Seismic driven reservoir classification using advanced machine learning algorithms: A case study from the Lower Ranikot/Khadro sandstone gas reservoir, Kirthar Fold Belt, Lower Indus Basin, Pakistan
“…For example, AI and SI are often used in conjunction with seismic inversion techniques to create high-resolution images of the subsurface, while Vp/Vs is frequently used to identify hydrocarbon reservoirs and structural traps. Lambda-rho and mu-rho are often used in rock physics modeling to predict rock properties and aid in the interpretation of seismic data [ 13 , 41 , 53 ]. Fig.…”
“…For example, AI and SI are often used in conjunction with seismic inversion techniques to create high-resolution images of the subsurface, while Vp/Vs is frequently used to identify hydrocarbon reservoirs and structural traps. Lambda-rho and mu-rho are often used in rock physics modeling to predict rock properties and aid in the interpretation of seismic data [ 13 , 41 , 53 ]. Fig.…”
“…One of the most crucial considerations in the appraisal of a formation is the distribution of the shale, as the presence of shale affects the reservoir’s effective porosity and permeability. The γ-ray index is used to find the amount of clay inside the formations, and this information is then used to calculate the volume of the shale. − …”
Petrophysical analysis is an industry-standard practice for reservoir evaluation as it provides critical inputs for characterizing subsurface formations and estimating resource potential. Khadro/Ranikot Formation sands are proliferous producers in the Central Indus Basin, Pakistan. The demarcate potential in intercalated sand shale layers that are thin and heterogeneous makes it a challenging reservoir. Conventional petrophysical interpretation is laborious and does not produce upto-mark results due to reservoir complexity, data limitations, and associated uncertainties. Hence, an emerging and delicate machinelearning (ML) approach has been comprehensively applied to analyze the potential and robustly interpret well log data while addressing the associated challenges. This case study entails a thorough evaluation of well log quality, assessing several algorithms such as least-squares support vector machines (one-class SVM), Random Forest Regressor (RFR), Extra Tree Regressor (ETR), Gradient Boosting Regressor (GBR), Decision Tree Classifier (DTC), etc. to compare their efficacy and reliability. One-class SVM helps to reduce outliers with great certainty, while the missing logs sonic (DT) and density (RHOB) are precisely predicted via GBR and ETR with 0.66 and 0.88 R 2 , respectively. Hence, providing reliable and optimized quality logs suitable for ML-based petrophysics. ML worked on these augmented logs by dividing the data into 60% training and 40% testing. The ETR outperformed the rest of the models with a correlation of 0.99 and 0.91 among conventional and ML results. Likewise, RFR performed exceptionally well for water saturation modeling, expressing the highest 0.93 correlation. Finally, DTC modeled reservoir facies with the best 91% accuracy and 0.935 F1 measures at the blind well. Excellent calibration of >85% is met with the estimates obtained by the predictive model compared to conventional methods. This comprehensive approach offers cost-effective and robust workarounds for modern formation evaluation with minimal uncertainty and resource-efficient multiwell interpretation within complex reservoirs and sets the stage for further research in the machine-learning ecosystem.
“…In this sense, inversion can also be considered a sophisticated method of integrating well logs and seismic data. Seismic inversion can be divided into two techniques: post‐stacking and pre‐stacking inversion (Bosch et al, 2010; Alvarez et al, 2015; Anees et al, 2022; Naseer et al, 2022; Manzoor et al, 2023; Naseer et al, 2023). In this study, the post‐stack inversion was used due to the lack of stacking data.…”
De‐risking the hydrocarbon potential in near‐field exploration is one of the most important procedures in the exploration of hydrocarbons, and it requires the integration of various data to predict the reservoir characteristics of the prospect area more accurately. In this work, wells and 3D seismic data from the Libyan producing oil fields were utilized to demonstrate how well this technique worked to improve and describe the hydrocarbon potential of the carbonate geobody that corresponds to the Palaeocene Upper Sabil Formation, which was revealed by new seismic data. This study integrates different types of data, including 3D seismic, seismic acoustic impedance, depositional history and geostatistical analysis, to predict the facies, reservoir porosity and permeability distributions and then visualize them in a 3D reservoir model. The 3D seismic data analysis revealed the presence of a clear seismic anomaly geobody (GB) that has never been penetrated by any well. The sedimentological analysis for the well adjacent to the GB indicated a deep‐water depositional environment as turbidites surrounded by deep‐water mud dominated facies. The Upper Palaeocene interval in the study area was subdivided based on the depositional facies and seismic stratigraphy into eight zones that were used to build the reservoir model framework. According to the porosity permeability relationships, the carbonate facies has been classified into five E‐Facies, that is, soft highly argillaceous limestone, hard argillaceous limestone, porous limestone (<20% porosity, and >30% shale volume), medium quality limestone (10–20% porosity, and >30% shale volume) and tight limestone (<10% porosity, and >30% shale volume). The rock physics and inversion feasibility analysis indicated that the acoustic impedance (AI) can be used to predict the porosity but not the lithology or the fluid content. The Bayesian classification has shown excellent results in predicting and modelling the reservoir facies distribution within the study area, utilizing the integration of gross depositional maps (GDEs), wells and seismic data. The reservoir quality of the GB was predicted by using the post‐stack seismic inversion, which indicated a high porosity interval (25%–30%). Moreover, the statistical analysis integrated with the well and seismic data was used to predict the GB permeability. The predicted permeability was reasonably high (40–60 mD). The final E‐facies show an excellent match with the input well data and an excellent match with the blind wells that were used for result quality control (QC) with higher vertical resolution. The developed model can be used as a guide for de‐risking the studied GB hydrocarbon potential in the studied basin, and it can be applied in other similar geological conditions worldwide for exploring underexplored reservoirs and de‐risking their hydrocarbon potential.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
