ML-based deconvolution method for high-resolution fast inversion of induction log data in deviated wells

Abstract

We built a deconvolution model for 1D induction log data in deviated wells using Machine Learning. Unlike iterative forward modeling inversion methods, the deconvolution model is extremely fast. Unlike linear deconvolution models in the past, the Machine Learning (ML)-based deconvolution finds the accurate layer resistivity and layer boundaries. For a unit induction tool 2C40 in a deviated well of deviation angle θ, the 10/cosθ-ft window deconvolution model would work satisfactorily.