Influence of ultra-deep solid bitumen on carbonate reservoir properties and gas saturation correction

• Digital cores were constructed using artificial thresholding and deep learning methods, respectively, and the digital cores were optimized from macroscopic, microscopic and analytical testing perspectives. • Based on the optimized model, two types of digital cores (bitumen-bearing and bitumen-free) were established. The influence of bitumen on carbonate reservoirs was clarified. • Using finite element resistivity modeling, correlations among resistivity, petrophysical parameters and bitumen content were established, and a correction model for resistivity and gas saturation was developed. • A method for evaluating the effectiveness of ultra-deep reservoirs and correcting gas saturation affected by bitumen has been established, providing a new research approach for the fine evaluation of bituminous reservoirs. Ultra-deep solid bitumen degrades reservoir physical properties and damages pore structures, complicating reservoir effectiveness evaluation and gas-saturation correction due to its high resistivity and low solubility. This study uses bituminous carbonate rocks from the second member of the Dengying Formation in the Pengla area of the Sichuan Basin as a case study. Microscopic thin-section observations, resistivity measurements, porosity and permeability testing, X-ray diffraction, and CT scanning experiments were conducted. These datasets were integrated with digital core reconstruction and finite element simulation to establish a comprehensive methodological framework for assessing the influence of solid bitumen on reservoir properties and evaluating reservoir effectiveness and gas-bearing potential. The results show that: (1) Digital core models containing bitumen were constructed using both manual threshold segmentation and deep learning–based modeling. Comparative analysis of digital cores, thin sections, and gas-measured porosity indicates that deep learning–based digital core model exhibits an average porosity deviation of 0.278% relative to measured values, and the threshold-based model shows an average difference of 0.852%. The deep learning model, therefore, more accurately reproduces reservoir pore characteristics. (2) Comparison between bitumen-bearing and bitumen-free digital cores demonstrates that the presence of bitumen reduces porosity by an average of 1.2%, permeability by an average of 0.876 mD, pore number by an average of 718, and throat number by an average of 321. These results confirm that solid bitumen significantly degrades reservoir physical properties and pore structure. (3) Finite element resistivity simulations were used to quantify the relationships among bitumen content, resistivity, and the cementation factor (m), enabling the calibration of petrophysical parameters specific to bitumen-bearing carbonate reservoirs. (4) Porosity and gas saturation correction models tailored for ultra-deep bituminous reservoirs were established. Before correction, the average discrepancies between log-derived and laboratory-measured values were 1.24% for porosity and 0.251 for gas saturation. After correction, these discrepancies decreased to 0.25% and 0.053, respectively. Overall, this study provides critical guidance for the effectiveness evaluation and gas-bearing assessment of ultra-deep bitumen-bearing carbonate reservoirs.