2.5-D Deep Learning Inversion of LWD and Deep-Sensing EM Measurements Across Formations With Dipping Faults

Deep learning (DL) inversion of induction logging measurements is used in well geosteering for real-time imaging the distribution subsurface electrical conductivity. We develop a DL workflow to solve 2.5-D inverse problems arising geosteering. The employs three modules: “look-around” fault detection module and two modules reconstructing anisotropic resistivity models presence or absence planes, respectively. Our approach capable detecting quantifying arbitrary dipping planes real time. compare performance considering only short logging-while-drilling (LWD) versus using both LWD deep-sensing measurements. latter provide enhanced depth-of-investigation while minimizing uncertainty. also obtain improved results when multidimensional inversion, especially nearby planes. This study verifies applicability across faulted formations