 So welcome to this session. I'm Richard Sassoon. I'm the executive director of the Strategic Energy Alliance and welcome to the session on buildings and facilities Before we get started, let me just say how delighted I am to see you all here. Jenny Milne and I Have organized the agenda for this research showcase and we really appreciate the interest and participation By so many of you who support and do so much hard work To make this reach energy research endeavor at Stanford as so effective So in this session we're looking at the energy and sustainability issues associated with facilities and What better place to investigate this than the Stanford University itself? It represents a perfect living laboratory for studying this this subject and Today we're very lucky to have two of probably the best people on campus who can address this subject first up we'll have Lincoln Gleabans. He's the executive director for sustainability and energy management at Stanford he leads the university's efforts in sustainability and resilience utilities and infrastructure and he will tell us about the Enormous efforts and investments that Stanford has made in decarbonizing its energy systems and operations after Lincoln Will have Jacques de Chalander He's an adjunct professor in the energy science and engineering department at Stanford And he'll broaden things out and describe his research on developing and applying computational tools for flexible Urban energy systems using Stanford as a test bed so I'll turn it over to Lincoln We'll he'll speak and then we'll have time for a few questions for Lincoln And then we'll do the same with Jacques and then if there's any time at the end we can bring everything together So they can I'm telling it over to you. Thank you Morning everyone it is a tremendous honor to be here and telling the Stanford story Which I absolutely love to tell what we've done here is pretty astounding What I'd like to do is give you an overview of the supply side the macro side of the Stanford energy system And then Jacques can do Get you through the really what I think is they're truly interesting stuff now Which is the more micro inside the buildings so I have the coolest job in the world just to give you a sense of this What my group does covers the entirety of the Stanford Ecosystem both internally and externally and we get to make sustainability real we get to make resilience real Across the utilities and the essential services the beating heart really of Stanford University We do that in the context of some very Significant goals. We're looking at zeroing out scopes one and two for greenhouse gas by 2030 and Zeroing out and this is the mother of all goals Scopes one two and three by 2050 The whole world is figuring that out as we go We're doing great work here, but this is truly a global effort to figure out both scopes one two and three by 2050 So when we talk about Stanford energy system innovations, which is the umbrella catch-all title for what we're what we've done I want to emphasize the word system It's very easy if you've been over to the central energy facility on the west side of campus to think of that as The big innovation But the truth is it's an innovation both at that level and an industrial facility But also a system that provides the the essential electricity and thermal energy services to the campus so when we looked at our end-of-life Co-generation plant about 15 years ago Cast your mind back about 15 years natural gas-fired co-gen Which is what we had was actually a very reasonable solution in The policy environment of the day it would have been very easy for us. I think very defensible for us To simply say you know what we got an old co-gen. Let's build a new more efficient co-gen Natural gas is better than coal etc. And so on and off we go. We didn't do that We didn't lean back on old solutions. We leaned in to new solutions We challenged the most fundamental assumptions That we had do we need steam? For example, do we need fossil fuels for example and the fundamental innovations that we came up with for the system None of these are rocket science But put together They're transformative going from a retail posture with the California grid to a wholesale posture That in turn allowed us to enter directly into solar PPAs across, California to transition from fossil to renewable Taking a whole lot of waste heat that was otherwise being evacuated the atmosphere and turning that into fuel Through a heat recovery chiller Moving from steam to hot water moving from a very high temperature energy intensive Product to a lower temperature much less energy intensive product to meet our needs We didn't we thought everybody thought you needed steam that's been the answer for what 250 years But the turn the the reality was we challenged that assumption. We didn't need steam and Then finally energy storage we went from basically zero energy storage To both thermal energy storage here on campus as well as electric storage at one of our solar farms So again, none of these things are rocket science, but put together. They're transformative Here's the centerpiece This is the central energy facility and if you haven't toured it I would encourage you to reach out to me I would love to take you through it. It is the most beautiful industrial facility. I've ever been in also really really cool And and you can really see the innovation, but this is on the west side of the campus We also have a smaller version on our campus at Redwood City about five percent the size of this facility And basically what this is is thermal energy storage Electrically driven chillers heat recovery chillers Creating hot and cold water that go out into almost 30 miles of pipe Throughout the campus to heat and cool the buildings We entered into our first of two solar power purchase agreements back in 2016 This covered about 66 percent about two-thirds of our campus electricity consumption we're offsetting