IP for Machine Learning

IP for Machine Learning

IP for Machine Learning

IP for Machine Learning

 

Our IP Machine Learning bundle brings together leading algorithms for performing classification and curve prediction based tasks.

Carry out data repair, prediction of missing curves and key reservoir properties from minimal source data using our Neural Networks, Fuzzy Logic and Multiple Linear Regression modules.

If you are looking to classify your well log data into distinct facies or rock types, then our Cluster Analysis and Self-Organizing Maps modules have you covered.

Unique to IP is our Domain Transfer Analysis (DTA) module which provides a mathematical advantage over other methods for prediction of missing and incomplete data. Using n-dimensional Partial Differential Equations, non-linear relationships can be revealed that other tools miss - giving its results that vital extra edge of accuracy.

A common interface between the modules allows fast and efficient model comparison.

 

Download The Flyer
IP-Machine Learning-1

Curve Prediction

Applicable at any stage of a well's life, Curve Prediction enables advanced interpretations from minimal source information. Employing several statistical methods, Curve Prediction's tools help you to generate new curves from offset wells, repair existing data, or even make continuous curves from discrete data.

  • Predict key reservoir properties including porosity, permeability and saturation from log curves and core data
  • Infer missing data or repair incomplete well data from bad hole intervals

 

Pictured: Crossplot your prediction results against actual measurements.

 

curve-prediction

Rock Typing

Rapid and easy to learn, Rock Typing employs powerful algorithms to build repeatable and accurate facies classifications. With both Cluster and Self-Organising Map tools, you can analyse core or raw log data to cluster, zone and focus on the best approach to lithology, porosity and saturation.

  • Build and apply a facies model to any number of wells
  • Cross-plots and starplots graphically show how key input parameters affect facies selection
  • Compare any Rock Typing models in Contingency Table, including core data and manual picks

 

Pictured: Visualise the results of Self Organising Maps in 3D or 2D.

 

rock-typing

Domain Transfer Analysis

Domain Transfer Analysis (DTA) is our premier solution for porosity, permeability and saturation prediction. Create multivariate and non-linear predictions based on real data from the current well, nearby wells or their sidetracks – you can even include previously unused information from normal deterministic analyses.

DTA's key advantage is mathematical. It employs n-dimensional Partial Differential Equations to reveal non-linear relationships that other tools miss – giving its results that vital extra edge of accuracy.

 

Pictured: Setup and Zone Selection within the DTA module.

 

dta-image