Artificial Intelligence
From machine learning to computer vision, deep learning to virtual assistants and autonomous vehicles to robotics, Shell has been focused on a range of technologies that have supported advances in AI.
Businesses that harness new data sources and use AI and machine-learning technology to provide insights will be in a strong position to shape future commercial development and influence how society changes. Many of the algorithms behind AI and machine-learning systems are not new but limited volumes of accessible data have hampered their application. The recent explosion in data volumes and availability has led to a step change in the training of algorithms and provided important new insights: easy access to vast data volumes is making AI algorithms smarter.
The Energy Podcast: Can AI get the world to net zero faster
How is the world meeting growing energy demand while reducing carbon dioxide emissions? From renewable energy to the world of AI, join us on The Energy Podcast to find out more. Listen and subscribe in your preferred podcast player below.
Artificial intelligence examples in Shell
The Data Detectives - Solving the 10-year-old problem at Perdido
Read the transcript
Read the transcript
Title: The Data Detectives - Solving the 10-year-old problem at Perdido
Duration: 05:20 minutes
Description:
Lying in the Gulf of Mexico is Shell’s Perdido Platform. It’s one of the deepest floating oil and gas platforms on the planet. This film follows how Shell must seek a new solution to help solve a decades old problem with Perdido – the periodic disruption of its pumps responsible for separating oil and gas. As the film reveals, the issue has stumped engineers for some time, so to solve the mystery Shell calls in a crack team – its data detectives.
The data detectives plan to use artificial intelligence to look for patterns in data returned from Perdido’s pumps. Each pump sends a stream of information, including temperature, pressure and chemical signatures. With AI, the detectives are looking reoccurring signatures in the data ahead of a pump failure.
After drawing an initial blank, they team up with pump operator Jimmy Johnson, who helps the team refine their computer code, which ultimately leads to a tell-tale chemical signature ahead of 70% of pump disruptions – a great win for Shell’s data detectives and AI.
[Music plays]
Rhythmic medium-tempo music swells
Shell's Perdido platform in the Gulf of Mexico.
It's one of the deepest floating oil and gas platforms on the planet, but for 10 years, Perdido has struggled with an engineering puzzle.
After this long to not have a resolution this supremely frustrated.
It's subsea pumps are mysteriously disrupted without warning.
Now, Shell are turning to a new solution to help solve the problem, artificial intelligence.
8000 feet below the waves, lie Perdido pumps.
Key to production, they separate a mixture of oil and gas that's sent to the surface.
But occasionally the pumps stop separating, causing a complete loss of production.
It causes a lot of headache so we can solve this problem, then we can be more proactive in trying to stop the event before it happens.
At these depths, inspecting inside the pumps is impractical.
Shell must find a different solution.
They're calling in a crack team.
Their mission, to predict partidos pump problems using artificial intelligence.
Meet the data detectives.
I think what makes me a good detective is always being skeptical of the assumptions.
We want to do well.
We want to find a possible solution through data science if we have it.
Each pump generates a constant stream of data, recording its temperature, pressure and chemical mix.
The data detectives plan to apply artificial intelligence to sift through the gigabytes of stored data to search for patterns which might predict when pumps fail, but it's easier said than done.
We had such a humongous dataset that it's like finding a needle in the haystack.
Three thousand tags, five years of data.
It's too much for the computer to process.
After weeks of work sifting through a decade of data, their search draws a blank.
They wanted prediction, they wanted possible causes.
Sometimes people feel that artificial intelligence is going to solve everything, and that's not really
the case.
Show me a sign that you're able to predict something for us.
It's research.
It just takes time.
But as the detective soon discover, it'll take more than time alone to solve this mystery.
We had some early wins, some early successes, and then it seemed like the progress plateaued.
There were times where I considered maybe we should stop.
And at that moment, it didn't feel like we were progressing.
In a bid to break the impasse, the detectives call on the experience of operator Jimmy Johnson.
