Geosteering interpretation using bayesian inference

Abstract

A geosteering subsystem has been designed to detect patterns in measurements that correlate to geological events and use statical analysis to interpret the geological events corresponding to control points and update a geological model based on the statistical analysis-based interpretations. The subsystem evaluates obtained measurements against predetermined patterns that correlate to geological events. If a pattern is detected, the subsystem labels the corresponding wellbore position and/or measurements as a control point and passes on the measurements for interpretation. The subsystem generates interpretations of the control point based on the measurements and historic control points. For each interpretation, the subsystem generates a pseudo log based on offset logs. The subsystem then generates a posterior probability for each interpretation based on evaluating the corresponding pseudo log with respect to the measurements and formation knowledge. High posterior probability interpretations are selected for updating the geological model.