070 Bill Bartling and Wes Cruver, Geo2Watts

Bill Bartling (00:00)
So we started looking at things called phase change materials.

So there's a lot of research going on in this, but phase change materials are really a pretty simple thing. And the simplest example of that is ice cubes, right? You throw an ice cube in your drink, it absorbs heat as it melts, phases change from solid to liquid, and your drink gets cold. And then you can put it back in the freezer and you can suck the heat out of it and turn it back into ice. So that's pretty simple. But there is a much more sophisticated versions of that. So like the blue ice you put in your lunch bag is a little more sophisticated.

that but there's industrial scale phase change materials that can go from cryogenic temperatures all the way up to thousands of degrees centigrade and their job is to as they go from a solid to a liquid or from a liquid to a gas is to absorb energy and then you can return the energy back when you go the backwards in that phase change and there's new materials coming out now which are actually tri modal or even more more efficient than that so they not only go from

solid to a liquid, but there's a thermochemical element to it as well, which is going to give about a 4 to 6x improvement over even the best phase change materials that you can get commercially today.

Mark Hinaman (02:17)
Okay, welcome back to another episode of the Fire Division podcast where we talk about energy dense fuels and how they can better human lives. We've got two special guests on today, Bill Bartling and Wes Krueger from Geo2Watts, which is a little bit of a different topic from what we've discussed before. Excited to learn about your guys' technology and kind of some of the stuff that you guys worked on. So Bill and Wes, welcome to the show.

Wes Cruver (02:41)
Thanks for having us.

Bill Bartling (02:41)
Glad to be here, Mark. Thanks.

Mark Hinaman (02:44)
Yeah. Before we dive into kind of our current ventures, you, or a current conversation, your guys' current venture, do you guys want to give the audience a little bit of background on yourselves, where you came from and how you got here?

Bill Bartling (02:56)
You go first, Wes.

Wes Cruver (02:58)
You do it. You've got the better background.

Bill Bartling (02:59)
Hi.

Mark Hinaman (03:01)
Fire away, Bill.

Bill Bartling (03:01)
Yeah, thanks. I'm Bill Bartling. So my background is in geology and geophysics. And so I started out my career in oil business and exploration production and research, then got into high tech. So quickly moved into high performance computing, advanced computational systems, eventually into fiber optics technologies, and then became a regulator and eventually got involved with this company.

which is very, very exciting technology, which is based on some of the things in the oil business, we'll get a lot more into later. But so that's basically where I came from and been in the industry around about 40 plus years and it turns out don't know how to retire. So here I am doing it again.

Mark Hinaman (03:41)
I love it. I hope to be there myself someday. Thanks, Bill. How about you, Wes?

Wes Cruver (03:46)
Yeah, my background is quite a bit different. I don't come from the energy sector in any way, any meaningful way. My background is more just on entrepreneurship. My dad's been a serial entrepreneur his whole life and tried to get me into it very early on. And so at 11 with his guidance, I started my first company. And it was sort of just a company to help spread my love of technology and science in general.

kind of a predecessor to kind of like Khan Academy in a sort of way. So was a lot of educational material that we would distribute. And as the internet became a more viable option in terms of delivering that content online, we started building out studios in Los Angeles and distributing the video content there. And that eventually morphed into my next company, which was

effectively a video based company that was there to address all the shortcomings and problems that I saw when trying to distribute it in the first place. so this is before YouTube really, or we started right around YouTube did. And we would call it just basically like a internal version of YouTube. And so we had large enterprise customers like Audi and Volkswagen and PNC Banks and Cox and a lot of pharmaceutical companies.

Bill Bartling (04:47)
you

Wes Cruver (05:00)
And they would use it for just the sort of generic things that you would want it for. So this kind of interaction, but also large corporate communication, training programs, everything like that. And one of the things that we noticed when we were trying to scale up this technology, because back then video is extremely difficult problem to solve. And we would have these huge spikes all the time of tens of thousands of people that would immediately log on over the weekend to

complete some curriculum that was assigned to them. And it was really difficult for us to scale to meet that immense demand that would just spike. And so the amount of traffic that we were getting and the data that we were collecting from that became extremely useful for us to sort of predict these waves and these huge spikes. And that started my sort of interest in AI and in machine learning and

And so, yeah, ultimately I'm here with Geo2Watts to help them down that path of AI and integrate both integrating it, but also assessing the upcoming giant boom in terms of, you know, energy demand that's going to be coming from these large AI systems.

Mark Hinaman (06:07)
When was that Wes, the big spikes that you guys were seeing and how did you guys tackle some of those problems?

Wes Cruver (06:14)
So luckily, by this time, we had moved to Amazon. And so it was pretty simple for us to leave some buffer, like 10 % buffer of idle compute, and hopefully try to spin up enough servers to meet that demand. But it usually could be challenging for us. And so we had to eventually predict that.

A lot of it was just turning out to be, you know, some sort of assignment that, you know, all of their like sales teams had to be trained at a particular time. They'd all push it off to the last minute. And so when it's due by Monday, they'd all log on and Sunday night. so some of those things, if we were talking with our customers and especially really large customers like that, we of course would. And so we have some upcoming notice of it, but eventually it became a very difficult problem to solve just in terms of all of the different customers that we have.

