Loading...

Constructing a Predictive Model for Water Saturation Using Artificial Intelligence Models and Conventional Well Logging and Core Data

Karimi Arpanahai, Azam | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57408 (06)
  4. University: Sharif University of Technology
  5. Department: Chemical and Petroleum Engineering
  6. Advisor(s): Jamshidi, Saeed
  7. Abstract:
  8. Water saturation is a critical parameter in hydrocarbon reservoir studies, directly impacting the assessment of recoverable volumes and production planning. Accurate calculation of this parameter in carbonate reservoirs is particularly important, as precision in its determination can lead to optimal decision-making in exploration and production. Traditional methods such as Archie and Indonesia equations, despite their widespread use, may yield inaccurate results due to their dependence on empirical coefficients and limitations in specific field conditions. In this study, we employed artificial intelligence models, including Random Forest and Symbolic Regression, to develop more accurate methods for estimating water saturation. The results demonstrated that the Random Forest model achieved a Mean Absolute Error (MAE) of 0.039, a Mean Absolute Percentage Error (MAPE) of 5.89%, and a coefficient of determination (R²) of 0.924, outperforming traditional methods. Additionally, the Symbolic Regression model provided highly accurate predictions of water saturation with precise mathematical relationships. These findings suggest that artificial intelligence models can serve as effective and reliable alternatives to traditional methods for estimating water saturation in carbonate reservoirs
  9. Keywords:
  10. Water Saturation ; Carbonate Resrevoirs ; Artificial Intelligence ; Well log Data ; Random Forest Algorithm ; Symbolic Regression

 Digital Object List

 Bookmark

No TOC