load and see how large of a battery we would need to meet it, to meet three days of autonomy. And what we see is that the existing battery banks provide three days of autonomy for about 75% of the systems. So you might say, well, the target was three days of autonomy, this seems to, you know, you're only doing that for three out of four homes. So maybe the design was off, but to really look at what it would take to provide three days autonomy for all systems, it would require more than doubling the battery bank capacity, which would result in gigantic battery banks. So here we're highlighting the trade-off, right? You know, if you're gonna have one design that you stick with, where on this curve, where on this table do you wanna be? And I would argue that meeting 75% of the homes with three days of autonomy is probably a good trade-off. The more homes that you serve, the larger the battery bank you need to have for each home. So let me just point out a few things before we talk about future work. What we've kind of shown in this design case study of looking, having the benefit of having now historical data shows a few things. First of all, if you didn't have the data, estimating the size of the load and the size of your components is very difficult, especially in contexts where, you know, there's not a lot of literature. Also point out the NTUA, when they rolled these systems out, we're less concerned about capital costs and more concerned about getting them rolled out quickly and having sort of a one-size-fits-all for just efficiency on their end. And then also, as Derek pointed out, the service costs are high. Getting out to some of these areas can take hours. So you'd rather have maybe your PVRA larger than it needed to be if that meant one fewer trip that you had to take a year. So we shouldn't ignore the maintenance and service costs. Now, we're just sort of scratching the surface on the analysis that we can do here. We just look, what I presented was just about electricity use. Muhammad is gonna briefly talk about some of the future analyses that we're looking at. So go ahead, Muhammad. Thank you, Henry. So as if our future analysis will include analyzing battery voltage profiles in order to give us insight about the reliability of the system, as you can see here in the figure, it's provide a typical voltage profile. So later after sunrise, there is enough solar power for the batteries to recharge, which is evident by the battery voltage increasing rapidly. And this is called the bulk charging stage. And once the battery voltage increases to predefined set points, as you can see upstairs called the absorption voltage, it's usually for between 46 to 50 voltage depending on the temperature here, it's above that. Then the charge controller regulates the battery voltage as so it doesn't damage the battery by over voltage. And this is called the absorption stage, which can be short as a few minutes to longer hour, several hours, for example. And when the battery is fully charged, transmission to float stage and the battery is maintained at a lower voltage, as you can see lasting from about 11 a.m. in this graph to 4 p.m. And this is typically, this profile just shows a typical behavior. It doesn't happen always exactly like this one, but this is a typical behavior showing over the year in multiple systems. So we'll be trying to investigate multiple systems, sample from system that we have over the year and see how the behavior is showing in order to give us insight of the ability of the system. This next slide show an example of time series analysis for the battery, this is 10 days, it's showing 10 days in January. You can see most of the days where the absorption stage reached. However, there are some days it didn't reach to the absorption or sometimes it's different times happen. For example, in January 1st, the absorption stage was only reached late in the afternoon. This could mean that the battery was never fully recharged on that day. So we are now investigating if the timing of when the absorption stage is reached and how often it's reached it. And also, which will give us an insight about the ability of the system. Next slide please. This graph shows the histogram for the voltage. As you can see here in the system spin significant amount of the time at around absorption voltage. So between 40 and 56 volt. So this indicates a higher reliability of the system. This graph shows an example of the time at which the absorption stage is first reached for four different units. Some get recharged earlier in the day, which again would suggest that the batteries are being fully recharged. So now we are at this stage of analysis. We will try to get a solicit at what time exactly the battery reached to the recharge to the fully absorption over the year for multiple systems. And we'll compare and see the reliability of the system across the year. So this is a type of analysis that you can do if you have such data for off grid system. And hopefully we'll get more insight from the result we have in this future research. Yeah, thank you Mohamed. So that's the work that Mohamed worked on as a E4C fellow and we're continuing on it. So to conclude this case study we actually looked at over 100 million data points to do this analysis of the consumption, average daily consumption on the AC side was 3.58 kilowatt hours per day, which is far lower than grid connected homes in the area, but also about an order of magnitude higher than what you often see in Sub-Saharan Africa. So it highlights sort of the difference that context makes not all homes that are powered by off grid systems use that electricity in the same way. There's a wide variety of consumption characteristics that we saw. We saw that with the benefit of hindsight probably the PV and inverter sizes could be reduced somewhat and still meet that demand or a segmented approach could be used. The battery bank though seemed to offer the desired days of autonomy for a reasonable number of the homes. As Mohamed talked about for next steps we're gonna look at really the battery bank voltage and what that can tell us will be also looking at load profiles in more detail. We're doing some surveys to figure out and explain some of the characteristics of the consumption that we've seen. We're deploying some data acquisition systems to look at sunlight and to try to figure out how well the PV panels are performing and we'll also be looking at