A lot is expected of Meindert Dillen and Philipp von Wussow by Wintershall Dea.
Their mission is to ensure that the company’s technical team is increasingly able to use advanced data analysis to find and produce oil and gas more productively.
The focus is on increasing the capabilities of those with traditional engineering and geology training. “Someone who can understand seismic processing can program a neural network,” said Patrick von Pattay, a vice president for Wintershall Dea and chairman off the Digital Transformation Committee of SPE’s Digital Energy Technical Section.
That is an apt description of Dillen, whose work as a geophysicist using advanced analytics led to his current job. In both roles, finding new ways to extract useful bits of information from massive data sets is valuable.
This team was created after the merger of Wintershall and Dea—with one member from each company. They have been working to scale up their influence by “building a community and pulling people together.”
A key part of their effort was creating a digital skills network a few years ago. This grassroots effort has helped bring together technical staffers with the wide range of traditional and digital knowledge needed to deliver digital change.
“We are in a lucky position,” Dillen said. “A lot of people are interested in this technology and want to apply it in ways” that can have a significant impact.
The number involved must be expressed as a range—between 100 and 200, because engagement varies—with people including geophysics experts who were learning about neural networks in the 1980s when the available computing power limited its uses and digital natives who wonder why their workplace is not using the tools found in everyday life.
A network’s worth of skills is required because the potential applications are as varied as the many disciplines within engineering, geology, and geophysics, among others.
“No one on Earth can define all the use cases,” said von Wussow, whose career began in subsurface roles and later included stops in business development and management along the way to analytics.
The company does offer digital skills training and discussions—which are done online because of COVID-19. A lot of the new thinking is spread by word of mouth in online communities.
“Now, people in Norway who know people in Argentina and Russia spread the ideas,” von Wussow said.
Von Wussow and Dillen sometimes play the role of matchmaker. Relatively simple requests can be met by connecting people with complementary skills inside the company or steering them to outside sources.
Other times, the problem is bigger and project evaluation and management skills are required, beginning with figuring out the root of the problem, how digital might help solve it, and whether the benefit of doing so justifies the effort.
If a project is a go, those involved need to think through the plan of attack and consider the people and resources required and whether they will come from inside or outside the company.
An example was a project to create the Exploration Advisory Tool—a digital application that would help search for valuable information in a mass of written reports when evaluating new prospects. The business case for this is that it can provide bits that reduce the uncertainties for decision-makers when deadlines loom on bid rounds or farm-in decisions.
They hired IBM because of the company’s expertise in cognitive search—a language processing method able to identify passages likely to identify key bits in documents lacking the structured formats normally required for computer analysis.
The result allowed exploration teams to “zoom in quickly on what they want” when evaluating an unfamiliar area, Dillen said. As the project’s data scientist, he worked to make sure the company would be able use what was developed to extend the artificial intelligence functionality of the tool for geologists to other disciplines in, and beyond, exploration.
As these tools are developed, the innovations are being scaled up. Those approaches include standards for coding and documentation allowing others to use previous solutions for new problems. To make that code accessible, it is stored in a corporate hub where users can find useful bits of code—a tiny relative of a major code source such as GitHub.
“What we want is for people to work in the same way on data and data that is accessible for everyone,” Dillen said. They are promoting digital methods that can be shared and taught more easily, allowing common digital training programs and interfaces among business units.
“Having a common data science environment enables us to structure and scale solutions” for others to use, Dillen said.
“We have to marry machine learning and engineering to show what is hidden in the data we have,” von Wussow said.
