AI Assisted Interpretation of Core Photos and Vshale Prediction

Cores are the “ground truth” data for subsurface reservoirs, providing invaluable information that is often underutilized in reservoir characterization. While well logs, such as gamma ray logs, offer indirect measurements of rock properties (e.g., Vshale) through empirical formulas, they lack the resolution to capture thin-bed shales in heterolithic reservoirs like the Tilje Formation in the Norwegian Sea. These environments, formed in nearshore to intertidal settings, contain thin-bed shales that are often below the resolution of well logs.

coreAI Workflow

coreAI, part of the SubsurfaceAI 2024 release, addresses this challenge by providing an innovative AI-driven workflow for core photo interpretation and property prediction. Here’s a step-by-step overview of the coreAI workflow:

1. Project Setup: Build a SubsurfaceAI project and import well geometry and true color core photos. Core plug measurements and existing core interpretations can also be imported.

2. Removal of Non-Geological Features: Remove irrelevant elements from core photos, such as foams, empty spaces from core plugs, and core cases, to focus on geological features.

3. AI Model Training for Rock Types and Lithofacies: Develop and train AI models to accurately recognize various rock types and lithofacies.

4. Quality Control of Facies Prediction Models: Test AI models with blind test data from intervals not used in training to ensure reliability.

5. Generation of Lithofacies Images and Logs: Produce detailed lithofacies images and logs using the trained AI models.

6. AI Model Training for Core Plug Properties: Train AI models to predict crucial core plug properties, including grain size, porosity, and permeability.

7. Quality Control of Rock Properties Models: Validate the accuracy of AI models by applying them to blind test data from untrained intervals.

8. Prediction of Core Plug Properties: Apply AI models to predict core plug properties along intervals with core photos.

Core Photos rightImage_AI Assisted Interpretation of Core Photos and Vshale Prediction

Delivering Results

Delivering Results in Industry-Standard Formats

coreAI ensures that all prediction results, whether lithofacies interpretations or rock property predictions, are delivered in industry-standard formats, such as LAS or ASCII column files. These results can also be used as training data for conventional petrophysical logs, enabling property predictions, such as Vshale, in wells lacking core photos.

An Integral Part of SubsurfaceAI 2024

coreAI is a vital component of the integrated AI solutions offered in the SubsurfaceAI 2024 release, designed to enhance subsurface workflow solutions. By integrating coreAI into your workflow, you can achieve unprecedented efficiency and accuracy in reservoir characterization and property prediction, positioning your operations at the forefront of technological advancement in the oil and gas industry.

Revolutionizing Core Data Interpretation

coreAI is set to revolutionize the way you interpret core data and predict reservoir properties, delivering faster, more accurate, and more consistent results for better reservoir characterization.

Related Articles

Groundbreaking Workflow Development in Collaboration
SubsurfaceAI Inc. Announces Groundbreaking Workflow Development in Collaboration with ENI...
Innovative AI-Powered Method for Advanced Reservoir Characterization
SubsurfaceAI Inc. & Eni Reveal Innovative AI-Powered Method for Advanced...
SubsurfaceAI 2024.1 release announcement
Take a look at SubsurfaceAI 2024.1 and read about what’s...