Article 5JMBF Until We Get Rid of Fossil Fuels, Can Data Make Them More Efficient?

Until We Get Rid of Fossil Fuels, Can Data Make Them More Efficient?

by
Steven Cherry
from IEEE Spectrum on (#5JMBF)

Steven Cherry Hi, this is Steven Cherry, for Radio Spectrum.

A few months ago, we had on the show an economist who specialized in the energy sector. She noted that while the Trump administration had put drilling rights to the Alaska Natural Wildlife Refuge, or ANWAR, up for bid, there wasn't much interest from the oil industry. The Arctic, and cold climes more generally, presented logistical-and therefore financial-problems for oil companies.

To be sure, there's been drilling in the frigid North Sea for decades, but that doesn't mean it's been easy. For example, at British Petroleum's Valhall oil field in the Norwegian sector of the North Sea, drilling began in 1982, and the company is still pulling 8000 barrels per day, but losses are considerable-or have been until BP began working with a data science company. Yes, a data science company.

Further out, in the middle of the North Sea, another set of BP oil fields, known as Alvheim, has been rediscovered to have greater reserves than previously thought. There, the same data science company optimized a calibration process and in so doing reduced production losses and saved BP a lot money.

More Signal, Less Noise:
The Radio Spectrum Podcast With Steven Cherry

Sign up to be alerted when
the next episode drops.

The data science company's work isn't limited to oil and gas.

For example, it recently won a research contract with the California Energy Commission to use modeling, and data analytics, to help it improve production efficiencies in wind energy.

The data science company is called Cognite, and my guest today is its Senior Director in charge of Energy Industry Transformation, Carolina Torres.

Carolina, welcome to the podcast.

Carolina Torres Hi, thanks for having me.

Steven Cherry Let's start with the Valhall oil field. Do I understand correctly that you added no new sensors and instead used a better physical modeling of the fluid flow and data analytics? Tell us about that work.

Carolina Torres That's correct. We didn't add any sensors downhole or anything like that for that. One of the issues that we have and that we've had in the Valhall field ... You can think of it as a reservoir collapse. It's a chalk reservoir, which is very unstable. And if you pull on it too hard or if anything changes in the subsurface, that chalk just collapses and comes into the wellbore and plugs everything up. It's very expensive to mitigate. Sometimes you lose the entire well and have to do a sidetrack or do something different. And so what we've been able to do is use data science to monitor the production itself and how the well is behaving in order to forecast or be able to predict when things might be becoming unstable downhole.

Steven Cherry And that's a general strategy-better modeling and data analytics are being used in the California wind energy research. And that's with another data company, Aker. Solutions?

Carolina Torres Actually, Valhall is now also Aker BP; Aker and BP have combined several of their assets in the North Sea. And Aker Solutions is another company, another Aker company. And correct, we have partnered with Aker Solutions, a wind branch of Aker solutions, on a grant that we received from the California Energy Commission to try to digitize and modernize the offshore wind industry. And again, it's using data science-machine learning and artificial intelligence and other optimization algorithms-to optimize wind power, offshore wind park production. And it's not just the actual production, but we can also optimize maintenance and we can also do remote sensing of wildlife and marine life to understand migration patterns and things like that that might be affected by the offshore wind parks themselves.

Steven Cherry And I gather a key element of this modeling is what Cognite calls a digital twin. What is that?

Carolina Torres So a digital twin, you can think of it as a model, like an airplane model, a replica of what is actually out there. But the replica isn't just like a replica of a physical model of it. It's actually a data model of it. What that means is you can have a 3-D visualization of, say, a turbine or the entire park itself with all the different parts and equipment that is on there. If there are sensors out there, you would have that data in there. If it's out in the northern coast of California, you can also include data from the weather, the currents, the temperature of that. So it basically combines the entire data set that is available for that entity into a single place which can then be queried, analyzed, do trend analytics, do machine learning, do AI with it.

Steven Cherry So interestingly, there's going to apparently be a big push to do a lot of offshore wind in the Northeast. To what extent is work like this transferrable to another situation like that and how much of it has to be redone? And in addition, since you're not adding new sensors, would you have like a wish list of sensor data when you create a project like that?

Carolina Torres Yeah, absolutely. If it's a ... You know, one of the things you mentioned in the intro was what are some of the challenges? And if you think about oil and gas, we have very often what we call a brownfield problem, which is, you know, we've got an existing facility; maybe it's been there since 1985 and we've got a lot of analog gauges and sensors. And so we have to retrofit if we want to have real time readings on equipment, for example.

