Fast inversion of logging-while-drilling azimuthal resistivity measurements for geosteering and formation evaluation

Abstract

We developed an efficient parametric inversion method for the determination of the distances to boundaries (DTBs), dip angle, and formation anisotropic resistivities from logging-while-drilling (LWD) azimuthal resistivity measurements (ARM) in real time. To increase the efficiency, the development employed a depth window, which will slide along the well trajectory during data processing. The local structure inside the depth window can be regarded as one-dimensional (1D) layered formation. A three-layer model (one transverse isotropic (TI) layer embedded between two isotropic shoulder layers) is assumed in the inversion to rebuild the formation structure. Then, the inversion problem is solved using a regularized Levenberg-Marquardt minimization approach. Multiple initial guesses are employed in order to avoid local minima and to find the global solution. Synthetic experiments show that the developed algorithm, which is capable of handling a large subset of the complex 3D conditions and arbitrary well trajectory, is robust and efficient. Discussions are also made to further demonstrate the advantages and limitations of ARM and the inversion processing in geosteering and formation evaluation.