There's no way we know when every single one of them is planning a large event. And so, yeah, it just came about just as our customer base grew, it became easier and easier to predict when certain events would happen. So a lot of it was just like, you had very traditional strict curves that would happen. And then occasionally, you'll see some big spikes.

Mark Hinaman (07:28)
Bill, when you mentioned high performance computing, was that in the oil and gas sector or were you kind of adjacent stuff? Talk to us a little bit about that.

Bill Bartling (07:37)
Well, it started in Northern Gas. And so there's a couple of really large data sets that we used to work with mostly in three and four dimensional seismic surveys. And these were usually in the offshore areas where they had to drag boats for 100 square miles. And you'd have to try to converge all of that into a single three dimensional image, actually with multiple variables and attributes associated with it. So just the computational side of that was really intense.

So we got into supercomputing. This was back when I worked for Chevron back then. And we got into the big craze and the big SGI machines back in the day. And then finally into the big cluster machines. So that was when I was there. Later, I got involved with it on the vendor side. I worked for a company which doesn't exist anymore. But you may have heard it. It's called Silicon Graphics. so we were into the very, very large graphics visualization game.

theater scale, IMAX scale, immersive with glasses, full immersive visualization systems and interactive stuff. So, when SGI came off, was a spin-off from Jim Clark, who was a professor at Stanford. actually invented the graphics board. And so that was his first big graphics company. All that IP has now gone on to, I think HP has it now. So I was actually involved in the supercomputing, big data.

big memory, big graphics stuff that was going on. And we experimented with things with Nvidia and so forth. And then when I got into the business of subsurface geophysics and the fiber optics business, we were starting to collect huge amounts of data from the subsurface. And so just in our own shop as a service company, we were providing, having to do a pretty heavy lift. And that's when we finally got into the GPU processing, which was really exciting. And we had gone through a lot of iterations of

Mark Hinaman (09:34)
you

Bill Bartling (09:24)
FPGAs and other such things and finally settled on. What's the big computational engine today, which is the graphics chip, right? And everybody figured out how much work you could do on those graphics chips. We just pushed it off as a side program, you

Wes Cruver (09:39)
So,

Bill Bartling (09:39)
know, just kind of leaving the organizational stuff to your main CPUs and learn a lot about how, you know, the limitations on data movement through these systems. And it was all very, very educational.

That was some time ago and obviously the technology is still abiding by Moore's law. You can fit it all in your phone now, the stuff that we used to sell for millions of dollars in a But yeah, that was how I got into supercomputing things which were, I mean, we called it the big data then. was terabytes, many terabytes back in the late 90s and today, put a terabyte on a thumb drive, right? So it's changed a lot.

Mark Hinaman (10:14)
Yeah. But for the relative

processing power available and the size of datasets available, you know, I mean, those seismic datasets are massive, right? Yeah.

Bill Bartling (10:25)
Yeah, yeah,

yeah. A lot of data. And then the reservoir models, you know, when you have something which is, it's actually a dynamic process too. You know, so you have time steps and you have to understand all the computational fluid dynamics through a very, very large model and look at all the iterations, this big Monte Carlo simulation. So those were processes that took many, many days. One project when I was working at SGI was with a major company.

oil company in Houston and their final processing leg step for their 3D seismic took 90 days on I think they had like 10,000 processors running in parallel this thing and it still a 90 day computational process for the last last migration step so yeah it took it took some CPUs

Mark Hinaman (11:13)
Yeah. Do you have a comparison for how much you think it would take now?

Bill Bartling (11:16)
Well, the problem is that with more computation and cheaper systems, you get bigger data. So it takes the same amount of No, no. Yeah, yeah, just get more data. Yeah.

Wes Cruver (11:18)
you

Mark Hinaman (11:21)
You just get more data. You don't actually solve the problem. We're like 90 more days, man. Well, the next quarter we'll have it.

Wes Cruver (11:29)
And

it's the same sort of metrics now for training these big LLMs. You want tens of thousands, even hundreds of thousands of these GPUs that run for three to six months or so on the training.

Mark Hinaman (11:40)
Yeah. Okay, well, I'm excited to talk more about the data stuff also, but yeah, Geo2Watts, what is this? What are you guys working on?

Bill Bartling (11:47)
So it's a pretty cool thing. So I got involved with this about a year ago with Phil Krueger, Wes's dad, who was the founder of the company. He has some ideas about trying to solve an interesting problem. And that problem is that there are hundreds of thousands or a million end of life oil and gas wells sitting around the United States. And so these wells, they're not producing. They've had a long economic life.

Some, many of them are emitting methane as a pollutant. And everybody wants to get rid of them. So the regulators want to get rid of them. The communities want to get rid of them. And the operators want to get rid of them. But just costs money to do it. And so there's a reluctance to get it done. So Phil had this idea, well, is there something we can do? Looking at the energy transition, is there something we can do with these wells to store energy?

And he first started out in geothermal and then we realized, well, what we need to do is put it in our own heat and put it in the well. So Ken does this idea that we take these old wells and then we seal off the bottom, which is where all the hydrocarbons are. So no methane gets out and we clean up the surface where all the pump jacks are and all the pipelines are. And then we use some or all of that rest of that borehole, which is basically just steel pipe. It's a container underground made out of steel.