reliability. This work was funded by the National Science Foundation. There is another partner that we have in Navajo Technical University that's been very much involved in this project although none of their faculty are presenting today as well as several other individuals that have contributed to this work in one way or another. If you're interested in this, this is a, this top citation for energy for sustainable development. Actually it was just published so it's not under revision anymore, is available and it contains all the details. So with that, I think we're ready for Q and A. Here are our contact informations and I'll turn it back over to you. Thank you everyone. Thank you, Henry. And a number of questions have already come in. I'm gonna let you keep sharing your screen just in case you wanna reference any slides and answering the questions that have already come in. And I welcome our listeners to go ahead and put questions into the Q and A so we can keep track of them. So a number of questions came in specifically for you Derek related to the NTUA and more of the administration in terms of who finances the NTUA, what the revenue model is. If you can speak to that briefly, please. Okay, basically everything in our area we work off of the Rural Utility Services. So it's a, it's a department that's located under the USDA, so United States Department of Agriculture. So they, luckily for us that we qualify for low-interest long-term loans. And so that gives us the ability to finance a lot of these projects that are happening. Recently with the administration that as it is, with the pandemic that we just went through are going through with COVID, they had released on the CARES funding. So the Affordability Act, I believe was the second one. So it was ARPA. So you got the CARES Act and you got ARPA. So it was a significant amount of money that was channeled through the administration here in the United States. So we got quite a bit of money that would help us with COVID relief. And as a result, we were able to direct that towards a lot of the families and use the funding to purchase a lot of these solar units for, you know, through the CARES Act as well as the ARPA. So which is the American Relocation Act, I believe it is. And I don't, I don't remember what the acronym stood for, but those are two significant, you know, funding vehicles that we're able to use. Fantastic. Thank you so much for providing some of that insight. We also have some very specific system design questions, probably more for Henry and Mohamed. Considering the huge variation in load profiles, have you explored the possibility of designing standalone systems for some households, the outliers in particular instead of fully interconnecting the system? Yeah, I mean, good question. You know, to be clear, all of these systems are off-grid and none of them are interconnected with any other system. So they're all standalone. I think it's really interesting to look at the outliers because those are the homes that are consuming lots of energy, which you usually want, right? If you're gonna install a solar system, you can want people to consume a lot of energy. Unfortunately, just looking at the data alone doesn't tell us what these people are using it on, what, you know, what conditions they're in, what demographics, slice they are. We're gonna be looking into that with some of the surveys that we do. But the very brief exercise that we went through in this webinar about, well, once you have the data, you can use that to figure out how big of a system you would need. It does uncover the fact that that there are some homes that are really out there on the tail. And if you design all the systems to meet the needs of that, that one home that consumes the most, you actually end up with very, very large systems that for quite frankly, would be a waste of money and just sort of impractical to install. So understanding that distribution of consumption, I think is really important if you were to try to have a segmented approach, right? If you didn't wanna do a one size fits all, if you wanted to come up with two options or three options, understanding where on the curve each of those options would serve is quite important. That's excellent. And there's a lot of questions about the batteries. Now, zeroing in on that, what is the lifespan of the batteries? Most especially maintenance, since it's peculiar that most of the areas where these batteries are used are low income areas. And also, as in most cases, these batteries are not produced, for example, in this instance from the question in Sub-Saharan Africa. Derek, do you wanna comment? I can start and then you can follow up on the specific scenery. So the batteries that we contacted or that we have contracted with has a warrantied lifespan of five years. And these are maintenance-free batteries. I believe they're AGM batteries, so augmented glass mat batteries. So with the whole point of the reason why we want them is to kind of ease the amount of time that we have our crews to get on site with them. So before we had these batteries, we had the lead acid batteries, so the flooded lead acid batteries. So for you all that are familiar with these type of batteries, you have to refill them with distilled water. And so that takes a lot of time. Then there has to have an established crew on site and with the reliable vehicle to constantly be out every three months to refill these batteries. So with the added amount of units that we had on, we just couldn't do that. So we made the change for maintenance-free battery, which would limit the amount of time that we go out and service these units. So currently what we have in the systems are the maintenance-free battery, AGM batteries, as well as a set of lithium batteries. So Henry can probably talk a little bit more on the specifics and talk about the report. Yeah, sure. So as Derek said, the batteries that we looked at here are, I think they're gel, not just AGM, gel batteries. There's another set of units that use lithium ion batteries. The ones here are gel. One thing you have to understand with battery life is usually manufacturers do tests in very controlled conditions with constant current and room temperature. And as Derek talked about, there's four seasons where these are exposed to and they're not in any sort of temperature-controlled environment. So they're exposed to the hot temperatures, cold temperatures. I'm sure