Machine learning

ML built for well data: clustering for electrofacies, supervised prediction of lithology and fluids, and reconstruction of missing or rejected intervals.

Clustering

• Unsupervised clustering — automatic electrofacies and fast first-pass lithotypes from log curves.
• Semi-supervised clustering — anchor on core-defined facies and propagate them across the field; repeatable as the model evolves.
• Family assigner — ML-based curve-type recognition.

Prediction

• Lithology prediction — supervised classification trained on described wells.
• Fluid reconstruction — ML reconstruction of fluid-related curves.

Reconstruction

• Log / curve reconstruction — recover missing or rejected intervals from neighbouring wells and a trained model.

Facies

• Facies by signal shape — facies classification from the shape of the log response.

This page summarises the Machine-learning section of the Wellbore.AI / Gamma technical manual.

← Back to features