Techlog Ipsom

Predict and propagate rock classification groups—use the 2D indexed and probabilized self-organizing map to help your geological interpretation of log data and facies prediction.

Support Contact Us Related Resources

Overview

Advanced classification

The Ipsom module's capability to predict and propagate classification groups helps to improve interpretation workflows. Its 2D indexed and probabilized self-organizing map (SOMap) is designed for use in:

Geological interpretation of well log data and facies prediction.
Optimal derivation of petrophysical properties (Ф, K, hydraulic units, etc.).

The 2D indexed and probabilized self-organizing map at initialization phase.

The sorting and organization phase of the grid of nodes.

Indexation phase with facies core description; each neural unit is given a geological meaning.

Applications

Comprehensive statistical tools

Results are quantified and refined through an iterative process during the final control and validation phase. The statistical and graphic tools are easy to use and lead to a better understanding of coherence between core description, log responses, and consistency on the map. This allows for the:

Insertion of Python scripts within the workflows, greatly extending the module’s flexibility.
Application of different weighting to the various inputs.
Projection of indexation samples within the SOMap/nodes, helping the user understand how the prototypes were populated during the indexation phase; transition or “ghost” groups are then clearly identified.

Benefits

Accurate model calibration

Calibration is key to the classification process. Its aim is to give each neural unit an actual geological meaning; a process known as the indexation phase. Each neural unit is then assigned a code that corresponds to a geological interpretation (detailed core description). This stage is essential for direct interpretation of electrofacies in sedimentological terms.

Facies output curve associated with a probability of occurrence of predicted facies at each depth.

Features

Outputs

Optimized partition of log data calibrated to facies described in core.
Facies output curve associated with a probability of occurrence of predicted facies at each depth.
Automatic batch reports for parameter tables and results plots.

Training