Geological Information Research Articles

Integrated coupled assessment of geostorage and geothermal prospects in the oil fields of Upper Assam Basin

This study proposes an integrated approach of assessing CO2 storage potential and geothermal energy prospect based on the data of seventeen depleted wells of Upper Assam Basin which could assist the global objective of net zero transition. The petrophysical properties of Tipam, Barail and Lakadong + Therria Formations from the seventeen wells have been utilised to perform the Monte Carlo simulation for probabilistic estimation of the CO2 storage in the Upper Assam Basin. This preliminary work showed that the mean storage capacity of 18.8 ± 0.7 MT, 19.8 ± 0.9 MT and 4.5 ± 0.8 MT could potentially be stored in the three geological formations of the basin. The corrected bottom hole temperature values for the studied seventeen wells were determined using the well log data and Waples and Harrison method; these values provided a static geothermal gradient for each well, which varies widely from 0.017 to 0.033 °C/m. In order to enable geothermal prospectivity, static formation temperature maps have been generated for the studied wells. The probabilistic assessment of stored heat-in-place and formation temperature maps delimited five prospective sites for the extraction of geothermal energy in the basin. The study also presented a risk assessment for CO2 storage development in the basin. Further, the study illustrated an economic analysis of the implementation of a CO2 storage project and geothermal operations in the basin.

Definition of a thermal conductivity map for geothermal purposes

The use of geothermal energy is spreading globally due to its many advantages, especially for heating and cooling. The correct design of a geothermal system requires knowledge of the parameters of the subsoil rocks, and particularly the thermal conductivity (k), which is the intrinsic ability of a material to transfer thermal energy as a result of a temperature gradient. A thermal conductivity map of the geological formations is time-consuming to produce, but can be of great help when selecting the location of a low-enthalpy geothermal installation, resulting in significant savings and an increase in the efficiency of that installation. The preferred option for determining k is an in situ thermal response test, but laboratory methods may be an alternative if it is not available or affordable. In this work, the needle thermal probe method has been used to measure the k of representative outcropping rocks in Oviedo (NW Spain), since it allows to obtain a rapid determination, its cost is comparatively low and it can be implemented in a portable device. 162 measurements have been carried out on a total of 27 samples, ranging from 0.2 (clay) to 5.4 W m−1 K−1 (quartzite). A relationship has been found between the k of the rocks and their characteristics, such as mineralogy, anisotropy or geological age and a thermal conductivity map was created.

Deep learning algorithm-enabled sediment characterization techniques to determination of water saturation for tight gas carbonate reservoirs in Bohai Bay Basin, China

Understanding water saturation levels in tight gas carbonate reservoirs is vital for optimizing hydrocarbon production and mitigating challenges such as reduced permeability due to water saturation (Sw) and pore throat blockages, given its critical role in managing capillary pressure in water drive mechanisms reservoirs. Traditional sediment characterization methods such as core analysis, are often costly, invasive, and lack comprehensive spatial information. In recent years, several classical machine learning models have been developed to address these shortcomings. Traditional machine learning methods utilized in reservoir characterization encounter various challenges, including the ability to capture intricate relationships, potential overfitting, and handling extensive, multi-dimensional datasets. Moreover, these methods often face difficulties in dealing with temporal dependencies and subtle patterns within geological formations, particularly evident in heterogeneous carbonate reservoirs. Consequently, despite technological advancements, enhancing the reliability, interpretability, and applicability of predictive models remains imperative for effectively characterizing tight gas carbonate reservoirs. This study employs a novel data-driven strategy to prediction of water saturation in tight gas reservoir powered by three recurrent neural network type deep/shallow learning algorithms—Gated Recurrent Unit (GRU), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), Support Vector Machine (SVM), K-nearest neighbor (KNN) and Decision tree (DT)—customized to accurately forecast sequential sedimentary structure data. These models, optimized using Adam's optimizer algorithm, demonstrated impressive performance in predicting water saturation levels using conventional petrophysical data. Particularly, the GRU model stood out, achieving remarkable accuracy (an R-squared value of 0.9973) with minimal errors (RMSE of 0.0198) compared to LSTM, RNN, SVM, KNN and, DT algorithms, thus showcasing its proficiency in processing extensive datasets and effectively identifying patterns. By achieving unprecedented accuracy levels, this study not only enhances the understanding of sediment properties and fluid saturation dynamics but also offers practical implications for reservoir management and hydrocarbon exploration in complex geological settings. These insights pave the way for more reliable and efficient decision-making processes, thereby advancing the forefront of reservoir engineering and petroleum geoscience.

