 Okay, good morning and good afternoon everyone and welcome to the research panel on the energy resilience challenge topic. And this is the third webinar that co-hosted by April and the Stanford bits and what's, and together to discuss the challenges opportunities on the energy system resilience issue and how the AI can help. And a little bit background, as we are speaking the electricity sector is going through Titanic shift due to the urgency to decarbonize improve the resilience against extreme events. Like what we have been in the last several weeks, all the hurricanes in the southeast area, and also provide affordable and equitable access to the at risk communities. And the traditionally modeling simulation post approach has been used widely by utilities to design the infrastructure. Then we think that AI could be a powerful additional approach be used to analyze large amount data and identified patterns and the trends. And by seeing that over the last 18 months, you know, every launched the AI initiative and been working with Stanford and the many other organizations put together a ground challenge of AI application for the electricity sectors. And the way identified five ground challenges. And the one of them is how the AI can help to address energy resilience issue. And also I'd like to introduce my co-host Sarah from April. Sarah and I have been doing this over the summer. And the beginning of the summer June 21. Together, three US utilities from both west from west coast from east coast from Texas, and the shared with us all different kind of climate induced vulnerability assessment down at different utilities. And how they collect the data, and what kind of use case they have been using the AI approach to address the resilience challenges. And a couple months ago, we invited the two European utilities and from from Italy from spender and the Portugal discuss how they are using different approach to address the climate resilience challenges. And today, we are very glad to have both April and the Stanford are going to share with the audience. So existing research effort going on at these two organizations. And without further ado, so let me do this way why not I introduce Alex, then you can introduce my colleague around how's that. Very good. So, we will start with Alex to share with us at Chris climbing the ready initiative. And Alex is CEO chief of staff at electrical power research Institute, where he supports the CEO's enterprise strategy and operations. In addition, over the last six months and a year. Alex is co lead at Chris newest strategical initiative called the climate ready. We are with the officially launched in April 2022 with a goal of developing a comprehensive and holistic investment framework on climate vulnerability analysis. So with that, I will stop sharing my slides and passing back to Alex, Alex. Thank you, I appreciate the opportunity to be here today I appreciate everyone's attention and the opportunity to talk to you guys a little bit about whoever he is and the kind of work we have ongoing. As we said, my name is Alex so me I serve as chief of staff for CEO. I'm based out of Charlotte, North Carolina, and I've been with the effort team for about 11 years. I have a nuclear and materials engineering background. Just a quick aside on who every is and what we do for those of you who may not be as familiar with us. We're a nonprofit independent think tank for the electric power industry. We're conducting research and advancing technology on every aspect of electricity from the generation of electrons to the transmission delivery all the way down to the end use of electricity so any and every asset asset and aspects are a facet of the utility industry we have some sort of ongoing R&D. And ultimately what we're trying to achieve is work collaborative collaboratively to advance and develop the technology and mature to a point where utilities and customers can ultimately apply it in their homes for the benefit of society and for the benefit of the world. We do utilize a collaborative approach. So we bring together stakeholders utilities vendors OEM consultants academia national labs from across not just here in the US but but across the globe with that singular purpose of advancing technology and science. We represent over 400 global utility members and execute roughly 450 million annually and research every year. The most important and key takeaway with regards to whoever he is is that we are purely an objective science and technology based organization. We don't advocate policy we don't we don't recommend policy. And as a result we believe it gives us a unique and really a highly credible position within the industry and within the broader discourse to inform policymaking and to inform dialogue with external stakeholders and ultimately important utility strategy. All that being said, I will speak a little bit today about climate ready for climate resilience and adaptation initiative. And I promise to only use one slide during this and before I even throw up that slide I want to I want to set the stage a little bit on the need the opportunity and the barriers that ultimately led us to launching this initiative and initiating this work. The need which which Lang has alluded to is really in two parts. The first is around extreme weather. We are all having this shared lived experience of extreme weather, not just here in the US but but really around the globe, and we're seeing extreme weather events increase not just in duration intensity but but also most notably frequency. What was once a one in the thousand or one in the one in 100 year event is now becoming a one in one in 20 or one in 10 year event or even more frequent. You can see this also we an organization attracts the number of individual billion dollar weather disasters each year just here in the US. It has massively doubled in the past 10 years and quadrupled in the past 20 so there's this this undeniable lived experience around extreme weather and the consequences of it. In tandem to this is this rapid decarbonization transition we're all undergoing, which is driving and accelerating the growth of electrification. It's on electricity for its final energy source is rapidly growing. It's currently estimated today that roughly 20% of our final energy use is from electricity that's expected to double or even triple by 2050. So we already have this exceptionally high bar for reliability