Improving the classification of rock samples with diffusion probabilistic machine learning

Artificial intelligence (AI) and machine learning (ML) have transformed numerous domains by enabling systems to learn from data and make intelligent decisions. Within the field of ML, probabilistic machine learning has gained significant attention due to its ability to capture uncertainty and provide predictions based on probabilities. This essay explores the concept of probabilistic machine learning and its applications in AI. We delve into the fundamentals of machine learning, discuss probabilistic modeling, explore various probabilistic machine learning techniques, highlight the advantages of probabilistic approaches, and examine real-world applications. Additionally, we address the challenges associated with probabilistic machine learning and discuss future directions in this exciting field.

Pressure drawdown management (PDM) has been heavily studied in offshore and onshore environments as a way to improve the well productivity and in general Estimated Ultimate Recovery (EUR). There has been a great interest in this topic, especially in unconventional oil and gas reservoirs, where multi-stage hydraulic fracturing and horizontal well technology is used to enhance oil and gas production. Recent studies on Inflow Performance Relationship (IPR) of shale oil/gas reservoirs have shown that unmanaged pressure drawdown (UPD) can potentially impact well productivity and profitability in different ways. Hydraulic fracture conductivity impairment is an example and can occur through proppant crushing, embedment, and fine migration. Uncontrolled pressure drawdown can also cause geo-mechanical effect in high-pressured reservoirs. Aggressive pressure drawdown has shown detrimental impact to EUR, flow capacity, and economics of the wells in Haynesville, Eagleford, and Point Pleasant/Utica shales. Wilson (2015) used a coupled geo-mechanics model with a large Haynesville data set that showed that effective stress on fractures is influenced by operational changes, and managed pressure drawdown can improve gas recovery. Deen et al. (2015) also illustrated that aggressive pressure drawdown can damage a well's performance in the Eagle Ford, while conservative pressure drawdown can impact the near term economic value. In shale reservoirs the impact of pressure drawdown management on well performance and EUR is mostly related to pressure dependent permeability. Unmanaged pressure drawdown has also shown rapid transition from transient flow to boundary dominated flow in unconventional gas reservoirs. There is a balance between the rate in which the wells are produced and economic feasibility of the wells over time. Optimal economic production rate must be found for each area. The goal of this study is to analyze well productivity between wells that use managed drawdown versus unmanaged drawdown in over pressured formations such as Utica/Point Pleasant. The production performance of the wells will be analyzed using various reservoir techniques including superposition plot, diagnostic plots, history matching, and decline curve analysis. In addition, economic sensitivity will be run to find out the incremental increase in EUR needed to justify the delay in production. The optimum economic rate to produce the reservoir while maintaining the integrity of the reservoir is the ultimate goal producing from shale gas reservoirs, which will provide the maximum value for the shareholders.

At present, the development of unconventional and low-permeability oil and gas reservoirs has become a hot spot of energy development at home and abroad. The typical feature of this kind of oil and gas reservoirs is that the reservoir permeability is very low, and certain transformation measures need to be taken before they can be put into operation. The reservoir volume transformation technology is one of the important technologies for the effective development of unconventional oil and gas reservoirs. In recent years, in order to improve the development effect of ultra-low permeability reservoir, Hailar complex fault block oilfield has carried out field application tests of volume transformation technology in different types of reservoirs year by year. By increasing the amount of fracturing fluid and fracturing sand, increasing the fracturing sweep volume and improving the reservoir seepage conditions, the effective production of the reservoir is realized. However, from the perspective of fracturing effects of different reservoirs, there will be great differences in fracturing effects under the same fracturing method. It shows that the effect of large-scale fracturing development of oil and gas reservoir is related not only to the fracturing mode, but also to the geological characteristics of the fracturing horizon. The in-depth study of the factors affecting the fracturing effect has important guiding significance for the development of low-permeability reservoirs.