And we fill it up with thermal energy storage media. And we heat that up during the day using solar panels or a heat pump. And then at night, we flip a switch and we divert that over to a power plant. We generate electricity. So the key points of this, the most important points of this are really two things. The well is a simple thing. It's just a steel pipe. It's seven-inch in diameter and is 5,000 feet long. So you have a lot of space that you can do things with.

But the materials that are in the well, the thermal energy storage material, is really crucial. And so we talked about using sand because there have been a lot of experiments, especially in Europe and Los Alamos National Laboratories, using sand. Everybody knows sand is a good thermal energy storage media. You go to the beach during the day and it's hot. The sun goes down and it's still warm, right? So you know it can store heat. But it's not the most efficient heat storing system. So we started looking at things called phase change materials.

So there's a lot of research going on in this, but phase change materials are really a pretty simple thing. And the simplest example of that is ice cubes, right? You throw an ice cube in your drink, it absorbs heat as it melts, phases change from solid to liquid, and your drink gets cold. And then you can put it back in the freezer and you can suck the heat out of it and turn it back into ice. So that's pretty simple. But there is a much more sophisticated versions of that. So like the blue ice you put in your lunch bag is a little more sophisticated.

that but there's industrial scale phase change materials that can go from cryogenic temperatures all the way up to thousands of degrees centigrade and their job is to as they go from a solid to a liquid or from a liquid to a gas is to absorb energy and then you can return the energy back when you go the backwards in that phase change and there's new materials coming out now which are actually tri modal or even more more efficient than that so they not only go from

solid to a liquid, but there's a thermochemical element to it as well, which is going to give about a 4 to 6x improvement over even the best phase change materials that you can get commercially today. They're basically going add a salt or the tri-modal is made out of boric acid and succinic acid, but the commercial ones in force today are made out of sodium and potassium salts.

basically, table salts. And it's pretty cool as you heat them up. It's called sensible heating. They add heat and they get hotter when you add heat. Then it goes into a point where this partly solid and partly liquid. And so that's what's called a different heating phase. And during that phase, can actually, you have to a lot of energy, but the temperature doesn't change. So that's a really important

That's the latent heat portion. That's a really important part of that cycle because you can really recover a lot of energy from that phase. Once it all goes to liquid, then as you add energy, it gets hotter and hotter again. So that's an important piece. And we've looked at a lot of things to really improve the efficiency on that. But the other part is the power plant. So the way the power plant works is just called a binary system. And so we will pull heat.

that we've stored in the well bore in the thermal energy storage media. put it in there all day. up to about 150 to 200 degrees Celsius. And there's a U-tube that goes through the well with a heat exchanger that heats it up. But then we use that same heat exchanger to extract the heat. And it goes to what's called a binary power plant. So that heated water is adjacent to another tube through a heat exchanger. And this other tube.

disconnected to the power plant. This other tube has a very low boiling point fluid in it. It could be some organic oil, could be supercritical CO2, could be hydrogen, could be helium. But because it has a very low boiling point, we can add that heat to it. It flashes to steam, spins a turbine or pushes a piston and generates electricity. We condense it into the side, recycle it back, turns it back into a liquid and we boil it again using this heat.

So it's a really neat program because it's a zero emissions energy storage and generation system. We're using the sun and we're using natural systems and natural substances such as salts and sand and water to do this. And it provides something which is really needed, which is a long duration energy storage system. So it'll generate energy for about 10 hours or more. So you can contrast this with a lithium ion battery.

These tend to discharge in two to four hours, right? So when you're storing energy from windmills or from solar panels, these big battery banks, that's great. You've got all that there, but it's really got a short discharge period. This has a very long duration discharge period, and that's really one of the things that people are pretty excited about. So you got these wells that you don't want anymore. They're emitting methane.

They're kind of a problem in the community. You need to have long duration energy storage, especially for states that are becoming more more dependent on intermittent and episodic energy sources, such as wind and solar. And then you have long duration energy storage at the back end. Taking a physical asset, repurposing it, giving it another 20 years of life instead of just pouring money into it and abandoning it.

Mark Hinaman (18:23)
That's super cool, Bill. Love the idea. I've got a bunch of questions. It sounds like, yeah, one of the primary innovations here is using an established tank, essentially large, skinny tube, as a holding system for your heat exchanging material, your salts. And presumably that's thermally insulated because you're in the subsurface, right? I mean, is that kind of...

the they are those the two advantages thermal insulation and then I guess heat exchange because it's a long skinny tube instead of a big fat tank on the surface because otherwise you just put some of these like a storage salt system like this on surface.

Bill Bartling (18:56)
Yeah.

They do that. So there's an example of putting them on the surface. And this is an organization called Polar Nights in Finland. And they basically have a grain silo on the surface. And they fill it full of sand. They heat that sand almost to melting point, of course. We're not coming anywhere close to that. It's pretty hot. But what they do is they harvest the heat and pump it into the city for heating homes and buildings and things like that. Because it's pretty cold in Finland in the wintertime.