and resilience, and it's only going to grow even higher. And not only do we have to maintain it, but we have to improve upon it as society becomes more and more dependent on electricity. You know, examples of this are stuff you're seeing customers you customers have in their homes electric heat pumps, water feeders cooking appliances. And then of course you can't say electrification without commenting on electric vehicles. Just over in California where where liang is right now they recently passed the law saying by 2035. All new electric all new vehicles have to be electric so it really, I think put the fine point on the need here, both from an extreme weather point of view and then also from society's growing dependence on electricity. So, you know, taking a positive look on this it's not all not all bad or sad is that there's with need comes opportunity opportunity for change opportunity for advancement and development. So our opportunity here is really rooted in science and technology, as I would argue most opportunity is as a as an engineer are climate science or modeling tools are capabilities. They have advanced and mature to a point where we can start having meaningful conversations and projecting out. What does the climate look like in 10 years in 20 years and 30 years. How does that climate, these fundamental variables translate into strict the probability of extreme weather events. And what these tools and what these insights do is they ultimately enable us to take a fundamentally different approach to resilience. This different approach is going from a reactive to a proactive thought process around how we invest in resilience in the electrical grid. So if you think about it historically, some would argue that the utility industry has been gotten quite good and fairly effective that reactive approach to resilience. And I'll give an example when when Hurricane Harvey hit the hit the Texas coastline in the Gulf several years back. It caused significant flooding in a number of areas that really hadn't seen it and the local utility came in and spent half a billion dollars to elevate critical infrastructure to ward off future flooding events in that area. And that's good and that is a success in some regards in the in the fact that we were responsive to the event. But what we also we need to do recognizing that that climate climate change and the these new extreme events are becoming a reality that we're living in. This is a proactive footing towards resilience investment we need to and we can with these new tools and methodologies anticipate these new extreme events and and these untested vulnerabilities within our on our electric grid. So the study I saw that I think really puts it into perspective is that it estimated that transmission and distribution infrastructure owners would save roughly 2.5 billion that's billion with to be annually every year. By the end of the century, simply if they use projected climate data to inform their design standards, instead of historical climate data. If that's not strong enough, quantifiable justification, I'm not sure what would be there's other studies that point to similar quantifiable impacts of us adopting a proactive approach towards resilience investment. Not only does it have this fiscal impact it has a direct impact obviously on the communities and society itself. The consequences of losing electricity as I mentioned earlier is becoming more and more dire more and more severe so there's a there's a quantifiable and then a qualitative reason to be exploring these these options and these solutions. So we've got these clear needs. We have a defined opportunity based on the science and technology capabilities, capabilities open to us. What are the barriers what's holding us back, and this really gets to the crux of why every has launched climate ready is to address these barriers. The first one is, is the approach, the approach right now is the spoken inconsistent. We have universities, national labs consultants are the organizations like every across the US and across the globe, all providing their own unique and custom analysis and methodology towards projection and then downscaling and localizing of climate data and then applying it to individual assets. And it's not to say that any one of these methods is right or wrong it's they all have their pros and cons they all have fundamental value to add to the equation. But we need to set some ground rules we need to provide some some guardrails you might say some consistency around the fundamental assumptions that we applied. So we as a group when we talk about this and we talk about it with external stakeholders, we're all seeing it from the same same sheet of notes so to speak, and we all have a similar understanding base understanding to guide the conversation in the discussion. And with that inconsistent approach, the other key barrier is simply the lack of transparency. Because there's this bespoke and individual approach across the board. There's often a lack of transparency, and so you have utilities and other the operators and owners trying to make informed and justified resilience investment strategies. And just as just as viable and just as justified questions from key stakeholders from regulators from others around, have you benchmark this, what's your maturity right now have you thought about these other tools and these other options that are out there. And so this all this, all this talk and all this, you know, these these needs this opportunity and then these barriers, ultimately leads us to this, this new initiative that we've launched back in April. So I'm sharing my slide hopefully everybody can see it here it's called climate ready ready as an acronym with an eye stands for resilience and adaptation initiative. And what climate ready seeking to do is ultimately provide a framework that points to the landscape of tools and capabilities that is out there and provides a step by step, sort of how to guidance document for our industry to coalesce around which to ultimately assess vulnerability to our assets and to our system and then make informed and most cost effective decisions around the mitigation and adaptation strategies that can be applied. So within this initiative, which, as I think I mentioned kicked off in April this year, we have put the work out into three distinct work streams. These are not happening in serial these are happening in parallel