This paper presents the methodology used to deliver a consistent set of rock types, with petrophysical outputs of porosity, water saturation, and permeability, using a class-based machine learning (CbML) method. This novel tool is designed for summarizing the unique characteristics of well log data.

A nuclear magnetic resonance (NMR) study was planned for a gas field offshore Western Australia. The study objective was to integrate all NMR data acquired in the field and provide a consistent set of inputs for use in a field model.

To overcome NMR data normalization challenges, the CbML method was incorporated into the workflow. The objective of this paper is to present how CbML was utilized, and to show the flexibility of using the CbML method to solve a petrophysical challenge.

This content is only available via PDF.
You can access this article if you purchase or spend a download.