The Junggar basin is rich in oil and gas resources, and recently, unconventional oil and gas reservoirs have become important targets. Significant amounts of oil and gas are located in Permian formations in the Shazhang–Dajing area, however, fewer studies have focused on unconventional oil and gas reservoirs, particularly on the factors controlling the distribution of unconventional reservoirs. In this study, a sequence stratigraphic framework of the study area was developed based on cores, well logs and seismic data. Source rock samples were analyzed, including tests of kerogen, vitrinite reflectance, chloroform bitumen, organic carbon and group composition. The physical properties and pore characteristics of reservoir samples were also analyzed using scanning electron microscope and high resolution electron microscope. The Permian Pingdiquan Formation comprises three thirdorder sequences in which individual sequence can be divided into a low-stand system tract (LST), a transgression system tract (TST), an early high-stand system tract (E-HST), and a late high-stand system tract (L-HST). Several source rock lithologies have high total organic carbon (TOC), and are matured, having excellent hydrocarbon generation potential. Clastic and dolomitic rocks are the main reservoirs, and nanometer-scale pores are the main reservoir space in the tight reservoirs. The distribution of tight oil is controlled by the sequence stratigraphy, palaeogeomorphic units, and sedimentary facies. The fine sediments, which were generally deposited near the initial flooding surface (FFS) and the maximum flooding surface (MFS), are favorable zones for tight oil. The distributions of the hydrocarbon source rocks are controlled by the depositional environment, depending on the palaeogeomorphic units; the deep depression in the Huoshaoshan area and the moderately-deep depression in the Shishugou area are two distinct hydrocarbon generating centers. The distribution of tight oil is clearly controlled by the sedimentary facies; the tight oil is mainly located in the fan delta front and lacustrine depositional environments. The distribution of tight oil in this area is summarized, and favorable areas of tight oil are proposed.

Volcanic rocks are unconventional oil and gas reservoirs, and the pore structure of the rock is relatively complex. In this study, we selected samples from CS105 well and DS9 well from Yingcheng Formation volcanic reservoirs in the southern Songliao Basin to study the pore structure. First, we performed experiments on two rock samples using computed tomography (CT), and then measured the porosity and density of the two rock samples using mercury intrusion porosimetry (MIP). According to the data obtained by micron computed tomography (Micro-CT), Avizo 2019.1 software was used to perform three-dimensional (3D) reconstruction on the micron level of the two wellhead samples. Analyze the microscopic pore structure of the rock sample and compare it with the porosity of the sample obtained by the mercury injection method. Compared with MIP, Micro-CT more accurately characterizes the pore structure and porosity of volcanic rocks.

Accurate modeling of hydrocarbon production is a necessary, yet challenging step for economic exploitation of unconventional resources. One of the main challenges is to model flow to a horizontal well from a complex network of hydraulic and natural fractures. Many unconventional reservoirs comprise well-developed natural fracture networks with multiple orientations and complex hydraulic fracture patterns based on microseismic data. Conventional dual porosity and dual permeability models are not adequate for modeling these complex networks of natural and hydraulic fractures. Also, it is neither practical nor advantageous to model a large number of pre-existing fractures with a discrete fracture model. Therefore, an appropriate approach to model production from low-permeability reservoirs is to perform discrete fracture modeling for hydraulic fractures and employ a dual continuum approach for numerous natural fractures. We have developed a coupled dual continuum and discrete fracture method to simulate production from unconventional oil and gas reservoirs. Large-scale hydraulic fractures (macro-fractures) are modeled explicitly using a discrete fracture model, called EDFM, and numerous small-scale natural fractures (micro-fractures) are modeled using a dual continuum approach. The hybrid model includes three domains: matrix, discrete-fracture, and continuum-fracture domains. A systematic approach is devised to calculate transport parameters between all three domains. Moreover, EDFM allows for not only transverse and longitudinal hydraulic fractures but also macro-fractures of any orientation. Thus, the coupled model provides an effective and reliable environment to improve stimulation designs and completion strategies. We present several examples in this study to show the applicability, robustness, and performance of the hybrid method for the simulation of unconventional oil and gas reservoirs. We examine multi-stage hydraulic fractures with multiple configurations in the presence of numerous pre-existing fractures. Simulations show a noticeable contribution from natural fracture networks on total production. Furthermore, for the tight oil reservoir examined in this study, the stimulation scheme with longer hydraulic fractures improves cumulative oil production compared to the scheme with larger number of shorter hydraulic fractures. We also examine production from a tight gas reservoir wherein hydraulic fractures partially penetrate the formation height. Simulations indicate that inefficient fracture treatment can result in significant loss of production.

