Accelerated condition monitoring for Wintershall Dea

IN SHORT

Wintershall Dea used Cognite Data Fusion to combine contextualized data with domain expertise and develop machine learning models to monitor the installed gas turbine at the Mittelplate oil field, detecting faulty shutdowns and alerting maintenance engineers to avoid a potential breakdown during restart.

• $865,000 saved per incident

• Less unplanned downtime, lower repair costs

• More condition-based maintenance 

CHALLENGE
Liberate data for condition monitoring

An installed gas turbine generates electricity to power Wintershall Dea’s installation at the Mittelplate oil field off the coast of Germany. However, the maintenance engineers and electricians responsible for monitoring the turbine rely on fixed thresholds for alarms and warnings.

These experts lacked an easy way to receive early warnings about failures, which would enable them to perform troubleshooting on the turbine and avoid breakdowns.

Wintershall Dea needed a solution that could operationalize machine learning models at scale, helping maintenance experts monitor the condition of the gas turbine.

SOLUTION
The solution saves Wintershall Dea $865,000 saved per incident

Wintershall Dea worked with Cognite and the turbine service company to liberate sensor data and events such work orders and collect all the information as a contextualized set in Cognite Data Fusion.

Cognite and Wintershall Dea’s data scientists then used the contextualized data to build and operationalize machine learning models using the model hosting service in Cognite Data Fusion. The outputs from the model are visualized in a Power BI dashboard.

The solution provides electricians and maintenance engineers with notifications and intuitive dashboards with data that describe the status of shutdowns, which help them better understand the condition of the turbine.

The dashboards also let users explore sensor trends and warnings for each shutdown, early warnings of deviations from normal shutdowns, and analyses of historical shutdowns.

IMPACT

Wintershall Dea’s rotating equipment maintenance experts estimate that the dashboards will help the company save $865,000 per incident by automatically detecting unhealthy shutdowns, thereby eliminating unplanned downtime and repair costs.

With greater access to data and insights, the maintenance and service engineers can move toward a condition-based maintenance strategy, making Wintershall Dea’s operations safer and more reliable.