A numerical model for the quantification of fluid inclusion property variations caused by heterogeneous entrapment and post-entrapment modifications in the H2O-NaCl system

The primary geological conditions for hydrothermal systems can be reconstructed using fluid inclusions, which generally requires that fluid inclusions trapped a single homogeneous phase and experienced no mass and volume changes after entrapment (Roedder's Rules). However, heterogeneous entrapment and post-entrapment modifications have been frequently identified in fluid inclusion studies, making the recovery of primary fluid properties from fluid inclusion data challenging or impossible. In this study, we constructed a specific tool, FlincPro, for the calculations of ideal and nonideal Fluid inclusion Properties in the H2O-NaCl system (ideal here refers to fluid inclusions that adhere to Roedder's Rules). Multiple modules have been developed to quantify the variations of fluid inclusion properties caused by heterogeneous entrapment, volume contraction/expansion, and water loss. The developed software illustrates possibilities in the variation of fluid inclusions in a specific assemblage with the input of a certain starting point.We found that compositions of the vapor endmember fluid are more significantly affected by heterogeneous entrapment than the liquid endmember, as the latter may have several orders of magnitude higher salinity and density. Homogenization pressure and temperature of halite-bearing fluid inclusions trapped on the halite liquidus will increase sharply with the addition of halite, but the homogenization behavior will not be changed. Volume contraction/expansion are effective ways to change the homogenization behaviors, especially for high salinity fluid inclusions. Water loss is not likely to cause the homogenization behavior change from total homogenization by vapor disappearance to total homogenization by halite dissolution for a halite-bearing fluid inclusion without additional volume contraction. Low-salinity and low-density fluid inclusions with total homogenization temperature around the critical temperature of water are least affected by post-entrapment modifications. Therefore, the interpretation of fluid inclusions formation depends on more parameters, e.g., petrographic observations, density and composition variations in fluid inclusion assemblages, and geological information, than only microthermometric data.

Identification of Geochemical Anomalies Based on RPCA and Multifractal Theory: A Case Study of the Sidaowanzi Area, Chifeng, Inner Mongolia.

The focus of exploration geochemistry is an accurate interpretation of geochemical data and the precise extraction of anomaly information related to mineralization from complex geological information. However, geochemical data are component data and exhibit a closure effect. Thus, traditional statistical methods cannot adequately reveal and identify the distribution of deep-seated anomaly information. This paper focuses on the Sidaowanzi area in Inner Mongolia and uses multivariate component data analysis methods to process 1:50 000 soil geochemical data. Using the Exploratory Data Analysis (EDA) method, the spatial distribution and internal structure characteristics of raw, logarithmic, and isometric logarithmic ratio (ILR) transformed data were compared and, coupled with robust principal component analysis (RPCA) and elemental component biplots, the association between element combinations and mineralization indicated by these three types of data was revealed. The S-A method was used to decompose composite anomalies of the ILR transformed RPCA score data to extract the characteristics of elemental combination anomalies and background distribution, and the Fry analysis method was utilized to analyze the dominant mineralization direction within the area. The results show that (1) data transformed using the ILR eliminated the influence of the closure effect, making the data more uniform on a spatial scale and exhibiting characteristics of an approximately normal distribution. (2) The S-A method was further used to decompose the composite anomaly of the PC1 and PC2 principal component combinations. The screened-out anomaly and background fields can essentially reflect the ore-causing anomalies dominated by Au and Cu-Mo mineralization. Moreover, the extracted anomalies and background information closely align with known mineral deposits (prospects) and can effectively identify weakly retarded geochemical anomaly information. (3) Fry analysis based on geochemical anomalies indicates that the dominant mineralization directions, by an assemblage dominated by Au and Cu-Mo, predominantly occur in the NE, NW, and proximate EW orientations. The combined application of the aforementioned three methods for the quantitative analysis of geochemical data aims to explore a transferable methodological system, providing new insights and approaches for further prediction of mineralization potential.