fossil consumption with renewables and Then just last year Our second project went online. This is near Lamar, California about three hours away And this is it's a we have part of a project called slate Which is a Goldman Sachs project ultimately and it is currently the biggest solar plus battery storage Project in the state of California So we have a share of the solar panels We also have a big share of the batteries and that takes us to about a hundred and twenty percent offset of Our electricity consumption on campus both electricity to keep the lights on and do things that electricity does But also to produce thermal energy This is my favorite graph in the entire world. This is our carbon journey and you can see I call it the carbon cliff You can see what business as usual would have looked like that's that black line going off into the distance But you can see the reality too and that is when that when the Thermal plant went online when that first solar project came online We fell off a cliff in terms of carbon production huge win We fell off another cliff when the second project came online We still have a way to go. We still have about 18 percent From our peak that we need to reduce by 2030 huge challenge because it's not big steps like this It's tiny little steps You know our larger energy portfolio, but an incredible outcome from a greenhouse gas perspective But not just greenhouse gas criteria pollutants Which drive energy justice environmental justice outcomes and and our ability to breathe the air our water consumption? Down millions of gallons a day our end user costs Down I I'm I probably shouldn't tell you the number out loud, but significantly reducing our end user costs one of those is The result of distribution efficiency and that should be going up. It should be distribution losses going down I apologize for that, but when we had steam running through the campus We were losing about a third of it a third in distribution Now instead of 30 some odd percent losses. We've got like three percent losses Imagine the economic and environmental impact of an order of magnitude reduction in losses and finally service interruptions a Cogen plant any sort of rotating equipment like that combustion equipment You probably max out in the low 90 percent We're just not seeing that anymore. We're seeing absent last June's power outage, which I'll talk about in a second We were seeing five six nines reliability Instead of 90 percent all energy is compromised All energy is compromised and this what we've done has a significant number of them a couple of that I want to point out one is Significant initial capital costs if you spent any time with green infrastructure You know that the initial capital costs are significantly higher Than traditional legacy fossil infrastructure that said the operating costs and the environmental benefits are Much much better going forward the question that I'm wrestling with is how do we get not just Stanford University? But everybody else in the world over that green monster over that green cliff of Capital so that they can enjoy the benefits The other thing that we've the other big Compromise that we made and this came true in June Was that we are entirely dependent on the California grid? I mean we have emergency generators But from a bulk power supply perspective we're entirely dependent on the California grid and That is obviously Not perfect as we saw last June in particular But you know a lot of benefits Compromises my team is currently working through the compromises now and trying to mitigate those As I said, we've got about 18% of The way to go on our journey to net zero scopes one and two a Lot of smaller stuff you see the big drop We had with the system the energy system coming online and a lot of cats and dogs that we need to figure out As we get to a hundred percent And that includes scope three emissions too Like I said, the entire world is figuring out scope three in real time We have a very aggressive effort here at Stanford to try to both understand it and make it happen but I was talking to some Business school students yesterday and they said what should I you know? What should I focus on? I said if you can figure out scope three you've got it the best, you know You've not only done well For yourself, but you've done great for the world This is a massive challenge for our planet. One of the things that we found in the pursuit of all of this is that technology is The glamorous thing and this is this is my like liberal arts major Formula up here. So take it with a grain of salt, but illustrative the technology the ones in the zeros It's glamorous. It's fun people love having their picture taken next to it, but without behavior change You don't get the outcomes that you want and without project management and change management None of those have things happen in the first place So one of the things that I've been telling a lot of students is if you get out of here without a project management course Or change management course you've done yourself a disservice, but this is how We've I've found we can solve for climate positive impact. It's not just the technology It's got to be the change management the project management, and it's got to be behavior change People love technology people hate to change their behaviors But that's what we got to do going forward Next steps we are in the process of a comprehensive update to our 2015 climate action plan Chasing near-term the 2030 goal longer term the 2050 goal and I'm happy to provide these slides, but this is a good summary of Everything that we've done and everything that we're planning on because we are committed We're gonna do this a Lot of it. I don't know how we're gonna do it, but we're gonna do it and we are gonna make a The right optimized solution for Stanford University But by doing so we're also going to provide an exemplar for the rest of the world in the same way that we have With