 So first off welcome to Stanford. I'm sure you've heard that a lot It's a wonderful place with a ton of opportunities I'll talk a little bit about what our group is doing I'll finish with a little bit of a plug for some of the classes we teach and I teach in ERE Which I think are pretty broadly applicable and interesting to students across Energy sector. I'm gonna talk today about some work that really comprises these two areas comprise maybe two-thirds of my group's research Other areas we're working on are solar thermal energy storage solar forecasting and some other renewable energy technologies, but I'll talk about oil and gas emissions intensity So renewables are growing incredibly quickly. I've been working on energy climate interactions for about 20 years I sort of got obsessed about 20 years ago I've never felt more Positive about the potential for the future. Things are looking really good There's a lot of work to be done There's gonna be decades of work to be done, but a lot of pieces are falling into place So this is very exciting. Renewables are great incredibly quickly But also gas demand is increasing quite rapidly both in North America and in Asia There's a variety of reasons for this One is clean power generation another is the desire for process industrial or commercial heating without particulate matter Another is flexible generation of firm renewables Every day in California including today between about four and seven Clock as Lynn said we ramp. We're gonna ramp today 12 to 14 gigawatts of Natural gas power over the space of a few hours That's more than the entire grid of almost every state in the country We'll turn that on right and so flexible generation of firm renewables is going to be needed a lot of interest in natural gas As an alternative vehicle fuel. So gas is is with us and in fact gas demand is growing And all credible scenarios show decades of oil demand Remaining and I'll show some examples of that on the next slide For that reason we need improved methods to understand and reduce emissions from oil and gas Operations while we're still going to be using those and I'll show some examples of our work in this area So Adam says oil demand is going to continue for the next few decades. Well, Adam is interested in oil So of course he's gonna say something like that. He's not a neutral observer. This is not just me So we took every projection we could find including intergovernmental panel on climate change projections from the year 2000 and the year 2016 2000 we filtered for ones that were already rendered Inaccurate by the history that's elapsed since then so we took all the ones that still could be possible 2016 this is the newest set of of integrated assessment model projections We also took every projection for future demand we could find from oil companies as well as from government and non-governmental organizations such as IEA the International Energy Agency in Paris I You plot those these are the Kim or these are the yearly demand with historical consumption by country there, so we have a hundred and something countries here Here's the projected future of production and then here's cumulated Okay, what you see here is that on the cumulative graph. We've consumed maybe 1.3 to 1.4 trillion barrels or 1,400 giga barrels over the course of since the industrial revolution Okay, this graph actually goes all the way back to 1870 You'll see that even in the most optimistic scenario. So here's cumulative consumption associated with this most aggressive IPCC mitigation scenario, so this is like a two degrees mitigation scenario where the oil demand essentially Craters we've got another trillion barrels plus to be used over the over the century and and let me just re-emphasize this This is the intergovernmental panel on climate change mitigation scenario. This is not ExxonMobil, okay So maybe another another trillion barrels of oil So how do we reduce emissions if we're oh and okay, so here's a here's a little Quiz that I sometimes give my students to illustrate this Let's say ex Commercial scale airplanes were sold in the year 2017 the last complete year X's I don't know what the number is 700 or something like this. How many of those airplanes? Let's say the number was a thousand how many of those airplanes are still going to be In the air flying in the year 2050. Do we have a sense of what that number is? We sold a thousand last year. How many are going to be left? Not zero Not quite 800 it'll probably be about 600 to 650 the average lifetime of a commercial aircraft is increasing in the age of 50% retirements is running about 35 years right now Okay, so of the airplanes sold last year 60 percent or so are going to be in the air in 2050 Okay, that's just the reality when you spend a hundred million dollars on an airplane You don't shut it off that thing runs. That's just the way business works. This is unfortunate, but true Okay, so life cycle assessment of oil and gas operations. This is one area where we're working So, you know, all of this is just to motivate, you know, why are we interested in working on this? When you evaluate oil resources, you need what's called life cycle thinking life cycle thinking basically says I want to evaluate the impacts of a particular option over its entire Value