Machine learning model to predict the contact of angle using mineralogy, TOC and process parameters in shale

A machine learning is needed to predict the contact angle in the shale using the process parameters and TOC and Minerology of the shale. Minerology and Total Organic Carbon (TOC) content are some of the important parameters to be evaluated for reservoir characterization. Wettability is the capability of a liquid to remain in contact with a solid surface affected by the balance of both intermolecular force of adhesive force (liquid to surface) and cohesive force (liquid-liquid). The study aims to investigate the effect of both parameter, TOC, and mineralogy on the shale wettability with a case study of Malaysian shale sample. The values for each parameter, TOC and minerology are obtained through thermal pyrolysis and X-ray diffraction, respectively. Advance application is carried out by applying the machine learning technique to predict the effect of shale TOC and minerology to wettability of the reservoir rock. The application aims to develop a machine learning program using the algorithm of Support Vector Machine or Gaussian Process Regression to successfully predict the contact angle. The developed model has successful in prediction the contact angle for different input variables of the machine learning model with high r squared values. © EAGE Asia Pacific Virtual Geoscience Week 2021. All rights reserved.