that affects and will affect their performance and overall lifespan. So the jury is still out on how long they're gonna last and how they're gonna be performing. It's certainly a consideration. If anyone else is thinking about doing a similar project is to think about battery, the battery selection and really understand how batteries are rated and it's usually not for environments that aren't temperature-controlled where they can get snowed on, they can get heated up pretty high temperature as well. Yeah, it looks like there's a lot of questions about just operation of various batteries experiences that people are sharing regarding struggles to recharge the bank sufficiently to maintain health. So stuff will need a common issue here. I am going to, I can tell the audience that we have a lot of questions and I suspect we're not going to be able to answer all of them on go. We, I do wanna see if our speakers will be kind of to perhaps do a follow-up in addressing some of these questions offline, but there are two questions have come in that are related to each other in terms of designing other systems based on these learning. So the specific question here is what tools and resources would you recommend in sizing a system without a data set like this? And are there any metrics that you didn't measure that would have benefited you greatly? Oh, wonderful questions. So, if you don't have the benefit of historical data, really the kind of the state of the art is to do surveys or try to do a bottom-up approach where if you have some control or the number of appliances that are out there, you estimate how many appliances of which power rating there's gonna be, how long they're gonna be used and you build the load profile up like that. There are several universities that have developed tools that will help you come up with a realistic-ish profile, but really having historical data is where it's at. I mean, there are some data sets that are out there. If you look at our paper, we do provide, I think, enough of the statistical characteristics to get you started, but it's extremely hard to do without historical data. If you do do the survey approach where you ask potential users what appliances you might have, how long are you gonna use it, take any response that you get with a grain of salt. We've done other research where we've shown the survey methods to be off by like 300% where people tend to overestimate. In particular, the tariff structure also affects how much people will consume, especially in extremely low-income settings, people might, when you survey them, might say they're gonna use all this electricity, but then they get the bill for it in practice and they say, well, we actually can't afford to use that much. And so you might be extremely disappointed. I will point out that a couple of years ago, I did do a webinar series for Engineering for Change and in one of those modules, we did talk about load estimation and how to incorporate that. The tools that you can use to design the system once you have an idea of the load, I mean, there's many out there, personally, myself and my nonprofit, we rely on Homer, which is a really good tool, again, to get valuable information from Homer, you have to put in good information and that includes the load. I think there was a second part of the question about what data do we wish we had measured. I think having some sunlight data, the irradiance data, pyrinometer installed with some of the solar arrays would be really helpful for letting us understand how the solar arrays are performing. This is something that we're retroactively putting out in some of the units. In fact, Navajo Technical University is working on that design and installation. They're leaving that effort, so we'll be putting some of those out there, but that's probably the one piece that I wish we had. That's excellent. And I think we're taking our last question and this is one that I think there's also a little bit of a recurring theme in the questions around, which was, can you describe the monitoring system a little bit more? How does the data get transmitted to the folks doing the monitoring and there was just generally interest in understanding data flows before the systems? Yeah, so NTUA, when they put out their requests for proposals, specified that the solutions had to have data acquisition capability. So the vendor that they went with uses SAMSARA and SAMSARA is an industrial data acquisition company. I mean, this is what they do. And so they installed it, I believe it's all, they collect data, they sample the data, I believe it's pushed out through the cell network to just the cloud, essentially. And we're able to access it through just an API. So I really just have a pipe on script that we run every so often that downloads the data and it ends up being, there's about 50 fields of data that come from each unit, sometimes multiple times a minute for hundreds of units, it ends up being a tremendous amount of data that needs to be ingested and processed, but I think it's been really a highlight of this work, it's being able to get such high quality data, it's really remarkable and amazing. Cause some of these locations are really, really remote, I mean, really remote. Indeed. Well, we have arrived at the end of our time together. I apologize, we have something like 20 more questions that went unanswered because this topic is obviously of tremendous interest to our audience and relevance globally. So I wanna thank all of our speakers today, Muhammad, Henry, Derek, it's been such a pleasure learning more about this particular case study and hearing the necessary details about how your work has advanced accessibility and how we'll help others design similar systems. So we will try to get to some of the questions that went unanswered. Please do follow us on E4C to hear more about where you can see some of these questions addressed and for snippets of this particular webinar. And we've shared the link to the webinar series that Henry generously provided previously that includes a lot of detailed guidance on how to load estimation amongst other topics. With that, I want to wish all of you a good morning and good evening or good afternoon. We're looking forward to seeing you on another webinar soon. And thank you all so much for your attention, your time and your thoughtful questions today. It's been a fantastic conversation. Enjoy the rest of your day, everyone. Take care.