Jimmy's worked in the oil industry for 25 years and his knowledge of Perdido’s pumps is second to none.
I said that you need to have individuals from the field looking at this with you, because we understand
what this means.
When you're saying this is out of normal.
Tapping into Jimmy's knowledge, the team refine their computer code and finally uncover what
they're looking for.
A telltale chemical signature seen in the data a few days ahead of a pump disruption.
So, Vanessa, how far in advance did we see something on this one?
It was about two days.
Two days!
Awesome.
Wow.
Is that something we can do?
Yeah, this is something that we can work with.
It wasn't until they showed off some of their results and I felt that there was a sense of breakthrough,
that they were truly on the verge of discovering something.
Continuing crunching through the data, the team's refined algorithm identifies the chemical fingerprint ahead of 70 percent of pump disruptions.
Awesome.
Good work, guys.
Helping predict when a disruption event might occur, means Perdido can avoid lost production, making even the most skeptical appreciate the power of artificial intelligence and the work of the data detectives.
It took the results kind of hitting me in the face multiple times.
You can't deny the results.
A treisman in me says Nobody's going to know this machine better myself.
The realism of that is that's not always true.
Once you start seeing the results, there's no refuting numbers.
That's the beauty about math, right?
I mean, math is math.
If we're able to tap into that, we can truly succeed and make this company a more successful place to work.
[Music]
Rhythmic up-beat music swells and ends.
Shell RechargePlus
Read the transcript
Read the transcript
Title: 5019P Shell - Case Study - EV CHARGING - ENG SUBS
Duration: 1:14 minutes
Description:
A case study showing how Shell is using digitalisation to make Electric Vehicle charging more efficient.
5019P Shell - Case Study - EV CHARGING - ENG SUBS Transcript
[Background music plays]
The Sound of Shell adaptation.
[Text displays]
Powering EV Growth
[Video footage]
Time lapse footage from a bird’s eye view of busy traffic at a city intersection.
Interview with Angie Boakes
[Title]
E Mobility Manager
[Angie Boakes]
The future shift to electric vehicles is pretty clear.
[Video footage]
Time lapse footage from a bird’s eye view of traffic on a multi-lane highway. This cuts to a closer high angle view of several lanes of traffic on a highway.
[Angie Boakes]
Shell is launching our RechargePlus solution, and it’s all about really how you optimize charging for their electric vehicle.
[Text displays]
Angie Boakes E Mobility Manager
[Video footage and animated sequence]
Mid-view footage of Angie Boakes standing in one corner of a room, speaking to an off-camera interviewer; the walls that form the background are papered in a modern-style collage wallpaper featuring fragments of images and words, and a screen displays against the wall at frame left. Low angle footage of a woman opening the door of a yellow electric vehicle and stepping out. Low angle close-up of a hand taking the EV charging plug out from the Shell banded RechargePlus charging point. Tracking footage of the women taking the charging plug over to her car and plugging it into the socket. As she does so, white line animated graphics and text appear on screen alongside the charging socket and at frame right to indicate the charging battery.
Interview with Phillip Villagomez
[Title]
Business Development Electric Mobility
[Phillip Villagomez]
Our optimization algorithm is pretty smart, that it moves the charge to a time when solar or wind are at their highest and prices are at their lowest, therefore helping the environment and the driver save money as well.
[Video footage and animated sequence]
Tracking point of view footage, as from a driver’s perspective, of the view through a car’s windscreen, seen over the dashboard. White line animated graphics display against the windscreen to indicate battery levels and usage.
[Text displays]
Phillip Villagomez Business Development Electric Mobility
[Video footage]
Close-up footage of Phillip Villagomez, speaking to an off-camera interviewer. Low angle footage of Phillip and two men seated and engaged in discussion in a meeting room, open laptops and documents on the table between them. This cuts to panning footage of the same scene, followed by a high angle pan of the table and Phillip’s hand as he writes notes on a writing tablet.