and iterative with the iteratively with each other as they inform one another. The first one is all around the climate data science and modeling itself. It's around identifying what data is out there what's the suitability of that data. If you're looking at wind speeds for your wind turbines turbines will you care about it at 10 or 20 meters not at two or five meters so the applicability of climate data what's the suitability of it and what are the gaps of it. And therefore, your particular region is is a critical part of the equation. Once you've projected out and you've assessed what the quality of data is and what you what you can and cannot infer from it. That's when you jump over to work stream to and you can begin to assess what is the vulnerability and what is the risk on specific assets within your system. Based on that vulnerability and that risk, what are the mitigation options what are they, what is the suite of adaptation strategies that are out there. And so you can lay them out side by side and begin to compare which ones are going to be the most effective for you and most relevant for your situation. And then what are some of the design criteria you should be one should be planning for as they think about their system operating in the weather of the future 2030 40 years from now. All this bubbles up into work stream three where we take this information these knowledge blocks, if you will, and we build out a system level perspective around the vulnerability. And around the adaptation and mitigation solution so that ultimately at the end of the day, a utility is able to provide do a risk informed probabilistic cost benefit analysis on all of the solutions. And come come to some sort of consensus and agreement around what's going to be the best approach, and then and then have a informed and much more efficient conversation with external stakeholders to help bring that adaptation strategy out into the public side. So we, we envision this is a three year initiative we envision it culminating with this framework that I've described. It's not a prescriptive framework it's not picking winners and losers it's pointing to all the options out there and highlighting their capabilities along with their pros and cons. Along the way will be coming out will be providing a number of deep dive technical assessments briefings informing the public discourse in this area, providing a network of pure network for the industry and external stakeholders to engage and talk to each other. And, and the, and then there's three sort of key tenants associated with climate ready that I think is important to highlight here. One is this framework needs to be comprehensive and needs to include all of the different assets types. Generating electricity transmission and delivery of electricity down to end use the electrification I talked about, plus some consideration for equity environmental justice and other community priorities that might be of importance in a particular area so it needs to be comprehensive. In terms of the assets needs to be comprehensive in terms of the weather variables and the weather student the climate variables and the weather events that needs to be considered. And then comprehensive in terms of the utility activities this is looking at planning through operations and maintenance of the electric grid. It needs to be consistent. And I say this also recognizing it needs to be flexible. This is a big problem but it has hyper local solutions everyone situation is unique and different. We need to be able to leverage a framework that consistently brings us to similar solutions or at least has the same assumptions built in so that when we talk and we benchmark and we attempt to do some sort of, you know, pure maturity, modeling, we are, we're doing an assessment of the levels comparison. Last and I would argue, probably the most important aspect of this entire initiative is that it needs to be collaborative. I talked earlier about everyone taking their own approach and the lack of transparency behind these approaches. What the government ready is doing is convening a stake of a what we're calling affinity group members really external stakeholders. This includes regulators, this includes government organizations national labs DOE includes academia this includes consultants OEMs financial and insurance organizations. We want to bring everyone and anyone who has a stake in this in this discussion to the table to provide perspective to point to solutions and methodologies that they're aware of, and to ultimately incorporate their thoughts and ideas into the broader framework. So that at the end of the day, when a utility leverages the framework, there is a fundamental level of kids if not consensus at least agreement on the baseline assumptions that go into the conversation. Well, all that to be to be said, we're very excited about the initiative. We've had significant success success so far in terms of participation and engagement. We have 33 utility members that have signed up for it with many more ongoing conversations. We have dozens of external stakeholders participating in the effort we are having our first in person workshop next week in New Mexico. So, yeah, I think that's, that's laying out probably pause here and open up the questions but I would encourage everyone and anyone the phone that if you have interest and you want to participate in this initiative, please reach out to us and we very much welcome your engagement. Wonderful Alex thank you very much. We have a lot of information here. Then, I think we're a little bit behind the schedule so let's move on I will have a cell to introduce another panelist here. We are honored to have in addition to Alex Professor Ron Roger Roger go Paul with us today. He is an associate professor of civil and environmental engineering at Stanford and the co director of Stanford spits and what's initiative. He directs stand the Stanford sustainable systems lab focused on large scale monitoring data analytics and stochastic control for in strip infrastructure networks. And in particular and what we're focused on here today. We have power networks renewables integration smart distribution systems demands I data analytics R&D. He holds more than 30 patents several best paper, paper awards and has advised for founded various companies along the same lines and power network R&D. Thank you for joining us today Ram, and please go ahead