Probabilistic Decision Support for Geosteering Under Geological Uncertainty

This work presents an ensemble-based workflow for geosteering optimisation under geological uncertainty. In many reservoirs, geosteering is performed by updating a single geomodel during drilling and steering the well to stay in the reservoir or near the top of the target layer. In the proposed novel approach, instead of forcing one interpretation to fit the data, we maintain a finite ensemble of plausible 3D static geomodel realisations and update their probabilities as logging-while-drilling (LWD) or measurement-while-drilling (MWD) information arrives.

We present a practical integrated decision-tree based workflow for optimising horizontal well trajectories under geological scenario uncertainty. We demonstrated that, given a properly designed uncertainty ensemble, the workflow can extract value from incoming information by updating model probabilities and improving the steering well path. We treat this value as decision support for the geosteering team: it helps resolve complex situations and can make decisions faster and more confident.