the work that we've done so far So with that I'm gonna hand it over to Jacques and he's gonna tell you the cool stuff that we're doing on the other Side of the value chain in the buildings Yeah, let's thank Lincoln and Are there any questions but Lincoln? Yep, there's one at the back over there Maxine, there's one over at the back. Hi, I'm a whole a is an alum from CS at Stanford. Um, I This is wonderful progress that Stanford's facilities have accomplished I'm wondering to what degree should Stanford feel responsible for the houses that its employees and students Live in their private homes, you know, what fraction of those have been decarbonized as probably in the single digits Whereas you are up in the 80s now. This can you talk a little bit on that perspective? Probably not as much as you'd like me to I Honestly, I've I've been here about two years in change. I Probably I know about as much as you do about that topic. I'm not being I'm not Running away from the question. I just don't know the answer A quick question Peter Rumsey adjunct lecturer I'm super impressed and and universities everywhere in America are Envious of this project on the central utility system. I guess there's a lot of us who Who are wondering what happened with the the cooling outages and maybe you could just address that a little bit Sure Before my time, but I know the story When we first planned this system in the central energy facility in particular, this was 2010 2011 that planning which was incredibly rigorous had assumptions for campus growth and Assumptions for what I call the climate curve how quickly is it going to get hot? and how hot Both of those assumptions turned out to be too low and so instead of running into capacity constraints on peak days in 10 15 20 years we ran into them in a couple of years. There's also a decision made at the time that We're not going to build for the absolute peak day we're going to build for the Kind of normal peak day and so when the growth happened and the climate curve happened much faster than we thought All of a sudden we were in curtailment the system worked fine But the system just wasn't big enough To handle that load and so what we've done in the meantime, and we just finished this a few months ago We've actually doubled the cooling capacity Within that building we have doubled the cooling that the ability to produce chilled water and that's gonna that's gonna keep us good for a long time But I can tell you especially me My blood pressure took a step change down Now that we're getting into hot weather and I've got double the cooling capacity at the CEF So it was I think it was a it was a policy decision around Building for the absolute peak, but then it was planning assumptions that seemed very reasonable at the time and turned out to be not conservative enough One more question over here, and then we'll move on to show Thank you. I'm Andrea Krause from PG knee and Really impressive stuff I really liked your formula at the end that you showed and I was curious whether you could share one example of maybe a less Shiny exciting Change management accomplishment that was key to what you walked us through from a technology perspective I'll give you actually the least glamorous example, and that's trash My group is doing waste management. We run waste management for the University and we have a zero waste by 2030 goal We also have a contract with our vendor that goes back to the invention of sliced bread Sliced bread sliced bread was invented in the early 30s. This contract goes back before that We literally we haven't re-bid it What we're doing is taking That same sort of approach How do we reorganize ourselves from a technology from an infrastructure perspective? whether it's different bins or and different trucks and different routes and data collection both for operational improvement, but also compliance But then how do we also transition? our campus and every all these individual human beings on our campus from a Oh, I just throw it here. It's right next to my desk. I throw everything in there to Frankly more what they do at home, which is oh, there's this bit and there's that bit That's the situation where there's a big technology improvement big infrastructure improvement but the key to the program is the behavior change and Understanding change management. We I just hired a PhD in in change management basically for my group to focus on things like this but then to project manage that I mean we're talking about hundreds of buildings and really I mean Really stuff that's really personal to people that really I mean it affects their immediate work environment and so we're bring trying to bring all that together not just on the glamorous energy side, but on the perhaps less glamorous waste side and Frankly everything in between Okay, thank you very much. Thank you and I think we'll now pass it on to check Richard so I'm going to talk about Computational tools and experiments for flexible urban energy systems and the research I'm about to Discuss is made possible through support from Sanford land buildings in real estate actually and that energies and This work perfect. Okay, so decarbonizing electricity and electrifying heating Cooling and lanterns today and light transportation Sorry, we'll get us about halfway there to net zero in the United States That means that the electric grid is changing very fast both on the supply side and on The demand side this map summarizes some recent research to track emissions through the US electricity system from Generation to consumption while accounting for electricity exchanges One one big reason I'm putting this up here is to make the point that the grid is one giant interconnecting machine And so what we do on one side affects the other Decarbonizing electricity means ejecting large shares of intermittent renewables at the site level and at the transmission level So correspondingly in the future grid Flexibility