Permeability log estimation
Challenge
- Critical to capture reservoir heterogeneity through permeability streak to increase 3D model accuracy.
- Automate and accelerate data integration of different data types (core data, open hole data, well test data). Improved data analysis and accuracy aiding in delivering consistent and accurate results.
- The permeability of the reservoir is dependent on the depositional environment and diagenesis.
- Limited data/lacking input data integrations for permeability modeling.
- Effect of different geological processes on reservoir permeability, sometimes causing misrepresentation of high streaks.
Solution
Our permeability log estimation workflow uses a machine learning based approach, taking into account core measurement, log measurement, and well test data to accurately predict permeability of the target reservoir.
Results
Higher accuracy of predictions by delivering more consistent models.
Integrating new drilled well data efficiently and cost effectively.
Prediction of permeability using ML methods like random forest and decision tree.
80% improved accuracy—actionable.
20% ↑ faster model generation to fast-track field development planning.
Integration among domains.