chain common examples of life cycle thinking or life cycle assessment that you guys are probably familiar with our questions of Do I use a disposable coffee mug or do I coffee cup or do I wash my coffee mug? Do I use disposable diapers or do I wash? Diapers things like this. So these are common plastic versus glass for Coca-Cola bottles things like this So these are common Applications of life cycle thinking but we need to apply this to oil and gas as well But we really wanted the minimal emissions across the entire value chain all the way from exploring for the primary resource drilling of Well is producing of oil Processing it shipping it to refineries refining it and then burning the fuels Some resources is important to take this full value chain approach because some resources will score Well in certain parts of the value chain, but less well in others And so what you really want is the resource that's best across the entire life cycle Here's an example of an oil value chain. You've got exploration separation and reinjection and you transport it refine it refine send the product To the gas station and you consume it some terminology here well to tank is usually here Well, the wheels includes the product consumption in your vehicle We built a model for this as we've been working on it since 2009 or so we call it OPG the oil Production greenhouse gas emissions estimator estimator is on bold here What OPG does it is it estimates emissions from an oil operation given engineering characteristics of that field Things like depth the quality of the oil how much water is going to be produced, etc A whole bunch of published papers the model is completely open This was developed with funding from the California Climate Agency the California Air Resources Board They've funded development this of this tool for about nine years as part of the climate regulation in California They need the numbers that this tool generates As well as Department of Energy Carnegie Endowment Ford and Saudi Aramco What does the model look like it's a flow sheet kind of model approach so any of you have taken a chemical engineering class We'll recognize this kind of flow sheet. It's nice because all the equations are accessible able to be modified examined fully open source We're shifting some work that we're doing in the group and some of our new students are going to be working on this One of the pieces of criticism we've received over the years from industry is they want to see Sort of standard stock chemical engineering tools used so what we've done is we've used One of these expensive process simulation software to run in this case This is an acid gas removal unit. We simulate the model or we simulate the process 10,000 times Then we can actually fit an equation to predict in this case It's just a quadratic regression to predict energy use in the boilers or electric work for pumps And we can put that into our lifecycle assessment model to calculate the energy demand So this is a big step forward for lifecycle models. Typically, they don't tend to operate at this level of Kind of chemical engineering detail So we're really trying to push this the state of the art forward in the modeling of greenhouse gas emissions We also have done a lot of work in trying to understand emissions from the global oil sector So this is a paper I'm very proud of it was just published in science two weeks ago or so On August 31st We had 24 co-authors from 16 institutions for the first time we constructed a global Estimate at the field level of emissions from global oil and gas operations This is essentially every oil field in the world about 9,000 oil fields representing 98% of global oil production a funny For those of you who have excessive ambitions and are entering graduate school in the summer of 2006 So I was giving my qualifying exam and I had the audacity slash insanity to propose in my oral exam Some of you guys are going to be doing this soon I propose that I'm going to build a model to assess the greenhouse gas emissions from oil and gas operations And I'm going to model every oil field in the world One of the guys on my committee said great project. This would be super interesting. It's going to take you a decade Literally almost 12 years later to the day So, you know 800 references years of data collection we built a lot of data integration methods To using R and other sorts of SQL languages to basically integrate many many data sources And we can start to target these high-emitting sources We think the potential reductions are something like 50 gigatons of CO2 over the course of the century That's in relation to about an 800 to a thousand gigaton budget to stay under two degrees So 50 gigatons is is non-trivial if we can get these high-emitting sources and make them look more like Norway For example, it's down here We looked at variation by country so some countries are really good Norway other countries are really bad Canada, Algeria, Venezuela Depends a lot on the resource Canada has great environmental management rules, but their resource tends to be heavy and hard to extract Okay, another area we work on is methane detection Methane is a potent greenhouse gas as Lynn said it's emitted from oil and gas operation that has a high global warming potential we