Marathon Oil Scales Intelligent Alerts to Over 4,000 Wells Using AWS Partner Seeq
With 4,000 wells to manage and intelligent alerts to build and run, Marathon Oil decided to use Amazon Web Services (AWS) to improve its alert development time from months to hours. With its existing environment, the team at Marathon Oil was developing on average four intelligent alerts per year. “We weren’t getting any traction because it would take months to get an alert created, and we would lose momentum with the business team,” says Mark Betts, IT manager of digital solutions at Marathon Oil. “We needed a better solution.”
The company wanted to use the production data from its four assets, each with its own data collection process, to make it possible for its users to build intelligent alerts while providing a similar user experience across each asset. Rather than hosting the environment, Marathon Oil decided to focus on integrating intelligent alerts and spend less time building them. To achieve this, Marathon Oil chose to use AWS Partner Seeq, which hosts its applications on Amazon Elastic Compute Cloud (Amazon EC2), a service that provides secure and resizable compute capacity to support virtually any workload. With this integration, Marathon Oil builds alerts on top of its time-series production data in Seeq and then uses AWS to integrate and initiate actions in its downstream applications.
AWS introduced Marathon Oil to Seeq, and in April 2021, the three companies designed a pilot program and rolled it out to one asset over the course of 3 months. “Working closely with the Seeq team was by far the biggest reason for being able to integrate all the various data sources that we have,” says Betts. “The low barrier to entry to adopt is what we needed.” Since the successful pilot, Marathon Oil has continued to build upon its use of Seeq.
Using AWS Partner Seeq to Build a Data Hub and Intelligent Alerts
Marathon Oil is an enterprise customer of AWS and subscribed to Seeq through the AWS Marketplace, the place to find, test, buy, and deploy software that runs on AWS. Seeq is an advanced analytics solution for process manufacturing companies. The company’s mission is to improve manufacturers’ time to value by consolidating data and providing analytics.
Seeq provides benefits to customers through three software-as-a-service applications that run on Amazon EC2. The base is Seeq Cortex, which powers the applications to provide calculations at scale, data connectivity, and administration features. The main application Marathon Oil uses is Seeq Data Lab—an application for data scientists and process engineers to access Python libraries to expand their analytics—to scale alerts on its nearly 4,000 wells and facilities and to integrate with other applications. The company has over 50 employees using Seeq Workbench (which includes features to expedite the full arc of the analytics process), with 170 workbenches for one-time reporting and analysis.
By using Seeq alongside its digital oil field tools, Marathon Oil can monitor assets at scale and perform root-cause analytics for the upkeep of its wells. “Working with Seeq has been great from every side, support included,” says Betts. “It handles everything on the AWS side and has developed a lot of the tools that we use to scale across our wells and facilities.”
Improving Overall Performance Using Seeq
Marathon Oil runs servers on Amazon Elastic Kubernetes Service (Amazon EKS), a managed Kubernetes service to run and scale Kubernetes applications in the AWS Cloud or on premises, to query the Seeq Data Lab server and process the alerts and initiate actions and notifications in downstream systems.
This is where the custom user interface built by Marathon Oil comes in, where users can control what they need in a particular capsule. If they want a capsule to initiate a task in the field service management tool, it uses an API that alerts Marathon Oil through its service management tool alongside Amazon Managed Streaming for Apache Kafka (Amazon MSK), which businesses use to ingest and process streaming data in near real time with fully managed Apache Kafka. Finally, alerts can notify any necessary employees through email and chat notifications.
All these services and applications come together to provide business benefits for Marathon Oil. The development teams at Marathon Oil can focus on improving the digital oil field project because the intelligent alert creation happens with better time to value using Seeq. It used to cost over $10,000 in development time and hours to create a single alert. By reducing the development team’s need to assist in building alerts, the cost similarly decreases, all while Marathon Oil teams gain more time for innovation.
Marathon Oil is now able to scale actionable alerts across all its wells. As of mid-2022, the company had over 50 alerts, and it expects to be able to increase that number into the hundreds. “Being able to grow at scale using both Seeq and AWS is our focus,” says Betts. This scaling happens across all wells and supports multiple workloads and multiple processes. The intelligent alerts are used across all assets and are helping the shift from reactive to proactive well surveillance. Marathon Oil is pushing the technical limits by checking for new alerts and new capsules from Seeq on a constant basis. “We’re pinging Seeq multiple times per hour and pushing the limits from a performance perspective,” says Betts.
Along with improved scalability and creating an integrated hub for its assets, Marathon Oil has a goal of improving overall production performance. By using Seeq for exception-based surveillance on AWS, Marathon Oil intends to help its assets keep wells online longer and limit deferred production.
Embracing the Future Using Seeq and AWS
Using Seeq and AWS, Marathon Oil has achieved its goals of improving scalability and time to value. The intelligent alerts function in near real time, informing Marathon Oil when a condition might take a well offline and guiding the company on preventive measures.
Marathon Oil plans to continue improving overall reporting and visualizations within Seeq and in the company in general. “We’re looking at adopting more prediction and modeling on the artificial intelligence and analytics side from Seeq for root-cause analysis,” says Betts. “By being able to conduct root-cause analyses to recognize patterns, we can make higher-level changes and decisions around our production operations.