Artificial-Intelligence-Based Well Placement Look Ahead Concept

Abstract

Abstract Artificial intelligence (AI) brings many benefits to the oil and gas industry and in particular the automation of the geosteering process with structural prediction ahead of the bit to optimize horizontal well placement and reduce the operational challenges. AI techniques are based on data analysis that learn from data, identify patterns and make predictions without the need of direct human intervention. Recent ultra-deep azimuthal resistivity (UDAR) logging-while-drilling (LWD) tools have reached a significant depth of investigation capabilities to provide real time information that covers a large volume of the rock around the wellbore. The objective of this paper is to provide details on the automated geosteering methodology used to detect formation changes in real time ahead of the bit based on AI approach. The new automated geosteering approach is based on a suite of real-time enabled pre-drill and data integration workflows. The latter combines UDAR technology with seismic data. The suite of both workflows as sequential process allows for precise well planning and prediction of a structural model ahead of the bit in real-time. At first, the pre-drill workflow consists of data quality assessment, preconditioning, building 3D tectonic model repository and signal consistent 3D stratigraphic model incorporating offset wells which is constrained by tectonic model. During the second phase with the data integration workflow, ultra-deep azimuthal resistivity inversion output, generating synthetic seismic, extracting from 3D seismic volume along well-path, estimating depth mismatch through non-rigid matching, apply displacement to real seismic and model repository. Therefore, enabling a structural prediction ahead of the bit in 2D and 3D and aiding geosteering decisions in real-time. The new well placement look ahead concept will enable a right strategy that will enhance the geosteering decisions in real-time. Examples highlighting the value of the AI-based well placement look ahead concept are provided in the paper. Finally, this AI-approach will evaluate the potential added value to the well placement activity. The proposed workflow for look ahead geosteering based on real time integration of LWD measurements with surface seismic is a major step towards unlocking new horizons of providing critical information hundreds of feet ahead of the bit. This proposed technique will allow the reduction of subsurface uncertainty and helps the achievement of geosteering objectives.