Los Alamos Labs was also experimenting with using rocks, you know, in the silos. Yeah, the subsurface provides an insulation factor and one of them is the wind's not blowing on it, you know, and the earth is a good insulator in most cases. There always will be heat loss systems and we're studying how to deal with that and how to maximize, you know, the efficiency of the system. But yeah, the earth is a good insulator and it actually has a lot of heat of its own.

So it will contribute some of that heat over time. But yeah, it's basically a, it's a container that's already there, right? And we can repurpose it instead of, know, the end of life of an oil well, basically all you do is you go and spend a lot of money and abandon it and bury it. So here we're saying, let's put it to a good purpose for giving us clean energy.

Mark Hinaman (20:05)
Yeah.

Wes Cruver (20:19)
There's also tax advantages as well. And so that makes it a much more economically viable model. just having these long duration energy storage systems, there's a 50 % tax credit on top of that. There's carbon credits that are focused on sealing off these leaky methane pipes. And so all of these also help contribute to a much more economically viable product.

Mark Hinaman (20:19)
Hot ice.

Where are you guys in the development stage of this? Do you have a pilot plant or are you looking at opportunities to go and test it?

Bill Bartling (20:50)
Yeah, so what we've done so far is we have a pretty robust technical team, various guild in solar systems, computational thermodynamics that have done a lot of preliminary modeling. The good news is that all of the components of this are very well known things. So it's not like we invented something new. We actually know the physics and the behavior of all of these things under all these different environments.

So we're moving forward on to a plan to do some digital twin modeling. understanding what are the combinations of phase change materials in sand? Is it 100 %? Is it 10 %? Is it 50 %? If we were to do that in a well, let's put a whole bunch of stuff in the well and let's say a 50-50 and see how it goes. Well, you find out, well, that's not optimal. So now you've got to dig it out and it costs money, it takes time. In a digital twin, you can iterate on all those properties really,

quickly. So that's one of the things we want to do is get into computational modeling to understand the behavior of all these systems so we can optimize them. The next thing we want to do is we call a prototype. So basically what a prototype is, it's a big long piece of pipe on sawhorses in a laboratory. And so we can fill that full with phase change materials, different kinds of heat exchanger systems, instrument the hell out of it so we can get a lot of great data.

So then we can optimize it again kind of in a pseudo real-life situation. Then we go to a pilot to put actually in the ground because the geology will add some additional uncertainties that we'll discover then. So right now we're in that mode where we're designing these next few phases. We expect that's no longer than a year in duration. But meanwhile, we've...

Mark Hinaman (22:32)
you

Bill Bartling (22:33)
who developed a pretty robust intellectual property portfolio for how do you combine all these things together in this novel concept of using a well borer. So one of the things which is really, which to me is really interesting from a technical perspective, and I'm ask Wes to talk more about this, is let's just say there's a million wells in the United States that are candidates

Wes Cruver (22:53)
So,

Bill Bartling (22:55)
for this process. Well, not all of them are gonna be ideal, right? There's gonna be variables in them that are.

that are good, bad, and otherwise. They're scattered all over the United States. so one of the things that Wes has been piloting, spearheading for us is, how do you understand all that information? How do you decide which well do you do first, last, never, second? So I don't know, Wes, want to take it from there and talk a bit more about what you're doing with computationally with the data?

Wes Cruver (23:22)
Yeah, I mean, so it always starts, it feels like, in these ML-based.

systems, it always feels like about 70 % of that time training a model is really just data collection and acquisition, cleaning it up. And what we're noticing in, you know, every state is different. And some have really great records and it's all digitized, whereas others, it's still on paper. so lots of issues just in terms of placing and identifying where these wells are, because they may just use a lat long

coordinate for a large field worth of wells. So there's a lot of data challenges involved with that. But right now we've been collecting as much of that as we can. And then ultimately using that to identify, ideally, its clusters of these wells. So if we can get not just one well, but multiple wells that are all sort of close together, that also helps with the efficiency. And so

Yeah, we've been doing a bunch of modeling around what are these ideal wells and how do we locate them? yeah, so that's a big part of what I guess we're starting with right now on the AI side.

Bill Bartling (24:30)
Yeah, and also these wells are all built differently. so the standard model that we've been using is based on a well that sits in Long Beach Harbor. It's a seven inch diameter steel pipe that goes down 5,000 feet. But some wells have bigger pipe or smaller pipe and are deeper or they're shallower or they have deviations to them or they have other complications to them. And so another part of the kind of the

the analysis of these populations of wells is are they clustering, right? So there's a thousand wells that look like this, there's 10,000 wells that look like that. So can we then apply specific engineering solutions to each of those clusters, as opposed to starting with the one and going, oh, okay, we did that one, go to the next one and go, oh, let's do a new design. So it all helps by the efficiency of the system by being able to have some more cookie cutter types of solutions.

to these populations of wells. And that's another big part of trying to understand how these wells are built, where they are, what's their life history been. And it's just a lot of data when you start to think about there's 20 or 30 parameters over a million wells. And how do you understand all that to build a business out of it?

Mark Hinaman (25:44)
Now I assume the well clustering you want so that you can group the heat transferred to and then off of the material and have presumably a single or minimum number of balance of plants, meaning like build one power plant for five or or X number of wells. Am I thinking about that correctly?

