Elastic powers Shell’s flexibility to thrive in the energy sector

This post is a recap of a presentation given at ElasticON 2020. Interested in seeing more talks like this? Check out the conference archive.

Shell International knows that it takes cutting-edge technology to thrive in the competitive, global energy industry. With projects around the world, in both renewable and non-renewable energy, Shell must always have insights into the future. From determining expected output to predicting equipment failures, there’s no room for guessing in an industry where downtime is unacceptable. 

This is why Shell is a part of the Open Subsurface Data Universe (OSDU), and why they use Elasticsearch for the analysis of a range of geospatial, full-text, and numeric data that is critical in the energy space.

During his ElasticON Global presentation, Johan Krebbers, general manager of Digital Emerging Technologies and VP of IT Innovation at Shell International, spoke to the importance of the flexibility that Elasticsearch provides the energy giant — including its ability to be cloud-native or on-prem, and its ability to be used with several hosting providers to enable Shell to comply with data retention regulations imposed by governments throughout the world.

What’s more, because of its capability to search and analyze so many different data types, Elasticsearch is at the heart of Shell’s observability, machine learning (ML), machine vision, and natural language processing solutions.

Predicting failure, increasing uptime

According to Krebbers, the best time to replace equipment is before it fails. If you wait until it’s too late, the cascading effects can drive up repair costs, increase profits, and impact revenue and growth. By leveraging the real-time ingest, ML, and observability capabilities of Elastic, Shell is able to use predictive modeling to replace or repair machinery before it fails. 

“We use ML for predictive maintenance of our facilities,” says Krebbers, who is charged with bringing new technologies to Shell. “You collect the real-time data. You have the ML models. And then [you] start predicting when is a pump going to fail? When is a compressor going to fail? If you can predict failure, you can predict downtime. And downtime always costs you money. So you want to increase your uptime.”

Another way to keep overall costs down is to make sure resources aren’t wasted. That’s why Shell uses a machine vision solution with Elasticsearch at its core to gain insights into potential leaks throughout their global infrastructure. 

Detecting spills, leaks, and emissions with robots

“Machine vision plays an important role in leak detection and emission detection,” says Krebbers. “You have robots driving around with cameras collecting spillage, leakage, emissions. [Then] you bring it into a cloud environment, apply machine visions and start looking for videos of leakage and spillage. And when there’s an issue, you can immediately raise that with appropriate staff.”

Going beyond detecting leaks, spills, and emissions, Shell must always be on the hunt to harvest new energy sources. 

As part of their mission to bring more resources to market, Shell employees are embracing and demanding natural language processing capabilities with Elasticsearch as the prime search engine. Many Shell experts, for example, are using natural language processing to quickly search through mountains of the company’s mission-critical subsurface data related to wells, development, and exploration to keep Shell competitive.

Watch the full conversation with Shell International to learn more about how Shell increases uptime, detects emissions, and surfaces exploration data faster with Elastic.
Source: Elastic

Leave a Reply

Your email address will not be published.


*