Petrel Classification and Estimation

Advanced train estimation tools and trend modeling

Overview

Petrel™ classification and estimation enables the use of neural networks and train estimation models to help estimate properties or probabilities.

The trend modeling processes enables the combination of up to five secondary properties, creating a realistic 3D trend property.

Using a supervised neural network, facies logs for all project wells have been automatically estimated based on petrophysical profiles.

Geometric trends conveerted to probability volumes help guide adaptive channels in the facies modeling process.

With neural networks you can combine several different fracture drivers in order to create a general trend property to guide fracture distribution.

Applications

By combining several different data objects or attributes, the module increases interpretation confidence and accuracy during reservoir modeling by ensuring integration of all available data, regardless of the domain.

Benefits

Lithology can be isolated and classified along the wellbore, on the basis of acquired and interpreted logs, and missing logs can be estimated based on other wells present in the project.
The 3D seismic facies classifications identify lithology according to combined seismic attributes.
Trend modeling capabilities, incorporate 1D, 2D, and 3D data, geological concepts, and analogs to enable users to reproduce realistic sedimentary environments.

Petrel Geophysics Classification and Estimation

Geometric trends converted to probability volumes help guide adaptive channels in the facies modeling process.

Features

A train estimation model is a set of tools for neural network analysis that enables users to train an estimation or classification model and store it as an object in the project.
Geometrical trend modeling is used to create a variety of geometrical properties from polylines to control the spatial distribution of facies and petrophysical properties.
By combining different fracture drivers, such as geometrical properties and seismic attributes, users create a final 3D property trend to guide fracture distribution in the 3D grid.

A final trend property combining several types of data including seismic trends can then be used as secondary input in the modeling process.

A train estimation model is a set of tools for neural network analysis that enables users to train an estimation or classification model and store it as an object in the project.

Geometrical trend modeling is used to create a variety of geometrical properties from polylines to control the spatial distribution of facies and petrophysical properties.

By combining different fracture drivers, such as geometrical properties and seismic attributes, users create a final 3D property trend to guide fracture distribution in the 3D grid.

Training