Wes Cruver (26:05)
Yeah, that's, mean, there's still a modeling to be done, but that's the idea.

Bill Bartling (26:06)
Thank

Mark Hinaman (26:09)
Yeah.

Bill Bartling (26:10)
There's a lot of places where wells were drilled very close together intentionally. And then there's a lot of places where wells were, who cares? just go out to Midland Odessa and most of that area out there, you just go put a well down wherever you feel like it, because there's nothing there. But in places like downtown Los Angeles or on offshore platform or an offshore drilling island and more and more, in some of these

multi well drilling pads for some of the oil shale and gas shale production. Those wells are placed very, very close to each other to minimize the surface footprint. So you could have 100 wells or 200 wells, very close together. And one of the ideas is that let's say if you have 300 or 300, you have 400 wells there and 100 of them have a hit end of life, you can convert all of those 100 wells to

to the thermal energy storage system with a single power plant. All so now you have a lot more storage and you can get more megawatts out of it in a longer duration and you can actually power the rest of the facility with that power. All so the operators like this because there's one operator we're talking to and they have a drilling island where they're paying millions of dollars a year in electricity charges to the provider there.

And they've got all these walls they have to abandon. We're saying, why don't we abandon a hundred of them for you to convert them, the battery, we can cut your electricity costs in half or more. And now you have, you're not dependent on the grid and you have this zero emission system and you can still power your facility for as long as you need to go. So those kinds of ideas are very interesting to, again, the regulators, to the city, to the population and to the operators of taking something that's asset.

and converting it, giving it a new life into something which is really important.

Mark Hinaman (28:02)
Is there, I know there's lots of variability, right? mean, you mentioned it, the diameter of the casing or the steel pipe and then the depth and how deep that goes, right? Those two factors influence.

the total volume and how much material you're going to be able to hold. Is there kind of a rule of thumb that you guys have for a power rating or a total energy storage, right? How many watt hours per foot or watt hours per cubic foot of material? How do we think about that?

Wes Cruver (28:29)
So,

Bill Bartling (28:31)
So yeah, but let me put it in the kind more general terms in kilowatts per foot. We're ideally looking at just using 500 to 1,000 feet at the top of these wells. And there's a lot of good reasons for that. One is it's operationally easier to deal with just that shorter distance, trying to string a one-inch stainless steel tube.

5,000 feet down in the well, which is deviated, you know, it's a little more challenging We're just dropping it 500 feet into the seven inch hole The cost of the materials are going to be cheaper for with a smaller volume and if you can get sufficient Energy out of a thousand feet which is going to be in the hundreds of kilowatt hours Then if you can then multiplex that across ten wells, you know, you can get more wells now you're getting into the megawatts so

But that's right now, that's the scale that we're looking at, but that's not really even counting fully the new developments. There's three new developments that we're looking at. One is this new revolution in phase change materials going from two phases to three phases. Another one is looking at the power plants. And then the other one is the inclusion of high temperature reversible heat pumps.

So the heat pumps will do two things for us. One is it can increase the temperature of the fluid going into the well during this charge cycle so we can get things hotter in there. But the other thing is that when the fluids are coming back out of the well before it goes to the power plant, it can also increase the heat there. And so heat and power are correlatable, right? So more heat gives you more kilowatts. So we can actually amp up the temperatures coming out of the well at 150 degrees C.

we can top that up using the heat pump and actually add more heat to the binary power plant as well. So those are new systems that are coming online. We've been talking to people about which are really important in the optimization. But I think he can also see by looking at how much research is going into these three elements that you can only expect it to get better. So people are coming up with the new discoveries all the time on how to do these things more and more efficiently.

A lot of it's around buildings, right? So trying to figure out how to insulate buildings using phase change materials and how you keep buildings cool and warm, right? Respectively using these kinds of devices and machines and substances. How do you power them? How do you harvest them? So there's so much research going on around energy conservation with that. That's driving a of the things that we're going to take advantage of too.

Mark Hinaman (31:10)
Man, I guess I'd never even thought about that. Having a phase change material that just sits at 68 Fahrenheit, right? And just you put in the installation in your walls.

Bill Bartling (31:19)
Yeah.

Mark Hinaman (31:22)
That'd be pretty neat. In the subsurface, is there specific jewelry that's going to go into these old wells, like a heat exchanger? I mean, you mentioned what sounded like a capillary tubing, like a little one-inch tube or less. Or are you just dumping these phase change materials into the hole and they just sit there?

Bill Bartling (31:40)
Yeah, so it's one of the beauties of this system is there's a lot of ideas on how to repurpose oil wells and most of them involve injecting and producing fluids from the subsurface or having mechanical systems that go to try to generate energy. This has no moving parts, it has no injection and no production of fluids. So basically what it is, mechanically what you do is you lower this YouTube into the well.