and get started. Thank you so much for inviting me for this session and I just have a few slides to share and go quickly based on an ongoing project we have here at Stanford but illustrates kind of the value of the AI and this conversation about resiliency. So here's the team for a project, which if you look across it covers many different areas and backgrounds and the faculty are listed below here. You know everyone was aware of the Texas power crisis, and it was one of these big climate related grid events that we have had recently. I remember that I really wanted to focus on was the kind of the human impact of this event you know 70 deaths, 5 million people without power, 12 million people with water shortages. And if you look here at the outage map of Houston you can notice that a lot of the areas where power was cut off were of those most vulnerable people, the low and middle income folks that really don't have any alternatives of where to go and so on. In terms of damage, it was a huge issue so one of the questions that we started out here in the campus was, well, in order to help utilities and the rest of the ecosystem to address this. What do we need to do to be prepared for the next time, exactly like Alex was describing. And our kind of approach though since we have the benefit of being in academia was maybe let's start from kind of the bottom up and when you do that, you know, we said well, we know that the energy system will need to be decarbonized. And what we are asking is for it to be resilient to climate change and extreme events. That's great. But the one aspect that is very key is that it needs to serve everyone, and it needs to be aware of how it impacts people and communities. So rather than taking a infrastructure centered approach, how do we take a community centered approach in order to better manage our system resiliency, as well as the carbonization. So typically the way we have designed infrastructure is we look at the aggregate of the demand we predict the demand and when we say we build the sufficient infrastructure capacity to satisfy that. But by looking at the average demand when it comes time to do resiliency and things like that we don't really know how it is individually impacting different people in different communities. And in consequently, you know, the kind of human impact and costs that we saw before. So how do we start moving towards a design that's much more people in community center. To do that, we first need to establish metrics, and we thought of four very high level areas where we need metrics so the first you know they're shown here. First, of course, we want the system to be equitable, meaning, if you're going to have any sort of resiliency to climb it in the system, it should be distributed equally among all everyone. Second, it should be inexpensive, but we need to also define what inexpensive and equitable mean. You know some more recent definitions of grid costs are urging for us to consider costs as a percentage of people's income. That's not how traditionally we have looked at that. So we want the grid to be resilient. As Alex pointed out, particularly to these new classes of events, the climate induced grid events. And finally, we want it to be 24 seven carbon free. So we started from this big picture, you know, vision and said okay what is needed if we were to try to do this in practice. In order to design such a grid, you need to take certain steps. The first one is, if you want to build that resiliency, we need to be able to map the climate risks at the level of communities. Not at the aggregate level, not that the level of infrastructure alone, but also at the level of communities. And how do we do that in a scalable way. So there's a lot of work in the climate science community, but typically the information they produce is not necessarily the information you need when you're calculating kind of impacts on people through, you know, other infrastructure, you know, be it electricity or gas. The second thing you need to do is then take these inputs of these climate risks and transform them into climate, in first induced infrastructure vulnerability on a community basis so I can take a census tract in California and say things like you know this is how it's like you are to have an outage during you know these type of situations. This is how much extra demand you will have when when you see a heat wave, and so on so forth. So, you know doing that connection at a very granular level is very important and how to do that. And finally, in order to kind of mitigate the impacts of these climate events. So based on the idea of adaptation or preparation. Of course, there is a lot of other steps, even during the event while it's happening and then opposed event and so on so forth. But the question we had is, what are approaches for a more equitable infrastructure adaptation, because typically as we look at the infrastructure adaptation in the grid. And it's focused again on average performance metrics across the whole footprint of the grid. And behind that you can actually hide a lot of these nuances of local populations their needs and the impacts they might have. And we also made it a priority that anything we do in this kind of project we have out here is shared in the open through the energy data commons. And we are also partnering with various organizations including every to do this. So to test these steps here and actually come up with some concrete ideas. We're going to look initially on extreme temperature events because those are the best understood ones, according to climate modeling. As well as wildfires. And I'll just show you some motifs that came out of our research and then we can have a discussion. I'm just going to click through these so you can see the slides later. And then is we tried to map kind of what is the risk of having a very hot or extremely hot day, according to temperature humidity happening in any given location in a map in California, for example. There's two ways to do this. One is you do take station measurements information and do some sort of local interpolation based on the historical patterns and try to say what's that temperature. One is through the reanalysis of satellite extrapolated data so you use weather models and so on. They are at a much coarser spatial level, and then to make it granular, you use reanalysis to do that. There's a very nice report listed in the bottom of the