Internet Explorer 9 and Safari users must enable TLS in their browser settings (The new Ocean website uses TLS for enhanced security). Accept
* Internet Explorer version 9: Enable TLS versions 1.0. 1.1 and 1.2 in your Browser settings (Internet options > Advanced > Settings > Security).
* Safari: Enable TLS versions 1.0. 1.1 and 1.2 in your Browser settings (Preferences > Advanced Icon > Change Settings > Advanced Tab > Settings).


Ocean does not support Internet Explorer versions 8 and below. Get Latest Version
If you still have trouble, you should also clear your browser cache: press Shift-Ctrl-Delete [any browser, except Safari] and delete all items.
Further questions can be submitted via Contact Us
IPLAB Machine Learning Well Log Prediction3D
By IPLAB LLC
Plug-in Attributes
Platform:
Petrel
Lifecycle
Exploration
Challenges:
Unconventional Heavy Oil | Carbonates | Deepwater Exploration and Production | Unconventional Shale
ECCN:
Russia origin, EAR99
Version
  • 2017
Supporting Documents
Installation Guide
Release Notes
User Manual
Overview
Well log property prediction in a layer applying a set of seismic cubes and well logs in boreholes using machine-learning algorithms (linear regression, nearest neighbor, classic neural network, and T-neural networks). Virtual cubes are the result of this plug-in application with average, standard deviation, P10, P50, and P90 predictions per several cross validations calculations.
Specifications
    Learning algorithms allows using multiwell cross-validation options
  • Applying all well logs like one learning dataset. The output will be one virtual cube.
  • Applying each well to estimate separate predictions. The output will be a folder with several virtual cubes according to well number.
  • Applying each well to estimate separate predictions to calculate average, standard deviation, P10, P50, and P90 virtual cubes.
  • Applying each well to estimate separate predictions to calculate weighting average with reverse distance weights to well position.
Features
    Predictive algorithms
  • Linear regression
  • Nearest neighbour
  • Neural network
  • T-neural network
Additional Information
    For predictions, it is possible to use a set of seismic cubes
  • Full stack
  • Angle stacks
  • Azimuthal stacks
  • Seimic attributes (include inversion results)