When you talk about renewables, like installing a new wind park off the coast of Massachusetts, it's a new install. And that's a greenfield problem, which in some ways can be significantly easier because you can ... First of all, most manufacturers of different types of equipment, heavy industry equipment, already have sensors built in. So, you know, any new wind turbines that are being built offshore will have a lot of different sensors already built in,

Steven Cherry So you mentioned before that there's Cognite, there's Aker BP, there's Aker Solutions, and you've explained sort of the relationships there. You came from BP and so you're part of that very glue, so to speak.

Carolina Torres My background is actually, not in production and operations; my background is in geoscience. I'm a geologist by training and most of my career at BP was spent in drilling and wells and geology, subsurface work. But my last 10 years there, it was really more around strategy, change management, and digital transformation, particularly around subsurface wells. And I became familiar with Cognite through that effort that we were doing to digitize and modernize and improve and automate some of our drilling operations.

Steven Cherry There's also a broader effort at Cognite to help Norway digitize its energy grid.

Carolina Torres Yes, that's correct. One of our customers is Statnett, which is the national grid operator for Norway. There's quite a lot of optimization that can happen and is needed actually for running grids. We have multiple data sources coming from multiple energy sources, potentially. We have weather and demand load that needs to be tracked and managed and that's real-time data. And then we need distribution systems and the equipment that allows us to turn things on and off so that the energy can be distributed. So there's a lot of different sources of information that need to be brought together fused, if you will, contextualized. And then there's a lot of opportunity for optimization there.

Steven Cherry I'm curious that you mentioned weather. To what extent do you rely on outside weather data-Weather.com or NOAA or whoever-not only for data but for prediction? To what extent do you do your own predicting?

Carolina Torres So we talked about the offshore of California Energy Commission, the offshore project that we're doing there, and we do take in existing forecasting. There's really amazing and great forecasting that's going on. It's very, very sophisticated using multiple ensemble models, for example, to get very statistically significant, longer-ranging weather forecasts. If you think about a wind ... running a wind park, however, operators need to supply a very, very accurate 72-hour forecast continuously to the grid. And if they underdelivered or over-deliver on that, there are actually penalties, and those penalties-because wind can be quite inconsistent-those penalties can be quite significant for operators. So maybe 10 to 15 percent of the revenue might get hit because of those penalties. So it's really important that we have an extremely accurate 72-hour forecast.

Now, the weather is one thing that might impact the production of energy from a wind farm, but also your equipment and the status of your equipment can impact it and currents can impact it. And so what we do is we combine the forecasts about the weather, but we also generate our own forecasts about equipment uptime, equipment efficiency, and the actual productivity of the wind farm itself, the different combination of the multiple turbines, for example. So there's both things, I hope that answers your question, but we don't generate our own weather forecasts. But what we do is combine those external weather forecasts, which are quite sophisticated and combine it with forecasts with other components of the system that may impact the production forecast that we give to the grid.

Steven Cherry You mentioned your time at BP, most recently in its digital transformation area. BP has a goal of being net-zero by 2050. On paper, it's one of the most aggressive plans towards sustainability in the entire oil and gas sector. How real is the zero' part of net-zero? Are there a bunch of exceptions? Or is this a real and difficult goal?

Carolina Torres I think it's a real goal and I also think it's a very difficult goal. And the reason it's very difficult is they are looking at three different levels of getting to net-zero, which also includes their supply chain on the input side and the customers on the output side. It's called the Tier Three. And in order to do that, there needs to be an enormous amount of collaboration within the industry. No company is an island. No company is going to be able to achieve net-zero by themselves. The only way that we can achieve-that any one company can achieve-net-zero is to actually collaborate with all of their vendor partners and with, you know, and their downstream and upstream partners in this. So it is real. It is possible. I really believe it's possible, but it's going to be hard.

Steven Cherry I gather that in effect more and more of Cognite's work would count as efforts toward sustainability, not just at BP, but in California and with many other clients as well. Some of those are in the manufacturing area. Can you speak to any sustainability-enhancing efforts there?

Carolina Torres So one of the things that I've discovered and starting to learn and research about this is that historically a lot of heavy industries have seen sustainability as something that ... a chore that has to get done. Oh, you know, we have to comply and we have to keep regulators happy. And they are not looking at it as an opportunity. But I think people are starting to change their mind about that because they're realizing ... Most of our customers started off wanting to use our products for optimizing their production, for reducing their maintenance costs, for trying to work remotely. But what they found is, as they've done that and as they have allowed us to fuse and contextualize their data so that they can do that, once the data's in there, it's not very hard to actually look at it and say, well, how has this impacted our fuel consumption? How has this impacted our materials usage? How has this impacted the number of flights we have to take off shore or the number of people that we have offshore that require vessels for supplies?

