Webinar: Machine learning based oil family classification above 62° North – method and some observations

Welcome to a webinar arranged by the FORCE Petroleum Systems network. This one-hour talk will be given by Benedikt Lerch. Lerch will be talking about Machine learning based oil family classification above 62° North – method and some observations.

Date Time Duration Register by Location
24.03.2022 15-16 60 Min 23.03.2022 Teams




Unsupervised and supervised machine learning algorithms were applied on two large geochemical data sets from the Norwegian Sea (N = 253) and from the Barents Sea (N = 86). The unsupervised approach allows the algorithm to find similarities among the samples without giving pre-defined attributes. Two different approaches were tested for the unsupervised method.  1) All available geochemical data i.e. light hydrocarbons, aromatics, biomarkers and fraction isotopes were combined in one single dataset and used without pre-screening, and 2) the datasets were carefully examined with regards to sample type (oil vs. condensate), contamination and data integrity. Based on this pre-screening, it was decided to divide the data in 3 compound classes (light hydrocarbon, C10-C16 aromatics, C20+ biomarker range compounds). It could be shown that by combining all available data, the algorithm classifies the samples based on large scale processes (biodegradation, oil based mud contamination, oil vs. condensate), rather than reveal sample classes based on i.e. source rock depositional environment, type of organic matter supply or maturity. For the supervised approach – which uses pre-defined labels for classification – we established a workflow that enables an expert to infer oil types and source rock correlation in an immature (low confidence) area given available samples in a mature (high confidence) area. We emphasize through examples the importance of data quality and data consistency when using machine learning for automatic oil typing. The results demonstrate that both unsupervised and supervised machine learning can be used to differentiate between different petroleum provinces in a very consistent manner, and furthermore give additional insight into outlier samples.


Benedikt currently works as a consultant at Aker BP focusing on petroleum system analysis offshore Mid-Norway. He received his diploma degree (equivalent M.Sc.) in geology from the University of Kiel (Germany) and holds a Ph.D. in petroleum geochemistry from the University of Oslo. After his Ph.D. he spent two years of postdoctoral research in Oslo, before starting in Aker BP in 2018.


How a webinar works

You register as usual through the registration button above. Once you are registered you will get an invitation via email to join this webinar. 
FORCE uses Teams Video for this webinar, and has proven to work successfully. 

We recomment that everyone joining turn off their camera and microphone when joining.
If you have any questions you can use the chat or wait until the end of the talk.

FORCE seminars have previously been fully booked with waiting lists so you are encouraged to sign up as soon as you know you will attend.

Once you have registered, you will receive an outlook invitation with the Teams link. You will receive the link a few days prior to the webinar. 
You cannot forward the link to people who is not registered. 


Participation fees:
FORCE members: Free
Non-members: NOK 350,-
University/student: Free


Important information:
You can register as a FORCE member and pay "FORCE member" price if you are an employee of a member company.
All FORCE member companies are listed here.

Payment is made online by credit card. Please note that no refunds will be given after you have signed up.
If you for any reason can not attend the workshop, you are welcome to send a representative, just inform Linn Smerud as soon as there are changes. 

Updated: 28/03/2022

Did you find what you were looking for?

Do not enter personal information here.

We don't answer request from this form. We only use this response to improve the website.