Fracture porosity is crucial for storage and production efficiency in fractured tight reservoirs. Geophysical image logs using resistivity measurements have traditionally been used for fracture characterization. This study aims to develop a novel, hybrid machine-learning method to predict fracture porosity using conventional well logs in the Ahnet field, Algeria. Initially, we explored an Artificial Neural Network (ANN) model for regression analysis. To overcome the limitations of ANN, we proposed a hybrid model combining Support Vector Machine (SVM) classification and ANN regression, resulting in improved fracture porosity predictions. The models were tested against logging data by combining the Machine Learning approach with advanced logging tools recorded in two wells. In this context, we used electrical image logs and the dipole acoustic tool, which allowed us to identify 404 open fractures and 231 closed fractures and, consequently, to assess the fracture porosity. The results were then fed into two machine-learning algorithms. Pure Artificial Neural Networks and hybrid models were used to obtain comprehensive results, which were subsequently tested to check the accuracy of the models. The outputs obtained from the two methods demonstrate that the hybridized model has a lower Root Mean Square Error (RMSE) than pure ANN. The results of our approach strongly suggest that incorporating hybridized machine learning algorithms into fracture porosity estimations can contribute to the development of more trustworthy static reservoir models in simulation programs. Finally, the combination of Machine Learning (ML) and well log analysis made it possible to reliably estimate fracture porosity in the Ahnet field in Algeria, where, in many places, advanced logging data are absent or expensive.

The net-effective stress is a fundamental physical property that undergoes dynamic changes in response to variations in pore pressure during production and injection activities. Petrophysical properties, including porosity, permeability, and wave velocities, play a critical role and exhibit strong dependence on the mechanical stress state of the formation. The Williston basin’s Bakken Formation represents a significant reservoir of hydrocarbons within the United States. To investigate this formation, we extracted core plugs from three distinct Bakken members, namely Upper Bakken, Middle Bakken, and Lower Bakken. Subsequently, we conducted a series of measurements of ultrasonic compressional and shear wave velocities, as well as pulse decay permeabilities using nitrogen, under various confining pressures employing the Autolab-1500 apparatus. Our experimental observations revealed that the ultrasonic wave velocities and permeability display a significant sensitivity to stress changes. We investigated existing empirical relationships on velocity-effective stress, compressional-shear wave velocities, and permeability-effective stress, and proposed the best models and associated fitting parameters applicable to the current datasets. In conjunction with the acquired datasets, these models have considerable potential for use in time-lapse seismic monitoring and the study of production decline behavior. The best fitting models can be used to forecast the petrophysical and geomechanical property changes as the reservoir pore pressure is depleted due to the production, which is critical to the production forecast for unconventional reservoirs.

Fossil fuels, such as coal and hydrocarbons, are major drivers of global warming and are primarily responsible for worldwide greenhouse gas emissions, including carbon dioxide CO₂. The storage of CO₂ in deep saline reservoirs is acknowledged as one of the top practical and promising methods to reduce CO₂ emissions and meet climate goals. The North Dakota Industrial Commission (NDIC) recently approved the fourth Class VI permit for a carbon capture and storage project in the Williston basin of North Dakota for the geological CO₂ storage in the Broom Creek formation. The current research aimed to conduct a comprehensive petrophysical characterization and rock physics modeling of the Broom Creek deep saline reservoir to unravel the mineralogical distribution and to understand the variations in petrophysical and elastic properties across the formation. This study utilized geophysical well logs, routine core analysis, and advanced core analysis to evaluate the Broom Creek formation. Multimineral petrophysical analysis calibrated with X-ray diffraction results reveals that this formation primarily comprises highly porous clean sandstone intervals with low-porosity interspersed with dolomite, anhydrite, and silt/clay layers. The formation exhibits varying porosities up to 0.3 and Klinkenberg air permeabilities up to ∼2600 mD. The formation water resistivity using Archie’s equation is approximately 0.055 ohm-m at 150 °F, corresponding to around 63,000 ppm NaCl salinity, which is consistent with prior data. The pore throat distribution in the samples from clean sandstone intervals is primarily situated in the macro-mega scales. However, the presence of anhydrite and dolomite impedes both porosity and pore throat sizes. The accurate prediction of effective elastic properties was achieved by developing a rock physics template. Dry rock moduli were modeled using Hill’s average, while Berryman’s self-consistent scheme was employed for modeling saturated moduli.

The purpose of this work is to compare two fracture prediction models with real-world data. The pure Artificial Neural Network (ANN) model emphasizes regression analysis, while the hybrid model (SVM-ANN) focuses on the combination of regression and classification analysis or Support Vector Machine. The results were subsequently tested against logging data by combining the Machine Learning approach with advanced logging tools. In this context, we used electrical image logs and the dipole acoustic tool which together allowed the distinction of 404 open fractures and 231 closed fractures and, consequently the estimation of fracture porosity. The results are then fed into two machine-learning algorithms. Pure Artificial Neural Networks and hybrid models are used to establish comprehensive results, which are subsequently tested to check the accuracy of the models. The outputs obtained from the two methods demonstrate that the hybridized model has a lower Root Mean Square Error (RMSE) than pure ANN. The results of our approach strongly suggest that incorporating hybridized machine learning algorithms in fracture porosity estimations can contribute to the development of more trustworthy static reservoir models in simulation programs. Finally, the combination of Machine Learning (ML) and well-log analysis do permit reliable estimation of fracture porosity in the Ahnet field in Algeria, where, in many places, advanced logging data is absent and costly.