IMAGE ’25 – International Meeting for Applied Geoscience & Energy

IMAGE '25 MeetingVenue: George R. Brown Convention Center, Houston, TX

Date: August 24-28, 2025

Short Course:  SC-01: Chemometric Tools to Establish Petroleum Systems, Predict Physical Properties, and De-Convolute Mixed Production

Course Leaders: Dr. Kenneth Peters from LSB NExT Training and Brian Rohrback from Infometrix will be leading the short course on August 24th.

This one-day course is for all geoscientists who want to extract hidden information from substantial amounts of chemical and physical data using multivariate statistical (chemometric) tools. The course emphasizes applications rather than the mathematics of various chemometric methods and will include a demo version of Pirouette 5.0 chemometric software. Case studies focus on the following topics of immediate interest to geoscientists:

• Hierarchical cluster analysis (HCA) and principal component analysis (PCA) of biomarker and stable isotope data for oil-oil and oil-source rock correlation to establish petroleum systems.
• Quantitative regression analysis of chromatographic peaks by alternate least squares (ALS) to de-convolute mixed oils derived from two or more sources. In exploration, ALS identifies mixed oils (e.g., pre-salt and post-salt oils in the Middle East and Southern Atlantic). In production, ALS allows allocation of mixtures originating from multiple reservoir zones.
• Prediction of physical properties by partial least squares (PLS) of data obtained by micro-analytical techniques. PLS allows investigators to predict API gravity, sulfur, and viscosity for reservoir zones where only small samples of cuttings from storage are available for analysis.

For questions or more information, contact info@infometrix.com.

Deconvoluting Mixed Petroleum and the Effect of Oil and Gas-Condensate Mixes on Identifying Petroleum Systems – AAPG ACE 2020

Virtual talk at AAPG ACE 2020

Watch recent virtual talk by Ken Peters at AAPG ACE on using Pirouette’s unmixing algorithm for evaluating oil production.

 

 

Two points made in the talk are:

  • You cannot use ratios as the input variables and need to use concentrations instead.
  • The alternating least squares algorithm performs well to untangle mixed sources accurately.