AI is sparking a smarter, cleaner energy industry
Digital tools boost efficiency, exploration and emissions data
The energy industry faces no shortage of challenges as the world emerges from the coronavirus pandemic. From falling petroleum demand to increasing pressure to reduce greenhouse gas emissions to the rapid growth of renewables, energy companies of all sorts are contending with new demands and shifting business conditions. And they are turning to a new tool to help them adjust to the brave new world: artificial intelligence.
The technology is helping producers tap oil and gas wells more efficiently, pipeline companies identify leaks, and power grids
integrate wind and solar energy with batteries and traditional generators. Artificial intelligence, or
AI, is making sense of vast troves of data collected from exploration activities, drilling activities and other operations that deliver energy to the global economy.
“Everybody in the world is trying to improve efficiencies,” said Ron Beck, energy industry director at the Boston software company AspenTech. “AI becomes very critical for that.”
Artificial intelligence is the term used to describe computers, software and machines that mimic human intelligence, processing information and learning from it to solve problems and adjust to changing circumstances. Although the energy industry has been slower than health care, e-commerce and other industries to embrace AI, that’s beginning to change.
The technology holds out the promise of solving many of problems vexing the energy industry, including precisely measuring the volumes of carbon emission created across an oil company’s entire value chain, from wellhead to retail sales.
The energy industry’s adoption of AI is still in its earliest days, but it is already proving its potential to make the industry cleaner, more efficient and more profitable said Brian Sallade, CEO and president of The Woodlands-based Kinsmen Group, an engineering information management company.
Perhaps the technology’s most immediate impact has come in analyzing the never-ending stream of data generated by energy operations and synthesizing it to determine how to get the most out of an oil reservoir or when to replace a part to prevent a failure.
“There’s just no human being who can evaluate the amount of data we have,” said Sallade. “Without AI, you’re going to produce data that’s going to be dead data, black data, because you don’t have the resources to do anything with it to turn it into value.”
Shell takes lead
Among energy companies, the European oil major Royal Dutch Shell has been a leader in AI. For example, the company is using AI technology at its Quest carbon capture and storage facility in Alberta, Canada, to ensure that the carbon dioxide the plant injects into the ground stays in the ground.
“Digital in general and AI in particular are very much front and center of Shell’s thinking,” said Dan Jeavons, Shell’s general manager of data science.
In 2019, Shell announced that its digital technology initiatives had resulted in about $1 billion in profit margin increases, cost savings and production increases. In 2020, that number doubled to about $2 billion, despite the impact of COVID-19 and personnel working remotely.
Artificial intelligence, Jeavons said, represents the next step in the coming digital transition, building on earlier technologies developed by the energy industry, such as smart sensors used to detect leaks along miles of pipelines. These sensors produce the vast amounts of data that can be scooped up and processed with AI.
“AI is super-data hungry,” he said. “We have a ton of data, more than a terabyte of data in cold storage.”
The energy industry in general has been slow to adopt digital technologies, in part because of concerns about cybersecurity threats. But as companies have hardened their defenses, the energy industry has moved more quickly to adopt digital innovations, including AI, analysts and industry officials said.
The push to embrace the digital technologies led Shell to partner with outside companies that provide AI software. In 2018, Shell began collaborating with the Silicon Valley software company C3.AI to embed artificial technology across its global operations.
Using AI tools, Shell can analyze vast volumes of seismic, drilling and geological data more easily and quicker than before. AI simulations of field development have improved the producer’s ability to plan the development of an oil field and help determine the best placement of wells.
The Houston oil-field services company Baker Hughes launched a joint venture with C3.AI in 2019. Dan Brennan, vice president of the joint venture, called Baker HughesC3.ai, said artificial intelligence allows energy companies to take advantage of two broad technology trends sweeping the oil and gas industry: the proliferation of cloud computing, which uses a network of remote servers to store, manage and process data, and the proliferation of data-collecting sensors.
“We’ve got north of 25 billion assets that are connected,” Brennan said.
In February, Shell and its partners took the collaboration further, announcing the launch of the Open AI Energy Initiative (OAEI), which creates an AI platform that is open to all companies to use and build on as they develop their own AI software.
“The spirit behind the Open AI Energy Initiative is: Can we not come together in specific areas where we’re not competing, but in areas where we have a vested interest, like reliability?” Jeavons asked.
AI in oil and gas
In oil and gas production, companies can use AI technology to combine computer models of oil and gas reservoirs with data sets collected from divergent sources, such as long-term weather and economic forecasts, to plan the most efficient placement of wells for 10 to 15 years in the future, said Shahram Farhadi, chief technology officer of Beyond Limits, an AI software company.
AI-enabled field development planning will allow oil companies to optimize their drilling and production plans and determine the best placement of injection wells used to stimulate production as the oil field becomes more mature, said Farhadi.
The use of AI in field development planning will also help producers substantially reduce their carbon footprint over time, by limiting the number of wells that producers must drill to produce the same amount of oil.
The technology could also prove useful in helping to store carbon dioxide produced in oil and gas drilling by identifying geological formations into which a producer could inject the greenhouse gas, he said.
Among other potential uses of AI in the oil and gas industry are analyzing pipeline flows and creating “digital twins” of refinery equipment. Digital twins are virtual models that can predict problems before they happen, allowing for plants to replace parts likely to fail before they do. AI-enabled drones and other robots can perform continuous inspections of pipelines to identify methane leaks or monitor for corrosion.
AI in power industry
Like the oil and gas sector, the power industry is beginning to see the benefits of adopting AI, particularly in modernizing the electric grid. The nature of the power industry is changing rapidly as wind and solar generation and battery storage are integrated into the grid alongside more traditional forms of generation, such as coal, natural gas and nuclear energy,
Larsh Johnson, chief technical officer of Stem, an energy storage company, said his company uses its AI software to determine when to charge the batteries and when to dispatch the battery power to the grid, depending on the price of electricity. When power is cheap, the system stores it in the batteries and releases it when prices go up.
AI could have helped alleviate the disastrous near collapse of the power grid in Texas during the big freeze last month, said Steve Kwan, Beyond Limits’ director of product management for power generation and grid management.
AI could have helped the Electric Reliability Council of Texas, the state’s grid manager, to better understand the potential impact of severe weather on the grid. This would have given ERCOT the opportunity to warn large industrial customers to cut back earlier in the days-long crisis to decrease demand on the grid.
AI also could have helped run simulations on the effects of rolling blackouts, and the best way to implement them to minimize the impact on consumers.
That “would have given consumers a little bit of advance notice so that people could have planned a little bit better,” Kwan said. “What we’re seeing now is the blackouts being implemented in a very reactive manner.”