slide here on the that shows that if I look at, you know, a certain percentile of the temperature on a particular location. There's a difference of 30 to 60 degrees centigrade depending on which percentile 95th or 99 percentile, depending on the method you used. So what should you do. So we started with that question. You know, we tried to understand, okay, maybe the right thing to do with risks is to kind of fit tail probabilities. The difficulty of hitting tail probabilities is you don't have a lot of extreme data points. So can you combine the principles of some climate science and statistics in order to do that. And one methodology is to use clustering and be identified these bands where weather behaves similarly throughout history. And now I take the data across those bands and build tail probabilities for that. And that allows me to calculate these, you know, risks in a stationary hundred year level event for a period in the future. Now, all of all of this, you know, shows the importance of knowing what data sources you have and then producing predictions that are appropriate for the application you want to have. And the benefits of combining data with some sort of weather models. We also have been using AI to try to do temperature distributions that are hourly so that you get the actual temperature during the day which is very important to analyze anything to do with electricity because demand and planning depend on hourly demand. Sorry, planning depends on the hourly demand. And our conclusion is that you can build methods of this type. The next thing I wanted to skip to is now I want to take this extreme weather events and then project them into impacts on the infrastructure. The last of things that that one can develop is, for example, how does the actual electricity consumption response to changes in temperature, and then project that with respect to climate, like Alex was saying, and the trick to do that is really to understand your past very well and build a model of your past and then take a moment where you have a change and say I'm going to use the model of my past to say that's how I should have behaved. There's a difference with respect to that. And when we do this kind of analysis one of the things that we learned is that identifying the right metrics for comparisons is very important. So in a particular demand study we looked at, for example, how much extra CO2 emissions will be caused by the growth in demand due to climate risks and climate change. And we also have looked at, you know, during heat waves, for example, how much additional demand is required in different counties or zip codes in California for residential consumers and for small and medium businesses. And it's, it can be quite a lot. I mean, between 20 and 40%. It also leads to a lot of adoption of air conditioning. And again, this increase is not really equitable so whenever we do these analysis we also looked at the distribution of the impacts across the population. And the last thing that we are doing is we're kind of collecting a lot of data from utilities, both taking existing data that's reported and collected by the federal government and then combining with data that is provided to us when we do utilities so that we can build an outage model that would allow us to calculate the risk of outage due to climate events in any location in the United States. And when I say a location, we are thinking about census tracts eventually. And now we have collected data from about 102 electric utilities in almost all states and covering almost all counties in the US. We now correlate that information with the extreme weather database from the US to shell this data. And another important insight about applying AI and using this information, the electric utilities report together with their outage information, which type of event caused that outage so it could be, you know, an extreme, you know, hurricane or something like that. What we learned was, there is also what climate scientists consider extreme events, and that's actually a larger set than what utilities consider to be extreme weather events. So taking the climate science definition and applying it here, you can start to build these models that explain reliability, and it's very, very interesting. This is still ongoing work, but some early insights are that of course investments on the distribution network pay off, and they certainly reduce your safety and safety due to extreme weather events. But for example, if you have a lot of wind and solar in your footprint, that actually makes you a little bit more sensitive to, to extreme weather events, and you have to think about how to contract that. But hopefully we will be able to do more of these models, particularly partnering with many who are attending the webinar today. And the last thing I wanted to share was about adaptation. And the first thing we did in adaptation was, we took one particular case study we wanted to understand how can we adapt the urban grids to with respect to wildfires. And the challenge in doing that is like we wanted to know, okay, what is it beneficial to do should I do vegetation management. Should I replace wooden poles by metal poles, should I underground the lines. Well the challenges, a lot of this information is not publicly available. Particularly for example in California we don't know the under grounding status of the distribution lines that's data that is owned by the utility. What we did was we used Google Street View information and we use computer vision to kind of map which lines are overground which lines are underground we correlated that with the existing publicly available maps for the distribution grids in California for SC and PG&E. And then we combined the, so that allowed us to calculate an under grounding rate per census tract in these, you know, Northern California and Southern California census tracts. And then we combined that information with other sources of information. The fire publishes a fire risk map, which is their annual wildfire probability prediction over a certain period of time. So we can compare that to where the, you know, the under grounding rates that you obtain. So, undergrounding rate means, what is the fraction of miles of your distribution lines that are underground. We use three canopy cover data that's published as well, and we can then compute the canopy cover and how close it is to the distribution line. So that