and demand at the transmission level and at the site level will have tremendous value more than that Will be needed if we want to achieve our electrification and decarbonization goals what it's going to take to get there is Practical solutions. I think I would say always a boring solutions and that's what I want to talk about today So first I want to talk about some evidence We found that there's already today in today's grid significant amounts of flexibility that we can use to reshape and decarbonize electricity consumption My case study is the campus energy system that Lincoln introduced Whose electricity consumption is roughly equivalent to a 30,000 people city in? California This this is my living lab. It's it's a perfect experimentation test bed because the 150 buildings or so on this campus are all managed by one entity Stanford land buildings and real estate and this is my schematic for our integrated energy system We have what's called a district energy system on campus, which was almost fully electrified Lincoln talked about this because Most of the heating and cooling that we're using in the buildings is produced on-site at the central energy facility Which uses large heat pumps? And to convert electricity into hot water and chilled water the first source of flexibility like to explore is in scheduling this heating and cooling plant and So we're facing a decision-making problem here, which is scheduling the different machines the hourly operations for the different machines at the central energy plant I want to minimize energy costs while meeting constraints in the plant and Meeting system demand for cooling heating and electricity and so on the diagram here I'm just showing you the chilled water side, but heating cooling electricity are all Interdependent and connected here. So formally what we're going to do is write down an optimization problem We want to minimize some objective function By choosing the optimal values for a set of decision variables while meeting constraints The reason I'm showing this is that all of the applications I'm going to talk about next involve modifying this optimization problem in some way whether it's changing the Objective function adding new decisions new constraints or some combination of the three The first example I'd like to talk about is a real-world experiment in the summer of 2018 We participated in a demand response program with Pacific Gas and Electric the passive bidding program So in this program we were paid to provide energy flexibility to the grid So we modified that optimization program I was talking about to include the payments in the objective function. We added new decisions and constraints in In the problem to model how we were delivering the flexibility The so in the plot on the left here the the green line shows the total Electricity that's going to the central energy plant. So that's our dispatchable load what we control the orange line corresponds to The rest of the campus load the electricity going to the buildings That's our non dispatchable load and the blue line is what PG&E sees That's the what we're accountable for which is total campus electricity load Most of the flexibility for responding to these three events in July of 2018 Is coming from flexibility associated with the tanks the thermal storage tanks that I think Lincoln already showed this picture This is really the bulk of the central energy plant these storage tanks This is our battery if you like this is providing a coolant service to battery, but at a much cheaper cost the second Type of experiments I'd like to talk about our thought experiments So here I'm showing the total campus electricity load So heating and cooling for this from the CEP and also the buildings over a year And so that the heat maps here each column corresponds to day of the year each row to an hour of the day in the top one This is the optimal behavior if you want to Follow the current electricity tariff which is telling us have load. That's as flat as possible every day And when the electricity is cheaper for us, that's typically in the in the afternoon consume a bit less And so here our peak load every month is somewhere between 30 and 35 megawatts under this scenario in The bottom one where I'm assuming we're paying a price on carbon That's so high that really all we want to do is shift as much load as we can from the night to the day because this is sunny, California So, you know, this is a proxy for saying how much of that solar power can we absorb if we wanted to reshape consumption with the thermal storage and here the lows dropped about 20 megawatts in the middle of the night and Jump to 45 megawatts in the middle of the day So next I'd like to talk about some options that Might not be available right now, but would be available very soon Given investment and the black box. I'd like to open what was until now a black box is the energy consumption From the buildings on campus so far. I've considered that's low that I have to meet no matter what for heating cooling and electricity So I'm going to change my optimization problem again Because now I can do demand response, but this means or what I'm calling thermal demand response This means I'm going to add new cost functions I'm going to add new decision variables. For example, now I can change thermostats head points inside the buildings If you know it gets to that I can start to do some chill water cutelments And I'm going to have new constraints. This is going to complicate my optimization problem because now I need to get in addition to decisions inside the central energy plant. I need to make decisions in the 150 buildings on campus These buildings are what are called commercial buildings So that cooling comes to the buildings in the form of chilled water that's used to cool down air At a central location in the building which is then pushed through a central duct to the different zones on campus occupancy whether The