Geomodelling TRIZ Cases
- Fully Convolutional Neural Network Magnetotelluric Inversion
Pre-trained synthetic data enables the network to resolve local extremum issues and reduce operation time while enhancing fitting precision.
- Variable Discretization for Reservoir Flow Simulation
A variable discretization method combines finite element and finite difference techniques within a single computational framework.
- Automated Well Placement Planning in Reservoir Simulation
Reservoir simulator ranks candidate well locations by predicted production rate to reduce manual planning time.
- Porosity Prediction via Variable Compression Index
A sediment porosity model uses a variable compression index to adapt compaction behavior across different void ratio ranges.
- Seismic Data Inversion With Perturbation-Based Uncertainty Analysis
Processes 4D seismic data through multiple model perturbations and statistical analysis to quantify uncertainties in reservoir property estimates.
- Machine Learning Reservoir Model Switching
A machine learning classifier selects reservoir models based on geomechanical simulation results.
- Geophysical Inversion via Convolutional Neural Networks
Convolutional neural networks extract subsurface physical properties from geophysical survey data, reducing computational time and improving accuracy.
- Continental Slope Foot Point Recognition via Terrain Grid Derivation
Automated terrain grid processing extracts continental slope foot points through multi-stage derivative analysis and geometric simplification.
- Non-linear Petrofacies Identification via Cross-Plot Partitioning
Automated cross-plot partitioning defines non-linear petrofacies boundaries using data frequency analysis for systematic reservoir characterization.