gives you an idea high level idea of vegetation risk. And then based on a particular data set provided by PG&E last year to the public, we were able to also look at the asset risks and polls and transformers and so on. So let me show you what this uncovered, which is really interesting. The first thing is that vulnerability of the grid is not equitable. So if you look at under grounding rates, that's the first figure here, and I stratified according to wildfire risk. You can see that undergrounding is always lower for lower income community census tracts. So the fraction of lines underground. And particularly if I look at high wildfire risk, and that is definitely true. Now this is obviously not meant to be done in purpose, but this is the result of us using tools for planning that look at things in the aggregate. The second thing that we learned is for example if you look at the fraction of overhead lines that overlap with the three cover within 10 meters. You can see that again, it is much higher for low and middle income communities. And similarly if I look at transformers and their ages, you know the transformers tend to be much older in low and middle income community regions, and finally the same for wood poles. So what we saw was, well there is this issue happening of the more exposure to risk for low and middle income communities. And so how do you go about fixing it? So we worked with a policy expert here on campus, Michael Wara, he also writes a lot about wildfires. And the idea that we had is really you know the most effective, one of the most effective mechanisms is line undergrounding. Of course you can't underground all the lines. So we use an algorithm to determine where are the critical lines that need to be underground. So if we have that then we have to figure out how to make the costs of this more equitable. And we use a definition from equity that has been proposed in the last five years by Severin Bornstein from Berkeley that says well we should start looking at these costs as a fraction of income. So we wanted to make sure that the increase of what you pay for electricity as a fraction of your income remains the same no matter where you are. And the method then it's pretty simple you look at the sense of strats where you have to bury lines. If you are below a certain income cut off, you're going to use a general cost sharing mechanism. If you are above that cut off then just the local community shares the cost for that upgrade, and that basically equates and makes it much more flat, you know the percentage increase of everyone's bill. And the next thing that we're trying to do now so so this shows you know the way we think about the solutions need to change if you're going to incorporate the impact on the population and this issue of equity. The next thing that we are doing is also looking at DRs. So that's kind of our labs main focus along the years have been distributed energy resources and how to integrate them to the grid and provide support in various ways. And the idea that that we are looking at now is, if I want to have resiliency using DRs what should I do. And so we set we took data that's available for basically across the US, you know, both residential and commercial data. And we were able to take utility rates in all these different regions and different costs of installation and everything and optimize a mix of diesel batteries storage and PV. So this is preliminary work we're also incorporating now a natural gas genset and see what is the optimal mix that a customer would have that's lowest cost, given that he has the goal to have a certain reliability level, a certain likelihood of outage, given the local expected outage rates based on our previous model that I just discussed. And one of the key findings is that, of course, with the cost of what technologies are today, you use a lot of diesel. But there is regions where you have a lot of adoption of solar and battery. And if you go out, I think 10% of the census tracts where you see a substantial number of consumers, basically, having more than 80% of their consumption being self produced. So once this model using all the NREL projections for costs of different technologies, what we found is that actually can grow and about 30% of all of these census tracts will have that so just a different view to resiliency and more bottom up. So, thank you very much, and we'd love to. Very good, then let's bring both Sarah and Alex back to the stage. For all the audience online, I would encourage you to type your question on the Q&A. I do see two excellent questions there, and we're going to address them later. I think both Ram and Alex, you mentioned Ram, you talk about a lot on the community resilience solution, and Alex, you touch a little bit on the environmental justice and equity will be very important part for AAPRIS design of this investment framework. So the question is how can AI, you know, data analytics approach to really bring this EJ and energy equity issue from just a concept to reality because from my perspective, I'm relatively new to the EJ and energy equity area. I'm learning a lot recently by talking to my colleagues here. And also, you know, from Ram's example is there's a lot of public available social behavior, social science data, like the income and the race, you know, from the census data. How this data can be leveraged to help you to make the investment decision. So probably, I think, let's, Alex, you go first, then Ram can chime in later. How's that, Alex? Great question. And I'll confess, I'm in the same boat as you. I'm learning a lot about EJ and equity from my friends and peers here at FRA. It's a fascinating topic, and it's especially for engineers because it's so qualitative and kind of squishy and subjective sometimes. That's not to say that it isn't just as incredibly important, but it's not as black and white as we might be used to. I think my initial thoughts here would be around the social science data itself and how AI can fill in some of the gaps and provide more granularity and depth around that data, particularly from a locational point of view. Ram touched on in the beginning of his presentation around the number of vulnerable and disadvantaged communities that were impacted by power outages or power disruptions. And so being able to understand where these communities are can directly translate into the types of resilient solutions that are going to be