buildings physical characteristics control systems all interact to determine what the building's cooling load is on a given day Which is why answering this question is not as trivial as it might seem if I changed The temperature set point by 2 degrees Fahrenheit in a subset of the zones so 2 degrees Fahrenheit That's 1.1 degrees Celsius and if I did that how much would the cooling load drop over the course of a day? So we went out and started testing some buildings and here I'm showing you some data From one of the buildings on the engineering quad so the y-axis here corresponds to The daily cooling load for the building as measured through the chill water loop The x-axis corresponds to the mean daily temperature as measured at the campus weather station Each dot is one day's worth of data the blue dots correspond to our base set point of 74 and the red One degree higher the difference between the two lines here That's what we were trying to measure. What's the flexibility in the system if I increase these temperature set points? So we tested three buildings in 2020 six in 2021 11 just last summer, but testing all buildings on campus at this point We would have implemented the program. So we also worked on tools to extrapolate The results that we found buildings to the rest of the campus. This is our basic recipe in four steps One run tests in a subset of buildings. I talked about that Second we collected a feature dataset for all the buildings on campus including the buildings. We tested and the buildings We didn't test third So what the feature dataset tells us is gives us an idea for how similar to buildings look so with that information and Information from the tests we can generate estimates for Flexible what we think the flexibility would be in the buildings. We haven't tested and then fourth We can combine our flexibility estimates for all the buildings on campus to estimate campus-wide load reductions Overall and I'll get to how I had that number in a second We think 14 percent is or 30 13.5 to 14 percent is the aggregate Flexibility we can get from campus. How do we get to that? So first we ran these tests We ran a lot of tests. We ran 1200 You know, we collected something like 1200 days of experiment data in 11 buildings over the course of three summers And all of this is what was done with distributed sensors and actuators that were already inside the buildings that allowed us to control about 1300 zones which pretty much equate to rooms and importantly we excluded 360 critical spaces so things like temperature sensitive labs We're calling these stress tests One one reason they're valuable just as is is because they provide very directly relevant statistical information to District energy system operators and electric grid operators they also are fed into these models I alluded to to extrapolate to the campus level and Give us estimates for what the campus-wide potential for flexibility is and so the the curve I'm showing here Some of you may be familiar with curves like these on for electricity. This is what we're calling a chilled water daily load duration curve each so he this is each day's worth of data for Campus cooling load in 2020 worst to easiest from left to right the blue curve is what actually happened and orange curve That's our counterfactual load duration curve under 10 days of demand response So this is what we think we could have seen we could achieve if we were doing demand response on campus And so this is where am I the 14% number I was quoting from Earlier and actually I want to so there was a question from Peter Rumsey earlier about chill water contaminants What did this was one of the drivers for setting up this project three years ago was thinking okay? You know back then as Lincoln said there was a Capacity shortfall we just didn't have enough and so part of the question that I was asking at the time Where I wanted to ask with with the my partners on this project was What if you know next time than I in the next five years or eight years down the road We come up with this sort of decision again. Do we need to spend? you know that money to buy all that extra capacity or Could we do things differently? So I just wanted to echo that To finish I'll just leave you with these three ideas the grid is changing very fast and that change means that Flexibility in demand will have tremendous value already today, but even more in the years to come In the grid of today. There's already significant amounts of untapped flexibility. That's available for pretty cheap You know, I think we've found evidence of that here and given investment tomorrow. We can have a lot more Flexibility if we if we go look for it We have about five minutes left. So any questions? Quick question on your Exclusion of your critical facilities what percentage of the facilities across campus would you consider our critical in that? model That's a very good question So and that's actually one of the black boxes I was trying to open with this project So there's currently there is a critical list of you know buildings being You know, and I think it's ranked from zero to five be or there's six or seven levels one of the difficulties with the current system is that This is not exactly the way that the the list is set So so Lincoln correct me this is the way I understand it You look inside the building if there's a critical zone in the building then that building becomes critical to that level You know, whatever the worst level is that's the level what that means is that I think it's 80 90% Of the campus cooling load right now is critical This is one of the black box. I'm trying to blocks is I'm trying to open and to say instead of having a prior list that is per Perbuilt, you know one, you know one building at a time. Can I have a prior list that is zoned by zone? to answer your question About 40% of the I think my least number I have in mind about 40% of the campus cooling load is for the hospital