Interview with Angie Boakes continued
[Angie Boakes]
And what we need them to do is basically use our app to therefore tell us what time they need to depart and roughly how many miles or how much energy they need by the end of the day, and then we will do the rest.
[Video footage and animated sequence]
Mid-view footage of Angie Boakes standing, as before, against wall-papered walls and speaking to an off-camera interviewer. High angle tilting footage of a woman tapping on her smartphone screen while standing alongside her car. Close-up of the woman’s smartphone screen displaying the “setup charge” steps on the app, with the woman’s finger tapping on the “Confirm” button. Reverse view close-up footage of the smartphone held in the woman’s hands; white line animated graphics of the car, its battery and other statistics display at frame-left alongside the smartphone, indicating charging or energy requirements. Mid-view footage of Angie Boakes standing, as before, against wall-papered walls and speaking to an off-camera interviewer.
Interview with Dana Madrid
[Title]
RechargePlus User
[Dana Madrid]
RechargePlus has given me the freedom and the peace of mind to park, plug in…
[Video footage and animated sequence]
Tracking high angle footage of Dana Madrid driving a yellow electric vehicle past other cars parked in a parking area. High angle footage of Dana taking an EV charging plug out from a Shell banded RechargePlus charging point. Side view footage of Dana standing beside her car and plugging the charging plug into her car’s charging socket.
[Dana Madrid]
And I’m charged, and I don’t have to come back out.
[Text displays]
Dana Madrid RechargePlus User
[Video footage]
High angle close-up of Dana speaking to an off-camera interviewer, her car and others seen in the out-of-focus background.
Interview with Phillip Villagomez continued
[Phillip Villagomez]
We are helping people save money. Not only the drivers, but also the site host and building owners.
[Video footage]
Low angle view of sunlight through leaves of tree. Close-up in profile of hands on a steering wheel, as scenery passes through the driver’s window beyond. Mid view footage of Phillip Villagomez, speaking to an off-camera interviewer, seen against the background of a meeting room, a long table surrounded by chairs behind him.
[Phillip Villagomez]
If you have a bank of ten chargers, you don’t want ten drivers plugging in at the same time. This causes expensive demand charges and also puts a strain on the grid.
[Video footage]
Close-up of running command prompt white lines on a black screen. Low angle footage of two young men standing alongside two RechargePlus charging points in an office environment; a computer screen above one of the charging points displays white lines on a black screen – both men’s attention is on the smartphone held in one of the men’s hands. Close-up of hands plugging a charging plug into a socket mechanism that is on the floor. Mid view footage of Phillip Villagomez, speaking to an off-camera interviewer, seen against the background of a meeting room, a long table surrounded by chairs behind him.
Interview with Angie Boakes continued
[Angie Boakes]
We’re helping them to switch to zero emission vehicles, helping the grid to have more renewables on the system. It’s more, cleaner energy.
[Video footage]
Mid-view footage of Angie Boakes standing, as before, against wall-papered walls and speaking to an off-camera interviewer. Close-up of Angie as she speaks to the off-camera interviewer, then cutting back to the mid-view angle, as before.
[Graphic]
Centre-framed Shell Pecten on a white background. Fade to black.
Predictive Maintenance Digitalisation case study
Read the transcript
Read the transcript
Title: 5019P Shell – Case Study – Predictive Maintenance – ENG SUBS
Duration: 1:39 minutes
Description:
Shell staff talk about predictive maintenance as a way to leverage technologies to forecast issues and predict failures before they arise. Unable to find an off-the-shelf solution, they built one themselves and now have 24/7 coverage. The model is now being replicated in other facilities and offers good insight, as well as allowing engineers more time to do engineering work instead of analysing mountains of data.
5019P Shell – Case Study – Predictive Maintenance – ENG SUBS Transcript
[Background music plays]
Instrumental music with synthesised effects
[Video footage]
Upward-looking shot of plant chimneys set against a sunny blue sky. Text appears on-screen as displayed below.