want to Develop ways to vastly reduce the cost of finding methane leaks and fixing it We went out to a test site where they have metered leaks from real oil and gas equipment They can basically turn on the leak and tell you exactly how much leakage is happening We gathered about a million frames of video over the course of a week at different distances different pieces of equipment Different leak rates different sun angles different wind conditions, etc. So about a million image Hours of video using a hundred thousand dollar infrared camera that lets you see Temperature sufficiently With a sufficient high fidelity that for example, you can see how much oil is in a tank You can also see normally Transparent methane which is very active in this mid IR spectrum. This is in the three micron Window where CH bonds are very active. And so that's actually methane coming off of the tank in North Dakota Okay, so we want to take those videos and do automatic detection. So here's six leak classes There's this vent stack. This is a big leak. This is no leak All right, okay, so those are some example frames of video. We want to be able to detect that We're using CNN's convolutional neural networks, which are an image processing technique And we're training the CNN to see plumes first few processes. So you take a raw video stream. Here's a leak I Can then binarize it or increase the contrast? No, yeah, that's contrast increase. That's not a binarizing then I can use edge detection to detect change So these are actually change vectors from one frame to another. How does that pixel move? So I'm detecting movement then I can threshold it and say select all pixels where there's a certain amount of movement Those are yes That's a yes pixel for a flame for a plume of gas And then we can recognize that and hopefully what we think is automatically detect that and then estimate How big it is so that you can have something like a camera watching every oil field and it would report in and say I see a leak on tank number six. Here's a video of it send someone out go fix it Okay So this is pretty fun Another thing is super cool. So this is my handsome postdoc Arvind. He's actually a professor now So he's I need to change the slide He's left Stanford onto greener pastures We're working with the environmental defense fund and what we call the mobile monitoring challenge This was three weeks of field tests with 10 participating companies who are all developing in about half the cases drones and the other half aircraft or Two cases of light aircraft or trucks that can drive and detect emissions without stopping So the idea is these are mobile Very rapid so that they can be much much cheaper You could survey a hundred well sites in the day, for example, okay So this could drastically drop the cost of methane detection So what we did is we went to two different sites in Sacramento in Colorado We were releasing gas at these various points Companies were driving their sensors and trying to say do I see a leak or do I not single blind format? So we were controlling it. They didn't see anything then all 10 companies had to report what they saw to us And then we're going to publish the results With names and everything attached. This is you know straight up exactly how these companies did We're seeing we can't talk about the results in detail yet But we're seeing some really positive results some companies have really high detection fidelity And do didn't a few of the companies just did an amazing job And so this technology this mobile detection technology. I think is is is here. Okay So this is we're releasing at that site in Colorado where we did the the videos for plume detection This is a rig that we set up in Sacramento to do larger leaks These guys are only permitted to leak a certain amount per day about a hundred and fifty standard cubic feet per hour in sane units One scuff is about 20 grams Couple kilograms an hour Yeah, standard cubic feet. I'm sorry. I don't know what to say Apologize That's too small for some of the aircraft So these guys here could release a much larger volume and these here is where we tested the light aircraft So here's the idea is Mounts a camera in most cases an infrared or hyperspectral camera on the underside of an airplane And then you can fly by it, you know 150 to 200 miles an hour at a few thousand feet And the idea is you could for example survey wide swaths of oil and gas fields Look for emissions of thousands of wells. So we needed to be able to release a larger amount So we did that up in Sacramento With all the appropriate permits this was out in farmland And so there was no potential for any sort of safety or human health concerns associated with the gas release So that went well. We did three weeks of releases We're analyzing the data. I write now Piccaro is a Stanford spin-off They do what's called cavity ring down spectrum Spectrometry very very high precision parts per billion level Detection of nothing. So they basically in our case, they drove a drone pass that have the center on it They're measuring the concentration and they use that plus the wind Fields that they're measuring to estimate the leakage rate These guys were drones Actually these two were drones as well these folks eras there are another bay area