Accurate inversion of directional electromagnetic logging-while-drilling (LWD) measurements plays an essential role in geosteering, a technique to actively adjust the borehole position on the fly to reach geological targets. In this paper, a new scheme based on Supervised Descent Method (SDM) for solving this inverse problem is proposed. Combining the conventional gradient based inversion and machine learning scheme, SDM has the flexibility to incorporate prior information, capability of skipping local minima, and accelerated convergence. Besides, by utilizing real-time feedback obtained from the logging process, the learned descent directions can be slightly revised with high efficiency to get closer to the true model. Further, the generalization ability is also explored. Numerical examples demonstrate SDM based inversion can achieve a higher resolution and faster convergence than conventional Occam’s inversion.
Presentation Date: Monday, October 12, 2020
Session Start Time: 1:50 PM
Presentation Time: 2:40 PM
Location: Poster Station 2
Presentation Type: Poster
