Artificial Intelligence-Based Model of Mineralogical Brittleness Index Based on Rock Elemental Compositions

Brittleness Index (BI) is among the key parameters used to evaluate zones of interest or “sweet spots” before drilling and hydraulic fracturing in unconventional reservoirs. Several methods can be used to calculate the BI, including mineral-based, log-based, and elastic-based approaches. This work presents a new approach that can provide a fast, accurate, and continuous profile of the brittleness index, which would significantly aid unconventional resource characterization and identification. Different artificial intelligence tools were utilized in this work to come up with the best model for evaluating the rock brittleness based on elemental data acquired using x-ray fluorescence instruments. Artificial neural networks, support vector machines, and fuzzy logic techniques were used to develop new models that can provide continuous profiles of quartz, calcite, and clay minerals using profiles of few elements. After that, the mineral-based brittleness index (MBI) was estimated using the predicted profiles of quartz, calcite, and clay minerals. Moreover, a new model was developed to assess the MBI profile based on the elemental compositions (Na, Al, Si, K, and Ca), without a need to determine the rock mineralogy. The results showed that the developed models could provide accurate predictions for the mineralogical profiles and brittleness index, with R² higher than 0.90. The proposed equation can be used to estimate the rock brittleness based on the logging data.