[Text displays]
PREDICTIVE MAINTENANCE
[Video footage]
Low-angled shot of Claudia Zuluaga Giraldo and Sander Buhling wearing protective gear walking at a Shell site.
Interview with Sankesh Sundareshwar
[Text displays]
SANKESH SUNDARESHWAR
SENIOR PROCESS ENGINEER, SHEARWATER
Sankesh Sundareshwar
Predictive maintenance is leveraging the technologies we have today to forecast issues before they can arise.
[Video footage]
Close-up of Sankesh speaking to off-screen interviewer at an office location. Footage of Sander and Claudia and another staff member walking past a series of pipes towards a staircase. Close-up of Sander and Claudia talking while Sander points at a piece of equipment.
Interview with Arnold Hes
[Text displays]
ARNOLD HES
INSTRUMENT ENGINEERING TEAM LEAD, PERNIS
Arnold Hes
We have to keep our plants running and our biggest problem is when the plants trip due to equipment failures.
[Video footage]
Close-up of Arnold Hes sat speaking to off-screen interviewer at an interior location. Slow motion, low-angled shot of Claudia and Sander talking while Sander looks at a document he’s holding. Back to Arnold speaking to off-screen interviewer.
Arnold Hes
If we can predict that, then we can prevent these failures and save the company a lot of money.
Interview with Claudia Zuluaga Giraldo
[Text displays]
CLAUDIA ZULUAGA GIRALDO
PRINCIPAL DIGITAL PRODUCT MANAGER
Claudia Zuluaga Giraldo
The more stable operations, the more energy efficiency we are able to obtain, the less emissions we generate.
[Video footage]
Close-up of Claudia speaking to off-screen interviewer at an interior location. Shot of several people sat around a table in a conference room as a lady at the front is presenting by a flipchart. Close-up of Claudia’s hands gesticulating. Shot of Claudia presenting in the conference room, the agenda displayed on a large screen behind her. A white grid and coloured bar graph is superimposed on-screen to the right. Back to close-up of Sankesh speaking to off-screen interviewer.
Sankesh Sundareshwar
We looked outside quite extensively if anyone else had a product that we could just take off the shelf and we couldn’t find anything that useful to us, so we decided to build it.
[Video footage]
Shot of Kenneth Innes and a male colleague standing by a desk with two rows of monitors as his colleague points to one of the monitors. An out-of-focus hand points to a screen with red and green markers. Close-up of Kenneth Innes speaking to off-screen interviewer at an interior location.
Interview with Kenneth Innes
[Text displays]
KENNETH INNES
SHEARWATER MANAGER
Kenneth Innes
Before we developed these systems, our people spent a lot of time analysing mountains of data which was very labour intensive.
[Video footage]
Shot of Kenneth standing by a row of monitors listening to a male colleague sat down talking. Close-up of Kenneth nodding. Close-up of Neil Forbes speaking to off-screen interviewer at an interior location.
Interview with Neil Forbes
[Text displays]
NEIL FORBES
OPERATIONS SUPPORT, SHEARWATER
Neil Forbes
Instead of that being reactive, we’re now proactively going out fixing issues before they become a bigger problem.
[Video footage]
Shot of a walking Neil, Sankesh and a female colleague on the office floor. Shot of Neil and Sankesh with their backs to the camera as they video conference with a colleague who we can see on-screen.
[Sankesh Voiceover]
The big benefits of this programme is 24/7 coverage.
[Video footage]
Close-up of Sankesh talking. Close-up of a screen showing Sankesh and Neil standing on the work floor as they continue to video conference.
[Sankesh Voiceover]
It looks at about 300 system parameters all the time. Human beings can’t do this kind of thing.
[Video footage]
Back to shot of Kenneth and his colleague discussing by the row of monitors, the out-of-focus hand pointing to a screen with red and green markers, and the shot of Kenneth standing by a row of monitors listening to a male colleague sat down talking. Back to Sankesh speaking to off-screen interviewer.