startup They've got a sensor on a vehicle Kairos is another bay area startup. They're a airplane as is ball aerospace Ball aerospace has a long-time contractor who's built a lot of the satellites for NASA Etc. So they've got a really nice camera system on an airplane As well as one university team from Calgary, you know, when we did this we were partnering with environmental offense They copped out the money for the experiments. We're thinking. Oh, we got to really hit the bushes and hope we get 10 companies to participate I was thinking we were gonna say yes to everyone We ended up getting 28 applications including a number from companies that I didn't even know we're working on methane We ended up selecting 12 and 10 actually did measurements They're gonna say do I see the leak Yes, no estimate the leak size And give us the location and time of where they think the the leak occurred Then we're gonna generate things like cumulative cumulative probability for detection as a function of leak size Quantification accuracy and the effective conditions. Hopefully something on the effective weather conditions for detection accuracy What we're seeing roughly and we're drafting right now to submit this fall what we're seeing roughly is Some of the technologies are very very good at Seeing the leaks Very few of the technologies to maybe none of the technologies are very good at quantifying the leaks It's actually a quite a difficult inverse Transport and dispersion problem So you have a set of concentration measurements for example you drive by a well pad and you're measuring methane and the methane increases and drops What kind of emission rate does that correspond to it's actually quite a challenging problem So I think yes. No. Do we see a leak or do we not it's really possible Very fast very cheap. This is coming. It's going to be great And but quantification is probably going to be hard So that's it. That's just a sort of a short preview of Bunch of stuff that's happening in our group This is time for questions There is one commercial satellite in the air presently it's called ghc sat. It's out of montreal The current version of the satellite isn't working super well They've released some preliminary images, but they're working on a second gen Environmental defense fund raised some amount of money. I'm not sure I hear numbers of 50 million dollars or so and they're actually building a satellite They're calling methane sat That's going to go in the air within the next three years or so and hopefully they're going to be able to observe About 80 of the world's oil and gas fields The challenge with satellite so this is amazing right if you have a satellite that visits the whole world every day This is like great. The challenge is typically you can only see large events And so I still think it's going to be very worthwhile because if you look at the statistical distribution of methane emissions A large fraction of the leakage comes from pretty infrequent large events So really it's going to be a big benefit even if you can only see really big stuff from space You can look every day. You don't need permission from companies. Anyone can see it They're going to put all the data online So hopefully the idea of what the edf is doing is it'll look like google earth you can zoom in Anywhere anyone can see it from the web and you can see a methane cloud over a facility And environmental regulators or other folks can anyone can access the data. So yeah, that's a that's a very good question Lots of people are thinking about it. They're saying 2022. I think is their target. That may be a little ambitious, but possibly doable Mr. Dany work on estimating to the extent Offset the benefits of fuel switching from called natural gas in terms of like Yeah, so the for those of you who didn't hear the question the question was Have we looked at the benefits of cold to gas switching? In the context of the methane emissions This really was the impetus for a lot of the the explosion and work on methane emissions Sorry for the bad pun bad metaphor A real sort of Explosion i'll do it again A real sort of explosion of research on methane Over the last five years a lot of it motivated by this question of should we actually switch to gas? Is it actually better? We haven't written any papers explicitly looking at that at the most recent We had a paper published in science. There was a minor co-author Rolling up about five years of work environmental defense did the roll up They think the leakage rates are about 60 higher than EPA says In the u.s. But still significantly lower than any kind of threshold you'd want To make coal versus gas good so it looks like coal versus gas trade-off is still quite good vehicles It doesn't look like cng and lng vehicles are going to have much climate benefit. That's probably worth things stand right now It may have other air quality benefits, right diesels can be polluting but but for climates doesn't look great Could you give us an idea of the estimate cost of the companies of We've we've got a field campaign up in Canada right now and we're paying Two to three thousand dollars a day For a team of two guys who have one of these hundred thousand dollar cameras and goes out to look They can see it depends on the complexity of the