With the continuous exploration and development of unconventional oil and gas reservoirs, volume fracturing technology becomes one of the necessary measures for developing shale gas and tight sandstone gas reservoirs effectively. Volume fracturing technology usually uses slickwater and drag-reducing agent as the core of the fracturing system. The composition of the fracturing system is the main factor determining its performance. Polyacrylamide has many amide groups in its main chain, high activity, and controllable performance, often in solid powder and liquid emulsion states. Furthermore, polyacrylamide which is the water-soluble drag-reducing agent is most widely used in applying current shale gas slickwater fracturing operations. Due to the low viscosity and poor sand-carrying capacity of slickwater, proppant easily settles at the bottom of hydraulic fractures. This phenomenon influences the stimulation effect of volume fracturing. Therefore, the law of sand carrying and placement of proppant in hydraulic fractures in volume fracturing plays an essential role in determining the success of the stimulation effect of volume fracturing. Through the visualization device of proppant transport in the fracture, the settlement of proppant in the fracture was studied experimentally. And through experimental equipment, the effects of different operation pumping rates, liquid viscosity, proppant type, and proppant pumping schedule on the stimulation effect were studied. The experimental results can provide strong support for volume fracturing into well material optimization and operation parameter optimization for unconventional oil and gas reservoirs.

The utilization of learning analytics to identify at-risk students for early intervention has exhibited promising results. However, most predictive models utilized to address this issue have been based on non-probabilistic machine learning models. In response, this study incorporated probabilistic machine learning for two reasons: (1) to facilitate the inclusion of domain knowledge, and (2) to enable the quantification of uncertainty in model parameters and predictions. The study developed a five-stage, probabilistic logistic regression model to identify at-risk students at different stages throughout the academic calendar. Rather than predicting a student’s final or exam mark, the model was focused on predicting the at-risk probabilities for subsequent assessments—specifically, the probability of a student failing an upcoming assessment. The model incorporated student engagement data from Moodle, as well as demographic and student performance data. The study’s findings indicate that the significance and certainty of student engagement and demographic variables decreased after incorporating student-performance variables, such as assignments and tests. The most effective week for identifying at-risk students was found to be week 6, when the accuracy was 92.81%. Furthermore, the average level of uncertainty exhibited by the models decreased by 60% from stage 3 to 5, indicating more reliable predictions at later than earlier stages. The study highlights the potential of a probabilistic machine learning model to aid instructors and practitioners in identifying at-risk students, and thereby to enhance academic outcomes.

The province of Neuquen in Argentina has public domain over one of the largest unconventional oil and gas reservoirs in the world. In 2012, with the promulgation of the Decree 1483/12, the provincial government made the first law for the exploration and exploitation of unconventional oil and gas in the country. In order to “prevent, mitigate and minimize environmental impacts” of hydraulic fracturing the Decree prohibits the use of underground waters for the exploration and exploitation of unconventional oil and gas. However, it provides for the use of surface water for these activities and argues that 94% of the volume of the main rivers in the region flows into the sea without being used. Through a political ecology perspective and using critical discourse analysis, the chapter intends to make sense of how the state shapes socio-natural relations and seeks to manufacture consent for hydraulic fracturing. The analysis shows that government and industry actors develop an environmental discourse centered on two main arguments: the need to exploit natural resources in order to reinforce development and the possibility of exploiting natural resources with environmental protection. This discourse thus aims to legitimize the exploitation of unconventional hydrocarbons. In particular, government and industry attempt to secure access to water for unconventional oil and gas activities by unfolding a sense of excess availability of surface water, which tend to ignore alternative ideas and usages of water.

<p>As the number of underground activities increase, the need for better understanding of the geospatial properties become more and more essential for correct engineering designs and optimal decision making. However, gathering subsurface data is still an extremely costly and imprecise endeavour. Geological modelling has played a crucial role for years helping to understand and correlate the complex geometries encountered underground but single deterministic models fail to capture all possible configurations given the limited data. Probabilistic machine learning allows to integrate domain knowledge and observations of the physical world on a rigorous and consistent manner. Inferences to the probabilistic model implements an automatic learning-from-observations process.</p><p> </p><p>In this work, we show how by embedding state-of-the-art implicit interpolants into probabilistic frameworks, we can integrate the information of distinct data sets in one single common earth model. We will present results from a minimal working example to introduce Bayesian statistics, to full 3-D probabilistic inversions. All the models used for this demonstration are implemented in the open-source library GemPy ( www.gempy.org) allowing full reproducibility of the results.</p><p> </p>

Predicting undergraduate student evaluations of teaching using probabilistic machine learning: The importance of motivational climate