Numerical Simulation of Hydraulic Fracture Propagation on Multilayered Formation Using Limited Entry Fracturing Technique

Hydraulic fracturing is one of the most effective stimulation methods for unconsolidated sandstone reservoirs. However, the design of hydraulic fracturing must take into account the mechanical and stress properties of different geological formations between layers. In this paper, a three-dimensional coupled fluid-solid model using the finite element method is developed to investigate multiple vertical fractures at different depths along a vertical wellbore under different geological and geomechanical conditions. The finite element model does not require further refinement of any new cracks, requiring much smaller degrees of freedom and higher computational efficiency. In addition, new elements were used to account for local pressure drop due to perforation entry friction along the vertical wellbore. Numerical simulation results indicate that hydraulic fracture connections are observed from adjacent layers. Furthermore, the low stress contrast and high Young’s modulus between the layers increases the likelihood of multiple fracture connections. Higher fluid leakage rates increase the likelihood of fracture branching, but decrease the area of fracture coverage near the wellbore. Increasing fluid viscosity is effective in improving the area of fracture coverage near the wellbore. These findings are useful for the design of hydraulic fracturing in multi-layered formations in unconsolidated sandstone formations.

Utilizing Random Forest algorithm for identifying mafic and ultramafic rocks in the Gameleira Suite, Archean-Paleoproterozoic basement of the Brasília Belt, Brazil

Geological mapping techniques have continuously advanced, particularly with the increasing use of remote data acquisition methods, such as aerogeophysical data analysis and the use of machine learning techniques, both has become essential for mapping and understanding geological environments. Mafic-ultramafic bodies, which may be linked to metallic ore deposits, can be mapped using these innovative techniques. In Brazil, the Tocantins Province is reknown for its primary Cu–Co–Ni–Cr ore deposits, which are found in mafic-ultramafic rocks. The Tocantins Province contains several mafic-ultramafic bodies that have received less attention than the well-known deposits in the Brasília Belt. The Gameleira Suite is a geological unit located in the basement of the North Brasília Belt. Our study aims to use machine learning to identify new potential mafic-ultramafic occurrences in the Gameleira Suite. To achieve this, aerogeophysical surveys data, including magnetometry and radiometry, were used in conjunction with remote sensing. The data collected from aerogeophysics surveys were compiled into a database. The Random Forest algorithm was used to analyze this database, with 1.96% (from a total of 84.535 samples) used for training and generating the predictive map. Three approaches were used to verify and evaluate the data: analyzing magnetic lineaments to assess the influence of structural tectonic factors on body positioning, collecting field samples, and utilizing Magnetic Vector Inversion (MVI) to analyze the magnetic properties of the projected bodies at depth. This was done because the Gameleira Suite bodies display magnetic remanence. Our findings suggest that the utilization of these methods, in conjunction with the verification techniques employed, aids in the mapping and identification of mafic-ultramafic rocks. Machine learning can improve cartography by identifying new occurrences or indicating areas that need more detailed mapping. The use of additional geological knowledge and information not included in the model is crucial because the predictive map does not inherently represent geological truth. It is necessary to interpret the results from a geological perspective.

Prior-guided adaptive generative adversarial network method for various types of borehole image inpainting

During geophysical exploration, inpainting defective logging images caused by mismatches between logging tools and borehole sizes can affect fracture and hole identification, petrographic analysis, and stratigraphic studies. However, existing methods do not describe sufficient stratigraphic continuity. In addition, they ignore the completeness of characterization in terms of fractures, gravel structures, and fine-grained textures in the logging images. To address these issues, we develop a deep-learning method for inpainting stratigraphic features. First, to enhance the continuity of image inpainting, we build a generative adversarial network and train it on numerous natural images to extract relevant features that guide the recovery of continuity characteristics. Second, to ensure that complete structural and textural features are found in geologic formations, we introduce a feature-extraction-fusion module with a cooccurrence mechanism consisting of channel attention (CA) and self-attention (SA). CA improves texture effects by adaptively adjusting control parameters based on highly correlated prior features from electrical logging images. SA captures long-range contextual associations across preinpainted gaps to improve completeness in fractures and gravel structure representation. Our method has been tested on various borehole images demonstrating its reliability and robustness.