site if it's got a lot of equipment on it They'll see one site a day if it's they're very simple those they can see 10 sites a day Really what you need is you need this to be that that's too expensive This is hundreds of dollars to thousands of dollars per site But you really need is 10 dollars per site visit something like that And so I think you need something that looks like a smoke detector in your house, but it just sits there You stick it to a piece of equipment it radios if it smells methane right It's just uses the mesh network or something else to send a alert home A lot of these sites aren't manned Frequently someone goes out once a week or less So you need something that sits there or you need something like a camera that sits on a truck And as men come to to do work on the facility the camera is just watching and radioing home with problems It needs to be effectively zero labor fast And cheap and I think actually with these sensors were probably a lot closer than some people are thinking And so then you're talking 10 to 20 dollars per site per visit And you can easily save that in the gas savings loan And so then the then the companies are going to start getting excited because hey, I'm paying 20 bucks a site to To detect gas, but I'm saving a hundred dollars worth of lost gas per site And so, you know, this is just this is win-win, right? We want win-win solutions where the companies see this as a no-brainer want to do it automatically Why would I not do that right now the perception in industry is that this is finicky Expensive labor intensive you got to get guys in a truck. They got to go out there the cameras a hundred thousand dollars you know So the perception in industry right now is that it's troublesome and But we need to get it to where it's effectively automatic. I think Um, you know, there's been a lot of work, uh, you know done on that over the decades I think the short version is and we've got folks in our in our Department who do lab work with high-pressure gases The short version is that nature abhors a vacuum actually what what nature doesn't like is a pressure differential And so that gas relentlessly 365 days a year 24 hours a day You know is that high pressure and wants to escape through any sort of orifice What you find is a lot of wear and tear. It's out in the elements. You get a lot of thermal expansion Cooling and warming cycles. You get a lot of vibration You get a lot of reciprocating motion in well pumps and in compressors And so just what ends up happening is that seals gas gets flanges things like this wear And uh You know to use the cleaner version of an analogy stuff happens, right? So sort of stuff happens and so really I think hoping that we're going to come up with systems that never leak is hard But thinking about ways to very rapidly find a problem when it occurs. I think it's totally doable and so So yeah, so that's where I've been focusing my Yeah, my efforts High-pressure gas wants to escape is the sort of the short version How Big improvement can you offer like the fields in Canada that you mentioned at the beginning that you were evaluating That just have you know sources of oil better low quality What we see globally is that what's driving a lot of that curve is flaring So we're doing a lot of work with satellite based flaring estimation I'm going to be doing more So these red ones are ones where gas is being burned essentially because there's no market use for it There's no pipe. There's no way to get it to people who will pay for it. So they just burn it So globally that's probably the big driver, but the second one is heavy Is blue here and heavy resources basically the way to think about it is that there's more carbon and less hydrogen So these are longer chain molecules. They're stickier. They're like a tar and asphalt or a bitumen rather than an oil that flows easily So when you have this sort of situation a couple things happen You end up steaming the oil or heating it to get it to flow So like butter and a saucepan so it's the melt and flows easily And then you have a lot of hydrogen addition or carbon rejection required to turn it into fuels that we use And so there's some inherent challenges with those resources There's a lot of people working on alternatives like low carbon hydrogen generation Or ways to essentially leave the hydrogen leave the carbon in the ground So for example, coke or chemically processed biologically in some cases the resource underground So that what comes out is very hydrogen rich And then you leave most of that carbon in the subsurface. So there's a lot of people working on it It is pretty challenging. I think this red stuff the flaring That's a management problem. We just need to decide that this is not okay This is three or four percent of global gas production You know, this is ridiculous. We need to move past this There's a lot of poor people in the world who could use the energy. There's a lot of Need to reduce climate emissions. Let's not burn three or four percent of world gas supply For no good reason. This is just crazy So I think this we just need to decide that this is not acceptable and put management and rules and regulations in place So that's where I'm most excited about Maybe I shouldn't be giving people but I don't know