most cost effective and useful for them, whether that is undergrounding, whether that's micro grid, whether that's putting in more feeders or changing the type of distribution pool that you're using or simply vegetation management. All these different solutions can be viable, and each one is going to be different depending on the unique characteristics of their location. But I think it's important for us to inject within the conversation this fundamental understanding. And, Rob, you related, I think, highlighted well that the impact to a disadvantaged community with a power disruption is massively more significant than to an effluent community. I would be so bold as to say most of the people on the phone here today can lose power for a day or two. We could lose all the food in our fridge, and we could, you know, jump in our Range Rover and drive 300 miles to go stay in a hotel where they do have power. It would be inconvenient. We wouldn't like it, but we would be fine. We would be able to move on and continue on with with our daily lives. That's not the case for the disadvantaged communities losing a fridge full of food for them can have a significant impact on their bottom line. They can't just pack up everybody in their family and go to afford to go to a hotel. So the extent to which AI can can help us leverage the social science data from a locational point of view and inform our understanding of the location of these vulnerable communities. And then also help us prioritize some of the response strategies when there is an outage. I think that's another important part of this too. We're not going to be fail safe necessarily all the time but but say fail is maybe a better way of thinking about it. Sometimes it's going to be more cost effective and more appropriate to recognize there will be failures on the grid. And we're going to design a response around those or anticipate them in a way that allows us to bring power back in a faster and more effective way so you know traditionally we look at feeders and prioritizing feeders back just a number of customers but with this sort of data we can instead overlay the types of customers and the severity of impact to them on which which feeders we necessarily you know prioritize and coming back from some sort of outage. I agree with what Alex said and maybe just summarizing and emphasizing a couple of his points. The first thing is you need to define which metrics you want to compute that capture these notions of equity and environmental justice and impact. When you're defining these things you know part of it is being able to compute them at scale across your entire footprint. And part is, you know, the, they need to be metrics that you can eventually actually, you know, take actions upon them, and meaning you can incorporate them in some planning and optimization. And that's another place where I think AI in the near future will be very helpful as we start to optimize with the goal of incorporating these metrics as an objective function. Another aspect that I think is very important when dealing with this issue of social justice and equity and so on is the ability to access data openly. And part of understanding equity is you need real the real granularity on on what's happening. And that means you need data. So that's been one of the challenges we've faced when trying to do that. And I think this data is not just sitting at any one entity so there will have to be a lot of coordination across many entities in the grid to. And outside of the grid as well, you know, building information and cities and all of that to provide the information. Great. Thank you for those responses. And we're going to move on to a second question. When we think about the challenges of integrating AI into systems and businesses. What are the key challenges that you both see in integrating AI into your practices. We may start with with profession with Professor Roger go Paul, while you're up on. Sure. The key challenges that we see in using AI, I think is the availability of high quality labeled data. So sometimes you have no data available, then sometimes you get some data, but the data doesn't have the appropriate information you need to build your tools and we spend a lot of time basically doing that, you know, adding that to the data. Just so you have an idea and all of these projects I shared today, maybe 90% of the time was spent on kind of finding organizing labeling the data. That is the biggest challenge. And, and I think a second but smaller challenge has been around the access to the, you know, different pieces of information that are disparate and putting them all together. Excellent. I was just saying I agree wholeheartedly with with ROMs points there and I might build a little bit off of your point around the data format itself. One of the things that we're going to be trying to do within the climate ready framework is take the downstream needs and data requirements and provide those to the data generators themselves. There's going to be a slow and painful and there's going to be iteration involved inevitable iteration on the data requirements and getting them to the place we need but we can at least short circuit some of it by taking a bit of more of a proactive approach and identifying and understanding better how the data is being used and what the data needs are, and then informing the data generators on how to implement those. And so that's that's something we'll be doing within the climate ready framework itself that will be will be trying to advance that conversation. The other aspect you mentioned to around data access I would build upon that in terms of data privacy and data security, owning the right to use the data and then aggregating it in a way that maintains and an enmity is not trivial. And then throw on top of that this massive requirements growing around cyber security. I don't know how much you guys are seeing it in your world. But in our world almost every facet of our business now is having its own individualized and unique requirements for cyber security that are becoming really quite paralyzing and burdensome. And that's not to trivialize their necessity they're absolutely it's actually very much required but seeing us as an industry in a society rally around some sort of standardization for cyber security data requirements would I think really open up the floodgates and enable us to be much more efficient. Excellent. Yang