To the depth you want to have it and then you add the phase change materials and if it's pure phase change material The one of the advantages of doing that is if you heat it up it turns into a liquid, right? So becomes very easy to deploy, you know as a liquid as opposed to as opposed to sand grains or something like that So to those parts are are simple There's a couple of pieces are going into the well. One of the most most important ones is a fiber optic cable

So the fiber optics could do, I don't know how familiar you are with fiber optics, there's three important data channels that fiber optics can give you. One is temperature. So obviously knowing how this battery is charging and discharging in four dimensions, spatially and temporally, gives you information about the performance of the system and how do you optimize it. The second one is acoustic. So if the...

if the heat exchanger, which has high pressure, high temperature water, it springs a leak, it'll start to hiss and we'll hear it. So we have early intervention and it's water, environmentally it's not going to be an issue, but that will ruin the performance of your battery if you actually have a leak in your heat exchanger system. But additionally, Wes talked about carbon credits and one of the ways you can maximize the value of carbon credits

is by being able to document that in fact you are successfully sealed off the methane. So methane does start to leak back into the well bore. It'll bubble through the fluid and we'll hear that with the acoustic channel as well. So we'll have real time and full time verification that the methane has been sealed off permanently. And the last channel is pressure. And so this may not seem obvious why that's important, but when the phase change materials go from a little...

Mark Hinaman (33:48)
you

Bill Bartling (34:03)
solid to a liquid they increase in volume a little bit. And so now you have a pressure field that's going to be expressed throughout that wellbore too. And this is to be non-uniform and non-linear throughout that wellbore spatially and temporally. So not only do you now see how the heat is being distributed around, but you also now see what's the status of the phase change materials? Are they solid, they liquid, are they transitional? Where and how and when and how is that thing going back and forth between liquid and solid?

So all those are really important measurements for understanding how the system works or go optimizing it and maximizing power and minimizing cost.

Mark Hinaman (34:40)
That's pretty cool. Yeah, but the YouTube that is installed is that your thermal power fluid that you would pump through that YouTube then or allow it to flow through or Just water. Okay, so you're pumping water through so water gets hot as it heats up and from the solar panels on the surface Flows down through the YouTube down one side back up the other

Bill Bartling (34:48)
Yeah, that's water.

Yeah.

Mark Hinaman (35:02)
and heats up the material. And then you've got your fiber optic cable, presumably on the outside, measures at depth, thermal, acoustic, and pressure.

Bill Bartling (35:12)
Yeah. Yeah. And then the other fluid, the working fluid in the power plant, right, which is where you go, this is binary transfer system. got the hot water here and you got the power plant fluids in here. That's some other exotic fluid. That's also a closed system, right? And so the original one we looked at was called an organic Rankine cycle engine. And it's the one that has a couple of million commercial hours on it. It only has about an eight to 12 % efficiency rate. So we're finding that that's not going to be

Wes Cruver (35:36)
.

Bill Bartling (35:41)
anywhere close to good enough. So that's what I've been looking at other things using supercritical CO2, helium, hydrogen as working fluids in the power plant. Much more efficient getting up in the 60, 70 % efficiency rate and then adding the heat pump onto that system. We can jack up the temperature of that water as it

leaves the well bore and goes to the power plant, giving it a lot more energy for the power plant.

Mark Hinaman (36:07)
Got it. So each hole, you anticipate on the hundreds of kilowatt hours scale. And so yeah, you're powering it with solar. I felt skeptical, but the Huntington Beach example or LA middle of LA makes sense. You've got a bunch of solar panels everywhere. And if you can plug in and tie into some excess.

power shave off some super cheap peaking power, then yeah, it could be a good application.

Bill Bartling (36:33)
Yeah, yeah, and that's a great example. another one we're looking at is drilling islands in the Port of Long Beach, And those are self-contained drilling and producing systems. And that would be a great place to do an installation like this. Urban environments there. In Los Angeles, there's buildings on Wilshire Boulevard that look like insurance companies or something. It's actually at Hoyle Field, right? It's completely contained.

It's got all the facade in the front with windows and things looks like it's an office building. It's not. It's actually an oil field.

Mark Hinaman (36:59)
Yeah.

Functioning pump jacks in the middle of roundabouts. Yeah.

Bill Bartling (37:05)
That's right. Yeah, yeah.

And those things, as they reach their end of life, you could convert that entire facility into a power plant. Things like there used to be an oil field behind the bleachers at the Beverly Hills High School. And they've now abandoned most of those wells. But you think about all of these urban environments where the wells are pretty closely contained because they didn't want to disrupt too much real estate for it. And they're all reaching end of life. And they're all.

being regulated out of business, those are all great candidates for doing this conversion.

Mark Hinaman (37:33)
Yeah. Is the salt pretty corrosive? mean, I feel like, or are you going to have to coat the tubing or coat the casing before you put the salt in?

