An Iterative Inversion Method for Mineral Composition Evaluation in Offshore Shale Reservoirs Based on Elemental Logging

The precise quantification of mineral composition is a basis for the accurate geomechanical evaluation, brittleness assessment, and completion design of shale reservoirs. Currently, elemental logging has become an indispensable technical method for acquiring key mineral composition information. The standard industry practice involves solving an optimization problem to invert elemental dry weight fractions, measured downhole, into mineral contents. The accuracy of this inversion is fundamentally governed by a predetermined transformation matrix, which ideally requires rigorous calibration against a statistically robust suite of laboratory X-ray diffraction (XRD) analyses from rock samples. This prerequisite poses a significant constraint in poorly cored intervals, leading to substantial uncertainties in the derived mineralogy.

To reduce the core-dependency, the key innovation lies in formulating the element-to-mineral transformation as a joint inversion problem. The proposed algorithm operates by treating the transformation matrix not as a fixed input but as an optimizable variable within the inversion framework. Starting from a geologically reasonable initial model based on regional knowledge, the method employs an iterative optimization loop. In each cycle, it simultaneously adjusts the mineral volumes and refines the transformation matrix to minimize a dual-objective function. The misfit between the log-measured and model-predicted elemental yields, and a regularization term that constrains the matrix adjustments to physically plausible ranges. The iteration converges when the global error is minimized, yielding a formation-specific optimal transformation matrix alongside the final mineralogy.

The efficacy of the method was rigorously tested using data from offshore shale oil wells in China. Comparative analysis demonstrates that the mineralogical profiles produced by the iterative method achieve an excellent correlation with those derived from the XRD-calibrated approach in intervals where core data is available. More importantly, in zones lacking core control, the iterative method provides stable and geochemically consistent results. A detailed comparative analysis indicates that this method significantly enhances the prediction accuracy for critical brittle minerals such as quartz and plagioclase. The reduction in error for these key components directly translates to higher confidence in computed geomechanical properties.

In conclusion, this study presents a robust workflow that significantly enhances the reliability of mineralogical evaluation from elemental logs in core-constrained environments. The iterative inversion method reduces the critical need for extensive, expensive core-based calibration, offering a powerful and practical tool for the accurate and efficient characterization of offshore shale oil reservoirs. This advancement holds substantial value for optimizing drilling, completion, and stimulation strategies, thereby supporting the economical development of complex unconventional resources.