Deep QI is a machine learning add-on for RokDoc that combines advanced neural networks with geological knowledge to predict missing subsurface data, automate rock property analysis, and accelerate quantitative interpretation workflows.

Key Features

Best-in-Class Machine Learning Algorithms Advanced neural networks guided by geological expertise

• Leverages deep neural networks and modern machine learning techniques for subsurface prediction
• Integrates XGBoost machine learning combined with grid search for parameter tuning
• Guides predictions with geological knowledge from petrophysics and rock physics
• Predicts missing data beyond training datasets through knowledge-informed algorithms
• Automates rock physics modeling functions (RPML) with machine learning calibration

Flexible Data Handling Processes varied dataset sizes and volumes efficiently

• Handles datasets of varied size and volume with flexibility
• Accommodates regional information analysis across different data scales
• Processes well data, rock physics models, and pressure trends seamlessly
• Manages complex, heterogeneous subsurface datasets
• Supports batch processing and automated workflows

Integrated Rock Physics and Petrophysics Combines quantitative interpretation disciplines

• Leverages well data, rock physics models, and pressure trends across the RokDoc platform
• Integrates petrophysical property prediction with machine learning
• Applies rock physics constraints to machine learning predictions
• Enables disciplinary integration through automated workflows
• Calibrates mineral volumes directly in rock physics models

Automated Property Trending Streamlines rock property analysis workflows

• Streamlines rock property trending for Ji-Fi facies-driven inversion
• Automates quality control (QC) and knowledge generation workflows
• Reduces manual interpretation effort through automation
• Enables rapid model refinement and improvement
• Supports repeatable documentation and summary plots for test comparisons

Petrophysical Property Prediction Predicts missing well-log and core data

• Uses trained machine learning models for petrophysical property prediction
• Predicts rock properties, lithologies, and pore fluids away from wells
• Fills data gaps in well logs and core analysis
• Improves property prediction accuracy through geological guidance

Enhanced Quality Control Improved functionality for data validation and analysis

• New correlation plots for easy quality control assessment
• Auto-built volume and saturation sets for streamlined workflows
• Input filters for easier data management
• Refreshed probability density function management system

Workflow Automation Reduces manual effort and accelerates analysis

• Automates repetitive interpretation tasks
• Enables faster project delivery through integrated workflows
• Reduces time required for complex quantitative interpretation
• Supports self-service analysis for geoscientists

Benefits

Accelerated Analysis Faster quantitative interpretation and decision-making

• Accelerates analysis of regional information through machine learning efficiency
• Reduces time required for property prediction and trending
• Enables rapid model refinement and testing
• Delivers faster project timelines for exploration and production decisions

Improved Prediction Accuracy Enhanced subsurface property forecasting

• Improves analysis through geological knowledge integration
• Combines data-driven machine learning with domain expertise
• Reduces prediction uncertainty through constrained algorithms
• Enables more robust exploration drilling and field development decisions

Flexibility and Scalability Handles diverse data scenarios and project scales

• Processes datasets of varied size and volume efficiently
• Adapts to different regional geological settings
• Scales from small projects to large regional studies
• Accommodates heterogeneous subsurface data

Reduced Manual Effort Streamlines workflows and frees geoscientist time

• Automates quality control and knowledge generation
• Eliminates repetitive manual interpretation tasks
• Enables geoscientists to focus on higher-value analysis
• Reduces documentation and reporting burden

Enhanced Decision Support Better-informed exploration and production decisions

• Provides robust property predictions for well placement optimization
• Supports drilling hazard avoidance through improved pore pressure prediction
• Enables more confident field development planning
• Delivers immediate value to energy company operations

Integrated Geological Workflow Seamless disciplinary integration

• Combines petrophysics, rock physics, and machine learning in unified workflow
• Enables consistent analysis across multiple interpretation disciplines
• Supports multi-disciplinary team collaboration