Data and AI are as important to Shell as oil

There are several motivations for Shell to revolutionise their company via the use of AI and data.

The oil and gas business is at a crossroads because to rising energy needs, disconnected environments, and pressure to combat climate change. Shell and other energy firms have a choice between maintaining the current quo and embracing a low-carbon energy future.

End-to-end processes must be optimised and kept up at scale as we move toward a more dispersed, diversified, and decentralised energy system. Therefore, it is essential to find solutions that can be quickly and universally implemented. Shell was forced to transform into an AI-driven technology business as a result.

Speeding up the digital transition

For instance, in order to hasten the digital transformation of the energy sector, Shell, Baker Hughes, Microsoft, and corporate AI firm C3 AI formed the Open AI Energy Initiative (OAI) in November.

The OAI provides business leaders with the chance to work together in an open, equitable, and transparent manner, according to Dan Jeavons, vice president of computational science and digital innovation at Shell. It enables them to develop AI application interoperability standards, speed up the adoption of digital technologies, and eventually achieve net-zero emissions.

We have pledged to reach net-zero emissions by 2050 or earlier and to cut scope one and two emissions by 50% by 2030, he added.

Although it may not be the magic solution, digital technology is one of the main tools Shell is adopting to speed up the energy transition. Jeavons continues, “We can also harness the data we have now and use it to alter the system, even though we’re going to need to modify a lot of technology to change the energy industry.

AI is important to Shell’s business strategy

The use of reinforcement learning in Shell‘s exploration and drilling programme, the introduction of AI at public electric car charging stations, and the installation of computer vision-capable cameras at gas stations are just a few of the AI projects the company has already put into action over the years.

The organisation has recently introduced the Shell.ai Residency Program, which enables data scientists and AI engineers to acquire experience working on various AI projects across all Shell companies.

At the moment, Shell releases more than 100 AI applications into production yearly. A central community of more than 350 AI experts has also been established, and they are now building AI solutions leveraging the enormous data sets that are accessible across the many Shell companies.

Shell benefits from AI for preventive maintenance

Safety and dependability are crucial, according to Jeavons. “It has been a goal for us to be able to recognise when things are going wrong and respond proactively.”

Shell may now utilise predictive monitoring to supplement their existing monitoring strategies thanks to AI.

To put it in context, Jeavons asserts that over 10,000 pieces of equipment, including valves, compressors, dry gas seals, instruments, and pumps, are presently being monitored by AI, which also makes forecasts about future failure scenarios.

Three million sensors collect 20 billion rows of data each week to keep track of all that equipment, and roughly 11,000 machine learning models enable the system to produce more than 15 million predictions every day.

In the past, Shell made similar forecasts using physics-based models. The business used to routinely replace components after a certain amount of time before implementing a predictive maintenance programme managed by C3 AI. As a result of this strategy, components were often changed while they were still in excellent shape. Another approach was to wait until something went wrong. When equipment failed, assets had to be temporarily shut down for repairs, which had an impact on output.

By deploying resources more effectively, reducing production disruptions, and preventing unscheduled downtime, AI-based predictive maintenance has allowed the organisation to reduce equipment and maintenance expenses.

AI is surrounded by a number of infrastructure and orchestration concerns, according to Tom Siebel, CEO of C3 AI.

Building machine learning models isn’t all that difficult, he said. What’s challenging is integrating two million machine learning models into a single application.

However, Shell’s data scientists could examine thousands of data points concurrently using a proactive technical monitoring method, allowing engineers and others to gain knowledge from the data.

“Our team uses that data to understand what normal behavior across our asset base looks like in particular cases, including equipment like compressors, valves and pumps,” Jeavons says. “Then we create forecasts of what we think normal is going to be in the coming periods. From that forecast, we can identify when normal conditions are no longer occurring and then link that back to historical events.”

Shell will next use AI for optimization

Shell has now released its C3 AI-powered AI predictive maintenance solutions for sale. According to Jeavons, the business is now entirely focused on optimization going ahead.

As a result, Jeavons added, “we can look at the CO2 footprint of these operations and try to optimise appropriately. This means we can find methods to produce more effectively, providing more output for the same cost.

He noted that Shell is also investigating how AI may be used to monitor carbon capture, storage installations, and methane levels in the near future.

In addition to improving our current operations, he said, “these enterprises are essential to our energy transition plan.”