Sankesh Sundareshwar
Now we’ve got to think globally and think about replication and replication at pace.
[Video footage]
Wide-angled shot of Neil and Sankesh standing in an office room as Sankesh points to a poster titled TAR. Close-up of Neil listening as an out-of-focus Sankesh continues talking. Close-up of Sankesh talking and pointing to the poster. Low-angled shot of a man sat at a desk with a double row of monitors. A series of screens are superimposed in white depicting bar graphs and other charts.
[Arnold Hes Voiceover]
We were able to use and replicate that model in our facilities and gave us a good insight in how our valves are behaving.
[Video footage]
Shot of two men, their backs to the camera, looking at a double row of monitors at an interior location. A series of screens are superimposed in white of further pie charts and bar graphs at the top.
[Sankesh Sundareshwar Voiceover]
This technology is going to make us better engineers because we have more time to do engineering work to prevent systems going offline.
[Video footage]
Close-up of data on a screen. An out-of-focus hands points at a cluster of coloured dots on a screen. Close-up of Arnold. Shot of a hand pointing at a screen displaying a line graph. Close-up of Sander Buhling speaking to off-screen interviewer set against a white background.
Interview with Sander Buhling
[Text displays]
SANDER BUHLING
INSTRUMENT PROJECT ENGINEER, PERNIS
Sander Buhling
I think the major change, it’s not really the prediction; it’s the scale-able part of this project.
[Video footage]
Still-frame of an electrical piece of equipment. A series of round diagrams are superimposed in white on-screen right displaying percentages of productivity, efficiency and system capacity. Zoom in on the productivity diagram as the percentage increases to 100%.
[Sander Buhling Voiceover]
You build a solution for 16 valves; it has the potential to increase for all valves within Shell.
[Video footage]
Slow-motion footage of Sander and Claudia sat at a desk as Claudia explains and gesticulates. Shot of Claudia’s handing pointing to an out-of-focus row of monitors. Slow-motion footage of Sander talking as Claudia listens. Back to close-up of Sander speaking to off-screen interviewer.
Sander Buhling
Why not, yes, use it on everything we have?
[Graphic]
Shell Pecten centred on a white background.
LNG Shipping Digitalisation case study
Read the transcript
Read the transcript
Title: 5019P Shell - Case Study - LNG Charging - ENG SUBS
Duration: 1:13 minutes
Description:
A case study showing how Shell is using digitalisation to improve the efficiency and lower the emissions of its’ fleet of LNG tankers.
5019P Shell - Case Study - LNG Charging - ENG SUBS Transcript
[Background music plays]
The Sound of Shell adaptation.
[Text displays]
Optimising Shipping
[Video footage]
Bird’s eye view of a vessel in silhouette on a vast glistening ocean.
Interview with Paul McStay
[Title]
LNG Fleet Performance Manager
[Paul McStay]
Shell has an ambition to drive down its carbon footprint. And within shipping and maritime, our vision is very much aligned with that.
[Video footage]
Panning footage of Paul McStay talking with colleagues, standing at a bank of large screens on a wall which display animated charts and data. Slow motion close up of Paul, in profile, speaking to his colleagues, one hand pointing to the screens in front of them. Front on close-up of Paul speaking to the off-camera interviewer, seen against the out-of-focus background of an office.
Interview with Samantha Lehel
[Title]
LNG Commercial Freight Operator
[Samantha Lehel]
The LNG Shipping accelerator essentially consolidates loads of information in one place.
[Text displays]
Samantha Lehel LNG Commercial Freight Operator
[Video footage]
Mid-view footage of Samantha speaking to the off-camera interviewer, seated on an oval table in a meeting room.
[Samantha Lehel]
So that instead of having to consult various different resources for each one of my vessels, I can see everything immediately.
