AI-Driven Lithotype Classification and Reservoir Correlation Utilizing Drill Cuttings: A Mesozoic Case Study from Europe

Summary This study demonstrates how the integration of AI-driven techniques can transform drill cuttings into a powerful tool for reservoir characterization. A proprietary workflow that builds a detailed lithotype model by combining X-ray fluorescence (XRF) elemental data and high-resolution image analysis was used to analyze 278 samples from two European wells that targeted Mesozoic reservoirs. Standardized imaging in both white light and ultraviolet light was part of the laboratory workflow, and 32 elements were subjected to XRF analysis. RGB and YUV color information, as well as particle metrics, were extracted from the images. Combining XRF and Image data, the results were used to classify samples into silicarich or carbonate-rich domains (via Si/Ca ratio) and further subdivided into 14 lithotypes based on brightness values. The identification of depositional cycles, patterns, and Gross Sedimentary Packages (GSPs) was made possible by this classification. For example in Well B, cleaner intervals in the middle GSP suggested higher reservoir quality, whereas dark-colored, clay-rich, and reducing conditions in the upper GSP may be advantageous for hydrocarbon preservation. With the aim of improving well planning and reservoir management throughout the region, this workflow provides a quick, quantitative, and repeatable method for understanding depositional environments and refining geological models.