Bill Bartling (37:43)
The ideally no, and the reason that you feel good about that is people think of oil wells, know, it's like Beverly Hillbillies, right? And the oil just coming out of the ground is all oil. At the end of life of an oil well is mostly salt water. And it's mostly the salts that we're gonna be using are in the brines that come out of the subsurface with the oil. So a lot of these wells, they might be producing

100 barrels of fluid a day and one barrel that's oil and 99 barrels of water so it's basically the same kind of same kind of fluids that that well is used to used to being in and by the way even not just the inside of the well bore but the outside of the well bore right so the geological formations that that this was drilled into contain basically ancient seawater right it says 30 to 40 thousand parts per million salt water

same kind of stuff you would go out and see out in the ocean. I these are old oceans or 20 million year old oceans, but they got captured in these rocks. So it's very, salty water. So it's the same stuff. So these wells were built in order to withstand that that kind of stuff. But having said that, wells do undergo corrosion. And this is also part of the stuff that Wes is doing and interrogating the life history of these wells, because there's a lot of wells that we're to get in and go, you know.

think so. It has too much corrosion, has holes in the casing, it's had a bad history. We're just gonna say this is not a candidate for doing this. So a big part of the data analysis for the wells, the wells history, the databases is answering that question too. What's happened to this well over its history? Because in California the data is pretty good about the permits that were given to the well. If you've got a

Mark Hinaman (39:15)
Okay.

Bill Bartling (39:35)
well that's had

you know, six remediation permits for sealing holes and stuff, you know, so you don't want that one, right? Because that's going to be just a problem down the road.

Mark Hinaman (39:44)
But those

are more expensive to plug, Those are going be more expensive to take care of,

Bill Bartling (39:50)
They are, and that's not our problem. Those problems are So what you don't want to do is create another problem. So when you're going to be adding heat and having all these heat-cool cycles over daily over the next 20 years, you want to start out with something which is competent.

Wes Cruver (39:52)
Check.

Mark Hinaman (39:53)
Yep.

Bill Bartling (40:11)
So that's actually a big part of the filtering system. as Wes said, the databases range from being pretty good to horrible. And so that's part of it.

Mark Hinaman (40:19)
Yeah, so I

guess, Wes, what kind of fields or variables are you looking for then to try and find some of these optimum candidates?

Wes Cruver (40:28)
We're still going through that. really big portion of this is just identifying.

how legitimate the data is, which we can start to get an idea of as we see more and more of this data. we've obviously looked at this point. I'm in Los Angeles. We've done a lot of work here, specifically in California. So we have all of the data around that. But as we start to look into places like Oklahoma, where all of the records are still on paper, some of this will also depend just in terms of,

Some states are recording different types of information. Some carbon credits require different metrics to verify how much or to predict how much methane could be leaking. All of these. So, you know, some are based on historical production data, which at that point you get into much larger data sets. And some are just sort of spatially how large is the basin or the reservoir. So it kind of depends based on the state.

But ultimately, a lot of it is just sort of filtering out, you know, a lot of these wells that, you know, from the California database that show up somewhere in Hawaii or, you know, on the North Pole because they just didn't have the, you know, somebody fat fingered a coordinate. And some of it's going to be based around some data imputation, obviously.

some operators are better at keeping up these records than others. And so as our data set grows, we'll have more more confidence in predicting what some of these missing values could be, as well as detecting the outliers. And so yeah, the whole cleaning of the data ends up being one of the biggest tasks of this project.

Mark Hinaman (42:00)
I completely understand. I've done that job a lot with specifically that those kinds of data sets. So,

Bill Bartling (42:03)
you

Mark Hinaman (42:06)
What's the timeline like for this guys? mean, it sounded like digital twin and the next year or two. then yeah, lab scale, bench scale facility and then prototype deployment tests, presumably some of your digital twin optimizations. But what are you guys thinking for timeline to go through those steps?

Bill Bartling (42:25)
So the first phase of all this is really kind of...

getting the idea into a form where it makes sense to communicate. The next phase of that is fundraising. And so we've got a kind of two-pronged fundraising program going on right now, which is composed of grants from the Department of Energy. Things are changing in Washington, DC. We think that the project that we are promoting is something that will still be in favor, even with the new administration.

Mark Hinaman (42:35)
Mm-hmm.

.

Wes Cruver (42:48)
you

Bill Bartling (42:59)
But on the side is a commercial fundraising effort to try to get investors to come in and help us take this to commercial scale very quickly. So the digital twin modeling really shouldn't take more than half a year, maybe a year. But certainly we've started getting sufficient results from that. In time to start doing, we're hoping the prototype within six months, I start putting

putting something in the laboratory where we've already asked the questions of what's the most important information we need to gather in that prototype system. So it'd be nice to know that going in so we can design it appropriately and minimize that effort. That should also be a very short duration project. And then we go into the demonstration or pilot program. And we have a number of locations already lined up with

with people in Louisiana and California where we would want to do this pilot program. And given the right circumstances, for instance, let's say a drilling island offshore, could very quickly morph into a commercial installation in a successful pilot program. Some of the things we want to learn from all of this. So you have a well that's 7,000 feet deep, 5,000 feet deep.

Can you actually use all 5,000 feet of that well? And could you get megawatts out of a single well then by doing that? Or is it operationally too difficult, right? Or is it too expensive? Those are a lot of the questions that still need to be answered going down the road. on the timeline, once funding's in place, I think we're less than a year from the pilot program. so we've got a number of universities already lined up to...

Wes Cruver (44:41)
.

Bill Bartling (44:48)
to do all the modeling work as well as one national laboratory. So they're computationally ready to go. We've got, as I said, wells in a couple of states where we can occupy them to do a demonstration project and a pilot project. And

we have multiple places where we can do the prototype. So several places have offered us facilities and work and science and so forth to do all those things.

