How Can Hybrid Physics and AI/ML Analysis of Formation Evaluation Data Benefit Reservoir Navigation?

Summary Geonavigation while drilling is a challenging task with applications ranging from steering in oil-gas bearing sands to mapping subsurface geothermal resources and gas storage. In this paper, we present the Energy-based inverse mapping, a hybrid physics + AI/ML approach, to tackle the uncertainties arising in logging while drilling (LWD) shallow-to-ultra deep electromagnetic (EM) tools and attempt to quantify the value of information. The new and improved capabilities that are considered include: probabilistic characterization of the impact of degrees of freedom, low-attention data- and work- flows, anomaly detection, early warning, first-hand classification, continuous risk analysis, data-driven decision making, advisory, transfer and federation learning and cost-reduced footprint of numerical evaluation. Future work will be aiming the new 3D formation classes with the strong features, application development and early customer support.