but in a large large fraction of that is critical loads, but they also have Offices in the hospital currently, you know, all that is sort of something that we don't want to touch in the future you know why should my office be different from the hospital office, I guess is my question and and Maybe just to add one more thing on that In so two of the 11 buildings I talked about that we tested were lab intensive buildings and still you know So we got less less flexibility from those buildings because I was excluding a much larger number of zones But still we got three to five percent percentage reduction loads by controlling just the office portion of those buildings So even you know even in this The buildings with sensitive loads, I expect you would you would get something a couple more layers of complexity there one is that As Jacques said a building might have a laboratory an office a classroom All sorts of different functionality with different levels of priority The second part is that the distribution system the electric distribution system is set up like most electric distribution Systems in that it is not oh here's a critical feeder and here's a non critical feeder You've got all sorts of different levels of criticality On the feeders on each feeder so it may be in a inner in a restart situation after an outage You're turning on an absolutely vital building, but you're also turning on a coffee shop And there's just no way around that without an army of technicians that would that nobody has The third complication maybe third and fourth complication is that we do have emergency generators but those are How those are wired into buildings is Not optimized for These sorts of events and then what ends up getting plugged into those red outlets is Up to the folks in the buildings and so there's a such at a minimum There's a situational awareness issue from the outside of the building into the inside of the building So this is something That we are all of these things we are trying to unpack but layers upon layers of complexity and then Ajax pointed out Everybody's work is important to them. Everybody's work is critical to them and that's completely valid But somehow we have to unpack that as an institution, so I think that's like seven layers of Complexity there, but that's the that's what we're trying to figure out Maybe just I can't resist to add one word to what Lincoln was saying about you know We can't control, you know, we selectively I would add the word currently, you know currently we can't do that But I you know that doesn't mean it's not possible. I think there was a question over there. Yeah Yeah, hi, Florian here. Yeah, it seems like you have found wonderful ways to play around with the electrons In a very flexible manner. So congratulations I'm a little bit worried about the data security of this system and the overall security because The more decentralized steering we have the more IOT devices we have in the system Security becomes an issue. So maybe you can elaborate very briefly on your strategy to Security in general, but also the data and the device side of security. I can talk a bit about that. Yeah So one thing I can say is when we were running these experiments So my you know my laptop doesn't have access to this there They're sort of different level the same way that there's an IT system that controls that central energy plant and that has some Safes around it you have the same things in the buildings And so there are different layers that you're adding at different parts that have different levers layers of access In other words, I'll say this is slightly different way I don't control the actual devices inside the buildings when I was running the experiments I what I do control our control set points, you know So I think it wouldn't be too difficult to add layers of protection within the building to say You know, there's a feasible range if the outside system is trying to push something to me. That's completely unreasonable I'm just going to reject it and build in a failsafe that way, you know in other words there and we're not I think there are there ways that you can work on this actually sitting right next to you is wrong radical ball who does some of this work around around Around around electricity and and thinks about that Okay, I think we have time. I think it is a concern, but yeah one more quick question and then we'll have to finish off this session I'm not familiar. This is David chin with marata again I'm not familiar with the building codes in California. This were there any building codes that you had had to challenge or Modify or you know address so No, I think is the short answer I do think one, you know one one thing that your question does make me think of is that there would be a lot more Value for flexibility on the electrical system side if the buildings code change one way I know in which they it would be nice for them to change to unlock a lot more value for flexibility is currently So I think that the rule is basically you need to have enough copper on the system to take the worst-case load and You can't have for example a failsafe that's built based on software to say when the loads get too high I'm going to start shutting things down so that it doesn't you know, it doesn't blow up Currently it has to be there has to be enough hardware in there if that but that's just a rule You know regulation if that you know was lifted and you could say well now I can count on the software to say if all the EVs are charging at the same time But you know when everything else is happening With software, I'm going to control that then you would lead a lot of not you know You could say I'm gonna have build a system with much less hardware. I think there you're gonna start seeing a lot of investment value Okay, I think we have to close this session So please join me again in thanking our two speakers and I'll move it over to sustainable transportation