[Video footage]
Tilting reverse view footage of Samantha seated at her desk, looking at the screens in front of her. Close-up of Samantha seated at her desk, her focus on the screens in front of her. Close-up of shaded graphics and data on a screen. Slow motion panning footage of Samantha’s desk surface, coming to rest on a close-up of her hand on her keyboard.
[Samantha Lehel]
The real value is, sort of, the time savings for myself as an operator, but also the level of accuracy and efficiency that that affords our business.
[Video footage]
Mid-view footage of Samantha speaking to the off-camera interviewer, seated at the edge of an oval table in a meeting room. Close-up of Samantha speaking to the off-camera interviewer.
Interview with Paul McStay continued
[Paul McStay]
What this allows us to do is see data live and in real time and therefore be able to proactively intervene when performance starts to drop.
[Text displays]
Paul McStay LNG Fleet Performance Manager
[Video footage]
Close-up of Paul speaking to the off-camera interviewer, seen against the out-of-focus background of an office.
Interview with Grahaeme Henderson
[Title]
VP Shipping & Maritime
[Grahaeme Henderson]
If we can improve our efficiency, we can reduce the amount of time that we’re waiting at ports…
[Video footage]
Wide angle view of a vessel in port, at first blurred and then brought into focus, seen through a wire mesh fence close-up, first seen in focus and then blurred as the vessel comes into focus. Panning footage of two vessels waiting alongside one another in port, seen against a background of blue sky.
[Grahaeme Henderson]
Then we can reduce our fuel usage. And by reducing our fuel usage, we can improve our emissions.
[Text displays]
Grahaeme Henderson VP Shipping & Maritime
[Video footage]
Mid-view footage of Grahaeme speaking to the off-camera interviewer, seated at the edge of a long table in a meeting room. Close-up of Grahaeme speaking to the off-camera interviewer.
Interview with Paul McStay continued
[Paul McStay]
This project allows us to understand where our footprint is and reduce that by setting appropriate benchmarks, and therefore reductions into the future, using data as a means to enable insights that will allow the business to deliver value by driving our vessels more efficiently and enabling us to deliver more molecules of LNG to our customers.
[Video footage]
Slow motion wide footage of Paul McStay and a colleague walking up to the bank of large screens on the wall which display animated charts and data; another colleague is seated in the foreground at a workstation lined with computer screens. This cuts to slow motion mid-view footage of the two men standing at the screens and Paul turning to look behind him as he points to the screens. Slow motion high angle close-up of the colleague seated at the workstation behind as he speaks and points forward. Slow motion low angle panning close-up of Paul looking up and speaking, one hand underneath his chin. Slow motion reverse mid angle view of the two men standing at the screens, and Paul pointing to data displayed there. Mid-view footage of Paul speaking to the off-camera interviewer, seated at the edge of a long table in an office environment. Slow motion low angle close-up footage of Paul and his colleague looking upwards, first showing his colleague in focus while Paul is blurred in the foreground, then bringing Paul’s profile into focus as his colleague blurs into the background. Fade to white.
[Graphic]
Centre-framed Shell Pecten on a white background. Fade to black.
Questions and answers
What is Artificial Intelligence?
What is Artificial Intelligence?
Artificial intelligence (AI) is a versatile technology that can bring about disruption by offering novel solutions to existing problems, as well as an indirect disruption by delivering complementary innovations.
Any process or business activity that involves monitoring, measuring and assessing can typically benefit from the implementation of AI technologies. Shell has successfully applied AI to various processes.
AI holds significant potential as an enabling tool for the developmentof new business models that facilitate the transition to lower carbon footprints. Importantly, AI serves as a crucial technology in helping to reduce the carbon footprint of our own operations, providing low carbon energy solutions to our customers and advancing the next generation of clean energy technology.
How does AI work in the energy industry?
How does AI work in the energy industry?
The world’s increasing demand for more and cleaner energy solutions necessitates significant changes in the way in which energy is provided. Digital technology, including AI, plays a crucial role in enabling this transformation. It can enhance various aspects of the energy industry, such as more efficient exploration and production, reliable manufacturing, agile trading, and personalised customer experiences.