Efficient seismic fragility analysis method utilizing ground motion clustering and probabilistic machine learning

Correlation and difference between conventional and unconventional reservoirs and their unified genetic classification

Numerous comprehensive well-logging suites have been deployed in recent years to assess unconventional oil and gas reservoirs. These technologies produce extensive and rich data sets that, while powerful, can introduce significant technical challenges during integration into formation evaluation workflows. A key challenge arises when integrating advanced dielectric and spectroscopy logs, each of which yields independent outputs for water saturation and salinity through distinct measurement physics. Present workflows designed to reconcile these and other quantities do not rigorously and consistently incorporate known measurement sensitivities and uncertainties and can vary between log analysts. In unconventional reservoirs, they also rely on user inputs of kerogen maturity and hydrocarbon properties, which can be difficult to ascertain a priori. This underscores the need for a consistent, quantitative interpretation methodology that can precisely integrate advanced logs into a formation evaluation workflow. A novel multiphysics inversion approach has been recently introduced to optimize the various formation property sensitivities of two primary advanced logs: multifrequency dielectric dispersion and geochemical spectroscopy. One of the primary limitations of standalone dielectric interpretation is its insensitivity to salinity estimation at high formation water salinity ranges, which is complemented by chlorine and formation sigma measurements in the multiphysics inversion. Additionally, nuclear magnetic resonance (NMR) porosity, when incorporated into the inversion with the spectral total organic carbon (TOC) measurement, provides unique sensitivities for constraining kerogen and hydrocarbon properties. The multiphysics inversion method quantitatively accounts for these differences in physical sensitivities, as well as their measurement uncertainties, in the solutions for water volume, saturation, formation water salinity, and kerogen and hydrocarbon densities. Combined with advanced mineralogy and porosity evaluations from the logging suites, this methodology produces unified, optimized outputs for advanced well-log interpretation while maximizing the extracted value of the acquired advanced measurement suite. We present case studies as a first exhibition of the new multiphysics inversion technique, integrated into complete petrophysical workflows, for utilizing multifrequency dielectric, geochemical, and NMR logging tools in unconventional reservoir evaluation. We explore reservoirs in the Permian Basin of the United States, including logged intervals over the organic shale pore-fluid systems of the Wolfcamp Formation. We also include an unconventional oil reservoir evaluation in the Vaca Muerta shale of Argentina. Through these studies, we demonstrate how the standardized multiphysics inversion approach supersedes individualized advanced log reconciliation and minimizes uncertainties in their integration. We provide comparisons to standalone interpretations of each advanced log to illustrate the benefits of a simultaneous inversion with discussions of the limitations presented by each approach. Validation of formation water salinity and kerogen properties is performed against available published local data. Finally, we discuss the conditions under which these workflows can be logically extended to conventional reservoir evaluation and carbon capture and storage (CCS) applications, broadening their applicability well beyond unconventional reservoirs.

CO₂ foam fracturing fluids, commonly used in fracturing deep unconventional oil and gas reservoirs, currently suffer from poor stability, high friction resistance, and inadequate sand-carrying capacity. This study selected three base fluids: guanidine gel (traditional system), G509 surfactant composite system, and SCF novel polymer system. Testing was conducted using a high-temperature, high-pressure foam generator and a high-temperature, high-pressure rotational rheometer system.The stability, rheology, and sand-carrying properties of carbon dioxide (CO₂) foam fracturing fluids were investigated under three different base fluid types. Using a high-temperature, high-pressure visual reactor, the foam stability in both liquid and supercritical states was evaluated via the half-life method. Rheological parameters were measured using a rotational rheometer, and the constitutive equation was fitted. Combined with clay aggregate settling experiments, the sand-carrying characteristics under actual operating conditions were analyzed.Experimental results indicate that under liquid CO₂ conditions, the SCF system exhibits optimal stability with virtually no liquid phase separation within 300 minutes. Under supercritical CO₂ conditions, both the guanidine-gum system and SCF system exhibited half-lives of 293 min, significantly outperforming the G509 system (14 min). Regarding rheological properties, all three systems demonstrated pseudoplastic behavior, with the SCF system maintaining superior viscosity retention under high shear conditions. In terms of sand-carrying capacity, the SCF system exhibited the longest aggregate settling time (>80 min) and the best sand-carrying performance.To provide reference for selecting fracturing fluid systems for deep unconventional oil and gas reservoirs.