Related Articles
Forward article link
Share PDF with colleagues

Rise of the machines

Technology is increasingly replacing the roles of humans across the energy supply chain

The application of machine learning to the oil and gas industry is well underway and has the potential to transform efficiency and working practices across the entire value chain. 

Atomiton works with majors, and other oil and gas companies, on industrial software that makes their operations more efficient and agile. It offers a “new category of analytics programs”, according to CEO and founder Jane Ren, across field, plant and infrastructure integrity applications.

What benefits do artificial intelligence (AI) and machine learning offer?

Ren: AI is a broad term for computer programs that leverage data to make better decisions than humans.

Machine learning is a narrower category within AI, a collection of methods that train programmers to discover phenomena behind large quantities of data. It is about understanding, for example, the factors wearing machinery or how terrain effects efficiency. 

An important category is optimal search, which makes predictions amid uncertainty. If 300 customers demand different products—what is the most efficient way to schedule the pipeline? Machines can do it using very sophisticated methods.

Which parts of the supply chain can benefit most?

Ren: The most immediate improvement is energy efficiency. Downstream plants are the biggest energy consumer—we usually see energy cost savings of 10-20pc. The problem is well-defined and there is enough data to run an algorithm. People use energy based on habit, leaving systems running 24/7 or generating more steam than they need. It is hard for operators to be efficient without predictive algorithms.

“The idea that algorithms will replace humans is nowhere close to being true”

Upstream is more human-driven, so data is not as available. But when half a day [of downtime] can cost $500,000, small improvements have big impacts.

How receptive is the industry to this technology? 

Ren: The mindset has changed at the executive level—it is no longer conservative. Curiosity and openness to new technology increased vastly in the last five years, due to oil price instability and uncertainty for the industry. 

But implementing is harder than believing. Knowledge is fragmented between IT people and field operators. It can be hard for field workers that have followed rigid procedures for 10 years to adapt to real-time intelligence.

Will the future workforce need different skills?

Ren: This may come naturally as we all use technology in our lives, not just work. People expect their working environment to change. There is a move to be cross-functional and collaborative within an organisation. AI integrates the entire supply chain and enables change.

Atomiton CEO and founder Jane Ren

How will machine learning develop?

Ren: The idea that algorithms will replace humans is nowhere close to being true. Algorithms suffer if they don't integrate with humans. But machine learning will get closer and closer to the human operator. The next wave of applications will allow humans to control when they run algorithms and be able to override AI decisions. 

How much further could this trend continue?

Ren: We are modelling how human operators make decisions so they will be less and less directly involved. Humans will be more focused on open-ended and strategic decisions. They will not need to go to dangerous places such as sub-sea pipelines.

The first phase was common analytics, which augments the human operator and provides cost savings, resource efficiency, positive environmental impacts and fleet safety. 

“Curiosity and openness to new technology increased vastly in the last five years”

The second phase is more entrenched and leads to bigger changes to the operating model. Decisions—how we store inventory, transport products, set prices and organise projects—are often based on infrequently available information and an inability to make system-level evaluations. This phase means daily pricing can become dynamic and storage inventory can be auto-replenished based on predicted demand.

Beyond that is the third phase. In other industries, we have seen the anchors of competitiveness shift. Oil and gas companies compete on their ability to discover resources but, as they transform into energy companies, this may shift to competing on efficiency, service quality or reduced environmental impact.

What role will technology have in the energy transition?

Ren: The transition will depend on the ability of infrastructure to accommodate alternatives. Majors are already diversifying into wind, solar and other sources. Diversified supply chains require control over not only petroleum but electrons. Intelligent software will help manage energy grid infrastructure and predict fluctuations in demand for all sources of energy.

Also in this section
Latest licensing rounds
10 July 2020
The industry's most comprehensive list of current and recent rounds for onshore and offshore licences
Congo’s disappearing discovery
9 July 2020
The Republic of Congo’s announcement of a major discovery appears to be targeted more at extorting EU environmentalists and multilateral lenders than attracting IOCs
Australian upstream rejects politicians’ call for A$4 gas
8 July 2020
Producers insist the costs of producing gas domestically—as well as prevailing spot prices—significantly exceed the historic benchmark price level