At Shell, we recognise the imporance of embedding AI throughout our organisation to enhance effectiveness, efficiency and competitiveness. We have already implemented AI in various processes across our operations. For example AI helps us to reduce the emissions from our own operations. Shell’s process optimiser for liquefied natural gas (LNG), for example, takes information from sensors and uses AI to calculate efficient settings for equipment. It also helps us offer our customers low-carbon energy products and services. Our smart-charging algorithm helps customers to charge their electric vehicles more cheaply. AI also helps to enable process improvements, cost reductions, production increases and increased customer margins across Shell’s businesses. For example, innovations such as predictive maintenance use data and AI to help us spot problems before they become big and expensive to fix. Shell is also using AI to sort through seismic data so we can speed up the process to help find oil and gas more quickly, at a lower cost.
What is natural language processing?
What is natural language processing?
Natural Language Processing (NLP) is a subfield of Artificial Intelligence that aims to create computational models capable of automated understanding and generation of natural languages to perform tasks in a similar way to humans.
Shell generates large amounts of internal data in the form of text or speech such as technical documents, reports, contracts, and emails. This unstructured data conveys valuable insights for operational and business decisions but requires significant time and effort to analyse.
NLP provides techniques for training computers to understand and extract information from text, leading to quicker and higher-quality decision making. One example of its application is drawing insights from past safety incidents in real time to support decision making.
Shell uses cutting-edge large language models, including GPT4 through Microsoft Azure OpenAI Service and MS Copilot. Additionally, Shell has developed a generative AI platform called Shell e, which allows teams to customise, build and deploy applications based on open source and other large language models.
We are continuously investing in research and development to further tailor solutions to Shell-specific datasets and domain knowledge.
What is Reinforcement learning?
What is Reinforcement learning?
Reinforcement Learning (RL) systems improve their performance through trial and error during training, without the need for human experts to set and adjust decision rules. These systems build mathematical models from training data to infer the optimal action or policy, learning which decisions lead to the greatest anticipated future reward.
Shell has long recognised the potential of reinforcement learning for decision support and is collaborating with business and academic organisations to develop this next-generation technology.
What is explainable AI?
What is explainable AI?
The convergence of physics, chemistry and AI could be the next breakthrough in modelling technology. Also known as augmented intelligence, explainable AI involves a working partnership between people and artificial intelligence systems to enhance cognitive performance. This hybrid-augmented intelligence enables scientists and engineers to address problems that pure machine learning cannot easily classify or train artificial intelligence systems to resolve.
For Shell, augmented-intelligence-based modelling supports traditional scientific computations based on fundamental physics and chemistry. This uses hybrid models that combine elements of data- and traditional physics-based models. Hybrid models offer increased computational speeds and accuracy in predicting the behaviour of complex dynamic systems. They are playing an increasingly important role in research areas such as computational chemistry, material science, image analysis, fluid flow and reactor engineering.
How do you scale AI in the energy industry?
How do you scale AI in the energy industry?
Anyone can develop a small-scale proof of concept with a machine learning model that meets a specific local requirement. But, to make an impact and maintain it at scale, we need to develop solutions which can be deployed globally at a rapid pace.
To enable AI at scale across Shell’s businesses, we are standardising approaches and aligning on common data structures, platforms, tools and ways of working. This allows us to share best practice, code and work seamlessly in cross discipline global teams.
One of our key differentiators in the space of AI is our domain knowledge and a deep understanding of physics and chemistry which we have built over the past 100 years. Our success lies in the fact that we are integrating data with our deep knowledge of physics. This that makes our AI models explainable and thus transferrable in the wide applications across our assets.
What is Shell.ai?
What is Shell.ai?
Shell.ai is an internal change programme, formed to drive a common approach to data science platform technology, develop consistent ways of working, and build a community of practice which can demonstrates the art of the possible, and shares best practices across the entire business.