So once we get the funding coming in, which we think is in pretty short order, we think we're probably less than a year from the pilot. And then from there, probably less than a half a year commercialization, given that everything goes well.

Mark Hinaman (45:29)
That's fantastic. Yeah, very, very exciting. I won't hold you to it, but surely you guys have kind of a target power price or costs that you expect to try and hit. How are you thinking about, yeah, cost and total value, value add?

Bill Bartling (45:44)
Super important question. And that's why the business model that we're working on is to not get into Queso or into the big grid, because it could be two or three cents per kilowatt hour, but to work behind the meter. So given an example, one of these oil-filled operations, they're paying $4 million a year to the power company. And if we could replace their wells, convert their wells to this power system, it cut the cost in half.

So we'll be getting 20, 30 cents a kilowatt hour, 10x what the power company would give us. And we're cutting the operative costs by a lot too. The cost of converting the wells, the abandonment process is actually not tremendously expensive. That ranges from 50 to $150,000. The materials for...

Mark Hinaman (46:19)
you

Bill Bartling (46:36)
for the phase change depends on what we use. Some are going to be a lot more expensive than others. And the power plants, more efficient is more expensive. that's, know, Wes talked about that kind of the idea of getting a lot of wells together with a lot of heat storage with a single power plant. We can then leverage that cost, that expensive piece of equipment across a lot of wells as opposed to having one power plant for every well.

So all those things will affect the economics. Like all things startup, it gets cheaper after you get going, right? mean, that's kind of the nature of experimentation and discovery is you get an idea, you use the best things you have, you learn a lot, and especially learn a lot about cost and efficiency.

Mark Hinaman (47:21)
Yeah, that's awesome. How big is the team, Bill? Can you share how many folks you got helping out with this digital twin?

Bill Bartling (47:28)
you got most of them right here. So we have two.

Mark Hinaman (47:31)
There you go.

Bill Bartling (47:34)
scientists, one of them is in Spain, and he is a postdoc researcher in advanced solar systems. We have another recent PhD at the University of North Dakota, and she just finished her degree in geophysics, specifically in geothermal exploration. The universities, they're kind of self-staffed. They're just going to be getting grant money from us or from the government.

Wes Cruver (47:53)
I can't you.

Bill Bartling (47:58)
The one in particular we're working with is in San Diego and they have a program where, you know, they combine the research with students for PhD candidates. And so there's a big value there to the university for bringing difficult projects in that are computationally hard. And they go through their system. They have very advanced mathematical systems that they use. They have access to supercomputing centers. So they have Big Iron to compute on.

Mark Hinaman (48:15)
.

Bill Bartling (48:28)
So that's who knows how many people that's gonna be in the university. That's kind of up to them and that was one of the we or the government gives them. But as far as our company right now, there's really only five or six of us

that are really spending a lot of time on this right now. So it's small. That's how all startup is small.

Mark Hinaman (48:47)
Lean and mean, I love it. Well, Bill West, give us your most optimistic view of the future. If all this kicks off and gets heading in the direction that you want to be going, what's it look like in five to 10 years?

Bill Bartling (48:49)
Yeah.

Wes Cruver (49:01)
Want to go first, Bill?

Bill Bartling (49:03)
Sure, so there are a million operating oil wells in United States today according to the EIA. These are not the idle wells. These are ones that are active. 75 % of those wells are what they call marginal producers. They define a marginal producer as less than 15 barrels of oil per day. All of the oil operators I know would think that was a great well. But still, that's what the EIA is classifying.

A lot of those wells are actually under five barrels of oil per day. So the point of that is that there's going to be a lot of wells, hundreds of thousands of wells that are gonna be reaching end of life, probably in the near term. Especially if there is a large surge of new production in the market. What that does is lower oil price. The administration has been adamant that that's what they wanna do is cut cost of energy, the price of the well.

the gas pump. By doing that, that takes a lot of these these marginally producing wells and makes them sub-economic which puts them into our wheelhouse where we can then convert them. So there's a huge market in place today and that market doesn't go away. It always gets bigger because more wells being drilled, more wells going to end of life every year. So that's kind of the big picture that there's a tremendous market opportunity for this in the success case.

Mark Hinaman (49:59)
you

Wes Cruver (50:24)
And there's a impending energy crisis with AI tools coming out. I know it's one or 2%, I think, of all of the energy usage right now, but most experts are saying it's going to start doubling every nine months or so. There's a lot of demand for this and with companies like Google and Microsoft being some of the larger deployers of these AI technologies, they're consuming all this energy and they have a lot of

Mark Hinaman (50:24)
All right.

Wes Cruver (50:53)
lofty goals to become carbon net neutral by 2030. And so there's going to be a lot of demand on storage of these renewables and the long duration energy storage. so we're pretty excited about how we can help in that.

Mark Hinaman (51:13)
I love it. I'm rooting for you guys. Yeah, I'm excited to see where he goes. Bill and Wes with Geo2Wats. This has been fantastic. Thanks so much for joining me today.

Bill Bartling (51:23)
Pleasure, Mark. Thank you.

Wes Cruver (51:23)
Thank you.

‍