Geological formation conditions for lead and zinc ore deposits in the People’s Democratic Republic of Algeria

Geological formation conditions for lead and zinc ore deposits in the People’s Democratic Republic of Algeria

The historical Corrente gold deposit, Minas Gerais, Brazil: Geological characteristics and a reconnaissance study of nugget effect

Corrente is a historical underground working in Minas Gerais, Brazil, mined for gold in the nineteenth century, for which little geological information is available. Retrieval of unpublished exploration reports from the 1980s has enabled the recognition of a 7-km-long linear cluster of historical underground and open-cast workings, here referred to as the Cata Branca–Corrente goldfield. Gold production ceased at Cata Branca, the most economically relevant deposit of the goldfield, in 1844. Both Corrente and Cata Branca are auriferous quartz-lode deposits. Geological characteristics of the Corrente deposit are presented, as well as reconnaissance whole-rock analyses for Au using a variety of analytical methods, including conventional fire assay, atomic absorption spectrometry, instrumental neutron activation analysis and laser ablation–inductively coupled plasma–mass spectrometry (LA–ICP–MS) on pressed-powder pellets (PPPs). Corrente is characterised as a gold-only deposit. The Corrente auriferous quartz lode, hosted in the Archaean greenstone belt sequence of the Rio das Velhas Supergroup, is extensively oxidised and has conspicuous wall-rock alteration to fine-grained muscovite. The resulting sericite phyllite shows overprint by low-temperature kaolinitisation, likely the product of wall-rock hydration mediated by acidic water derived from sulfide oxidation at depth, below the modern weathering zone. The Corrente gold deposit has a high nugget effect. Time-resolved results of LA–ICP–MS on PPPs indicate the presence of particulate gold even in barren kaolinitised wall-rock phyllite.

Insights into the Impact of Physicochemical and Microbiological Parameters on the Safety Performance of Deep Geological Repositories.

Currently, the production of radioactive waste from nuclear industries is increasing, leading to the development of reliable containment strategies. The deep geological repository (DGR) concept has emerged as a suitable storage solution, involving the underground emplacement of nuclear waste within stable geological formations. Bentonite clay, known for its exceptional properties, serves as a critical artificial barrier in the DGR system. Recent studies have suggested the stability of bentonite within DGR relevant conditions, indicating its potential to enhance the long-term safety performance of the repository. On the other hand, due to its high resistance to corrosion, copper is one of the most studied reference materials for canisters. This review provides a comprehensive perspective on the influence of nuclear waste conditions on the characteristics and properties of DGR engineered barriers. This paper outlines how evolving physico-chemical parameters (e.g., temperature, radiation) in a nuclear repository may impact these barriers over the lifespan of a repository and emphasizes the significance of understanding the impact of microbial processes, especially in the event of radionuclide leakage (e.g., U, Se) or canister corrosion. Therefore, this review aims to address the long-term safety of future DGRs, which is critical given the complexity of such future systems.

Geomythological sites as potential geotourism destinations on the example of Lublin and Roztocze regions, SE Poland

Geomythology is an important element of the region's immaterial culture and the cultural identity of its inhabitants. Thanks to geomythology, tourists' interest in geotourist sites is increasing and their perception of these sites is changing. Geotourism is one of the forms of specialized tourism with a relatively narrow audience. To attract visitors from other groups, e.g., schoolchildren, families with children, tourists interested in the cultural heritage of a region, information on geological (or more broadly, geographic) resources can be expanded to include cultural information. One way to attract tourists and get a wider audience interested in geology is to provide geological information along with legends about the places visited. At the same time, geomythology increases the cultural identity of the region's residents.

