Full Waveform and Rock Physics-based Machine Learning for Reservoir Property Prediction

Course Overview

Limited well control is the main barrier to reliable machine learning in exploration. This course shows how to generate physically realistic training data by combining rock physics model scenarios with full waveform synthetic modelling, then train convolutional neural networks and transfer them to real seismic for reservoir property prediction.

What you will gain

• A practical way to overcome sparse well data by creating large, geologically realistic training catalogues
• Training data from full waveform modelling that captures tuning, transmission losses and mode conversions missed by simple convolution
• The skills to train, validate and apply CNNs for predicting elastic properties, porosity, saturation and facies
• Rigorous QC methods and benchmarking against conventional inversion, so ML results can be trusted
• Hands-on experience with a complete, repeatable workflow on a real dataset

Software covered

HampsonRussell GeoAI

Who should attend?

Geophysicists, geologists, engineers, and technical staff who want to understand the theory and learn how to apply these increasingly critical techniques

Pre-requisite

Experience with HampsonRussell Software is not a prerequisite for this workshop; however, students will become familiar with its functionality through participation.

Duration

2 days (4 hours per day)

Format

Instructor-led, workflow-based, virtual training

*Please fill out the form to receive invoice.