I'm going to pass through one of the questions that we had come through chat. I had a link shared with us by Craig Lewis, an article about a market mechanism for financing community micro grids, namely a resilience energy subscription, RS straightforward approach. And I think this goes to consistency of approaches ability to provide transparency defensible decisions more easily defensible. The question from Craig is why not simply prioritize resilience based on the type of facility with consistent application across all communities. And Alex if you care to provide any comments. Yeah, it's interesting article Craig and I did note that and I'm going to read into it in depth, a little bit more to better understand the premise there. I guess my immediate reaction is it's, you know micro grids not always going to be the most cost effective or, you know, really reasonable solution for for built for providing resilience and certainly one of the options. But there might be simpler and more cost effective options out there so I would want to better understand the context of when this would maybe this is after you've already identified that my group is definitely the best solution. And you can provide some sort of community financing which sounds like a really interesting idea. I appreciate you sharing the article and we'll be looking at it in the context of our climate ready effort. I think that when most of the analysis I've seen and even the ones we're doing out here at Stanford, they do take into account the load prioritization and the different levels of importance and different types of loads. In terms of solutions. I do agree there is a diversity of solutions, but one of the things that will impact which ones get adopted is going to be the speed and the scale at which we can deploy things. And for example what we see in California is I talked to several friends that have been subject to outage and if they have enough wealth, they go and buy a natural gas little jet and sat and they have their solution right there. It doesn't really matter it's way more expensive than the grid power for them. So, I think how do you incorporate these kind of concepts in your picture in your planning picture. I think this is a new type of challenge for power systems, and it's super exciting to to to to figure this out, you know bring ideas from from public policy sociology and so on so forth into this mix here. Very good. We only have two minutes left and have a still have a couple questions in the kill. So, let me pick one of them. I think it's an excellent question. This is from Lawrence Garwin. And it's a long question. Let me go to the point specifically so talk about by direction EV charging so are you are you like a electrical public utility commission and the utility effect in the imminent direction EV charging into their infrastructure planning. And with the background on the electrification transportation side electrification building side know when to charge the car. And I also want to add whether or not the EV itself can be utilized as a reasoning solution for the homeowner. I will open the floor anyway, anyone want to go first run Alex. I was going to say that Alex is, is, you know, working with all the utilities at every so I just thought you could say something. I could take a step. Yeah, it's, there's certainly awareness, and there's a lot of discussion around this, this opportunity here, it is a, it is not trivial. The market design required the technology required just in terms of the communication with the grid. It's, it's going to be it's a big problem. And it's, I'll just say this, I don't want to be too, too negative, there's awareness of it. And we are, there's not just awareness but there's a sense of urgency, because it's obvious that adoption of electric vehicles is only going to be exponentially increasing. It's clear that there's some type of opportunity here, how we execute on it and how we get over some of the hurdles with the different manufacturers and they all have their own unique approach they want to utilize. Again, they can announcement recently that they wanted to challenge Tesla and, and, you know, providing energy store energy back to the grid and in some regards, it's, it's, it's not going to be easy. I think one of the areas where we have started to play around and do some demonstration, where it can potentially be the first first area we kind of adopt this sort of thinking would be around fleets and electric buses. For example, where they have these particularly large energy storage capacities, and there's these highly defined times when the buses are being used and not being used. So those might be some of the earlier adopters or at least testers of this type of philosophy. Yeah, I, I just wanted to add to what Alex said I think when you are considering particularly vehicle to grid. One has to remember that these technologies are quite expensive. Many of the homes require additional things like you know panel replacements and so on to support this so there will be issues around the equity of an availability of that because the person has to have an electric vehicle and infrastructure due to connect back to the grid. In terms of the one way charging and kind of alignment of solar, which is kind of was also commented on on that the question. And Liang and I just had a study that came out that shows that definitely, you know, putting emphasis on daytime charging is a good thing. And it's a very simple way to kind of increase the lower the cost of adaptation on the grid. And I think one question that's left open is does this also increase the resiliency intuitively the response seems to be yes but that's a study that we will have to carry out to to see if it's true. Excellent. And really great way to end this webcast because if we talk about the most challenging of the biggest challenges, just all kinds of great interesting stuff for us to be digging into. Great places to start. So want to thank everybody who was able to join us today for the webcast. And in particular thank you to Alex and Ron for sharing with us your experience and also closing out our webcast for the year. And to send a follow up email that gathers all of the diverse and distributed resources from our webcast to share those. And we thank you very much again, and look forward to seeing you as we move this work forward. Thank you. Thank you all. Thank you very much. Thank you.