So every piece of efficiency that we gain, which translates to dollars or incremental production, also translates to a lower impact on the environment. Pretty much. It's almost difficult to find an example where that's not the case. So you asked about manufacturing in particular. So we have an example of a manufacturing company that's one of our customers that wanted to hire us to help them optimize and automate their manufacturing, their tooling. So the machining of different tools, what they do is they you know, they create equipment, pieces for machinery, for heavy industry. And they have something like 2200 different machining tools that they use for cutting different shapes into the metal. And what they found after they installed our maintenance optimization software. And the data platform underneath that is that they were actually building a lot more of these machine tools that they needed to-they had stuff in inventory that they could have used that they didn't know they had. And so there was a significant reduction that, you know, they reduced their machining tooling by 80 percent. But that also translated into a lot less wasted metal resource that they were overbuilding.

Steven Cherry And sounds like they would also save some time instead of swapping out equipment to use equipment that's already out.

Carolina Torres Correct. Time and effort and people and energy to run their factory. We haven't calculated every single last bit of CO2 or solid waste that was saved. But, you know, you could see that it would be quite significant if you've actually reduced your machining and tooling by 80 percent.

Steven Cherry There seems to be a general sense that the environmental movement over the last 50 or 60 years and and sustainability efforts more recently and certainly moves toward reducing global warming all harm the business world in general and certainly the oil and gas sector specifically. It's never been obvious to me that that it's been anything but a net plus, that is to say, concern for environment and sustainability. I'm curious what your impression is.

Carolina Torres I mean, my philosophy is we should all strive to leave no trace and to make the world better for ourselves and for everyone. And that includes, you know, animals and the ocean and the air and everything else. So I think that anything that gets us gains in that space is a good thing and how we should all live our lives. We shouldn't treat our resources as though they are a bottomless pit for us to exploit. And we shouldn't treat each other that way and we shouldn't treat the ocean that way.

Steven Cherry It seems, if I understand you correctly, to improve efficiency and to improve sustainability are really almost two sides of the same coin. If that's the case, then improved efforts towards sustainability would generally be efforts toward greater efficiency. And that's going to only save a company money or make a company money. Is that is that a fair assessment?

Carolina Torres I don't know if this will answer your question, but I believe that Fourth Industrial Revolution technology, which is IoT, data science, remote work, robotics, a whole long list of things, can actually help us to transition to new energy. And that could mean doing what we're currently doing right now around hydrocarbons in a much more efficient, efficient and less impactful way and also transitioning to new types of energy that are going to be even less impactful. So, yeah, does that answer your question?

Steven Cherry I think it does, yeah. I guess the reason I've been asking is that more and more this podcast series has come to be about sustainability more than any other one thing. And that doesn't seem like such a bad thing, since it's the central problem facing humankind these days. Cognize bills itself as an industrial data company. The more I hear about it, the more I think of it as an efficiency company, a waste reduction company, and in some ways an industrial sustainability company. I hope that's a fair way to look at it.

Carolina Torres I do. I think that one thing that I would add to that is that-this is a particular love of mine-I think we need to and I mentioned it earlier, it's around having to be a lot more collaborative within a company, across an industry, of multiple companies, and even across an entire value chain. So multiple industries working in concert to deliver energy to the world. We need to be thinking about circular economies. We need to be thinking about how do we reduce the impact of an entire value chain. And we can't do it without the data, without contextualizing and fusing the data. It's impossible. So I see the data fusion and contextualization across these ... within a company, across multiple companies, within an industry and across an entire value chain of multiple industries as being a necessary first step for us to be able to truly get to zero.

Steven Cherry Well, that is the goal and I thank you and Cognite for its work in that direction and thank you for joining us today.

Carolina Torres You're very welcome. Thank you for having me.

Steven Cherry We've been speaking with Carolina Torres, who drives Energy Industry Transformation at the data startup, Cognite.

Radio Spectrum is brought to you by IEEE Spectrum, the member magazine of the Institute of Electrical and Electronic Engineers, a professional organization dedicated to advancing technology for the benefit of humanity.

This interview was recorded April 2, 2021 using Zoom and Adobe Audition. Our theme music is by Chad Crouch.

You can subscribe to Radio Spectrum on Spotify, Stitcher, Apple, Google, and wherever else you get your podcasts, or listen on the Spectrum website, which also contains transcripts of this and all our past episodes. We welcome your feedback on the web or in social media.

For Radio Spectrum, I'm Steven Cherry.

Note: Transcripts are created for the convenience of our readers and listeners. The authoritative record of IEEE Spectrum's audio programming is the audio version.

We welcome your comments on Twitter (@RadioSpectrum1 and @IEEESpectrum) and Facebook.

KxI27qPiHf8
External Content
Source RSS or Atom Feed
Feed Location http://feeds.feedburner.com/IeeeSpectrum
Feed Title IEEE Spectrum
Feed Link https://spectrum.ieee.org/
Reply 0 comments