Skip to Content
Author's profile photo Former Member

SAP AI: Machine Learning in Oil & Gas

Oil & Gas Analytics

Analytic technology has been used in the Oil Company Upstream segment since the early 90’s with the analysis of collected data pertaining to such areas as downhole conditions, drill bit performance and reservoir dynamics for improved efficiency in operations and decision making.  Implementation however generally required a significant investment in tools, software and resources and would often be used on only high producing wells.

With oil prices experiencing a downward trend and continued tightening of margins, oil companies have been motivated to implement innovative ways to improve their efficiency, reduce costs and minimize unplanned downtime; effectively shifting the focus from increasing the production to optimizing it. (1)  Today’s advanced software solutions and technology can help. Machine learning algorithms are making computers progressively more intelligent as they learn how to efficiently access big volumes of data, analyze and make connections, and subsequently make predictions and provide recommendations for operations optimization.

SAP AI – Machine Learning

SAP is planning to step up its machine learning and artificial intelligence efforts in hopes that its applications will have a broader reach when it comes to automating processes.  At the SAP Capital Markets Day in February 2017, SAP announced Machine Learning as one of the key advancements to its S/4HANA ERP suite, which also includes a new architecture of in-memory technology, combined with contextual analytics, and digital assistant capabilities.  SAP stated that its ERP suite will be enabling customers to instantly adjust and adopt business processes and models, and act on real-time insight and advice.  (3)

There are three key reasons why Machine Learning is rapidly coming into play.

  1. Access to huge volumes of Big Data stored on larger more efficient networks.
  2. Technological advances in computing power enable the processing of massive amounts of data sets from a multitude of sources – such as text, images, and IoT devices.
  3. Improvements in machine learning algorithms which leverage the ever-increasing wealth of information available across business networks and the cloud. (6)

SAP CLEA is the recently launched Machine Learning Solution of SAP that provides embedded intelligent systems on the SAP cloud platform with a portfolio of applications involved.  CLEA is the first front of the future of SAP’s Machine Learning systems which are to include several applications, tools and services, all developed in collaboration with and feedback from the various co-innovation customers across domains and regions. (7)

“Data is the fuel for machine learning. SAP systems touch more than 70 percent of the world’s business transactions. With this asset, and our deep knowledge of business processes, we are poised to create an end-to-end intelligent enterprise.”  (8)

Example:   The key functionalities of the SAP Clea for Cash Application shown below demonstrate how a CLEA app processes tasks based upon Machine Learning analysis and prediction.  Benefits include the reduction in human effort, increase in efficiency of invoice processing and improved quality of service.

Machine Learning in Oil & Gas
Machine Learning opportunities in the Oil & Gas market range from traditional back-office tasks, to equipment monitoring and predictions regarding events that impact overall operations.  These opportunities can be achieved with machine learning advanced algorithms and analysis of data as; quick identification of trends and patterns, processing of billions of data points in real-time and layering of information combining multiple variables.  The relative success of AI-based technology in other fields such as healthcare, finance and manufacturing has also grabbed the attention of the oil and gas industry, leading many companies to experiment with it in their own operations.

Use cases:

  • Collecting and analyzing multiple factors relating to the overall drilling strategy such as equipment ratings, seismic vibrations, strata permeability and thermal gradients for determination of optimal drill bit direction and control as it bores.
  • Monitoring of the overall performance of wells and analyzing the data to make more intelligent decisions aimed at improving operations.
  • Predictive software can also be used to analyze data to determine if downhole conditions are conducive to potentially catastrophic events such as a lost circulation, stuck pipe or blowouts. Leveraging data to understand what contributes to the likelihood of such an event, algorithms can provide recommendations to the control system and operating personnel to minimize the odds of such an occurrence. (1)

In 2015 SAP Co-innovation Lab in Silicon Valley partnered with Mtell and Rolta on a solution for predictive machine maintenance in oil and gas. Mtell created a machine learning and neural network that recognizes patterns associated with machine performance. Their software automatically generates notifications and sends them to SAP Plant Maintenance, which planners use to schedule machine repair and replacement. This improved the ability to analyze volumes of O&G asset data by integrating the solution with SAP HANA. The in-memory technology and analytics of SAP HANA enable real-time analysis to pinpoint potential machine failures. (11)

Analysts at McKinsey have identified $Billions in potential savings in the Oil & Gas Industry with operational improvement opportunities for machine learning–based systems covering a wide range, from invoice processing and logistical optimization, through failure prediction for equipment and control rooms.

Analysts at McKinsey identify $Billions in potential operations improvements.  (1)


Regarding Oil Field Services, the next steps I see for applications of SAP Machine Learning include:

  • Predictive maintenance based upon job history where determination will analyze the downhole conditions such as high pressure and acidity impacting the degradation of equipment.
  • Automatic invoice matching based upon customer payment patterns and typical variances disputed.
  • Capacity constraints based on historical demand patterns that predict capacity bottlenecks in terms of equipment, human and component resources.

I would be interested in your thoughts and feedback for continued discussion as we rapidly move into the next phase of Big Data analytics and Machine Learning.

Assigned Tags

      Be the first to leave a comment
      You must be Logged on to comment or reply to a post.