Synergistic fusion network for landslide segmentation: the vital complementarity of geological information to remote sensing imagery

In the contemporary technological environment, access to a multitude of data sources, such as digital elevation models (DEM), slope data, and other geological information, has become essential for augmenting remote sensing imagery in landslide analysis. These datasets enhance landslide identification capabilities, but existing methods often fail to fully leverage their unique characteristics due to oversimplified fusion strategies. To overcome this challenge, we introduce the Synergistic Fusion Network (SFNet), which is specifically designed to utilize the complementary nature of geological data to significantly improve the accuracy and efficiency of landslide segmentation. SFNet incorporates Multi-Source Fusion Modules (MSFM) and Reverse Feature Refinement Modules(RFRM), through an innovative network architecture, enhancing remote sensing imagery with integrated geological information. In our experiments, SFNet outperforms existing mainstream networks, demonstrating superior F1 scores, thereby showcasing its ability to effectively segment landslide regions with high precision. This advancement highlights the importance of considering the unique contributions of each data source, especially the complementary effects of geological information.

Impact of formation conditions and well configurations on underground hydrogen storage

Converting surplus renewable energy into hydrogen is considered an efficient technique to balance renewable energy generation and demand, which necessitates large-scale underground hydrogen storage (UHS) in geological formations. While the H2 trapping mechanisms and fluid flow behavior have been rigorously reported in past studies, the performance of UHS is strongly influenced by the formation conditions and well configurations, which have not been explored thoroughly. To this end, this study investigated the sensitivity of UHS efficiency to variations in formation temperature, brine salinity, well placement, and well perforations. For each scenario, key aspects such as plume migration dynamics, residual trapping, dissolution trapping, structural trapping, H2 recovery factor, and water production were thoroughly analyzed. The results indicate that factors such as formation temperature, brine salinity, and well perforations have a negligible impact on the shape of the overall H2 plume. However, these significantly influence local H2 saturations. Formations with higher temperatures, central dome well placement, and production wells in the top layers exhibit a higher H2 recovery factor (as high as 69.93 %), whereas salinity plays a minor role in H2 trapping and withdrawal (maximum change observed is ∼1.7 % and ∼1.4 % for trapping and withdrawal respectively with a 20 wt % increase in salinity). In addition, brine production is strongly influenced by the formation conditions and well configurations (e.g., ∼3965 m3 brine produced under at 343 K and ∼26,000 m3 for non-central well placement), due to variations in H2/brine properties, such as density and viscosity, and the extent of H2 accumulation.

Numerical simulation study on enhanced efficiency of carbon dioxide geological storage with nanoparticles in deep saline aquifer

Geological storage of carbon dioxide (CO2) is a promising technique to reduce large-scale greenhouse gas emissions. There are some candidates for geological storage formation while deep saline aquifers have attracted tremendous attention because they have the largest potential of all possible subsurface storage sites. CO2 injected into deep saline aquifers migrates upward until it reaches a sealing formation due to its buoyancy effect, which causes poor storage efficiency. Over the past few years, some novel research has been published to utilize nanoparticles for CO2 geological storage. The addition of nanoparticles to injecting super-critical CO2 (scCO2) changes its physical properties, such as the density and viscosity, which could mitigate gravity segregation and viscous fingering. However, the studies related to these techniques are still extremely limited, and the potential advantages or their impacts are poorly understood. In this paper, we investigated the effects of the physical properties of scCO2 containing nanoparticles on storage efficiency by using a versatile commercial simulator (CMG-GEM). In particular, the effects of nanoparticle types (aluminum oxide, silicon dioxide, titanium oxide, and zinc oxide) and concentration on CO2 storage efficiency were presented quantitatively by considering possible operation scenarios. The results showed the migration of CO2 to the upper part of the aquifer was suppressed, and the storage of CO2 in the lower part of the aquifer was promoted when scCO2 was injected with nanoparticles. It also showed some effects in delaying breakthrough time and mitigating breakthrough amount when we assumed the pressure control well. It was indicated that the cumulative breakthrough amount after 30 years was reduced by 8.2% in the minimum case and by 79.3% in the maximum case. Although the proposed method still has economic challenges, the innovation in the production processes of nanoparticles or combining with added value will lead to improving feasibility.

Investigating the response of Urban Underground Utilities (3U) within an elastic and elastoplastic geological formation: Employing numerical and analytical techniques

Investigating the response of Urban Underground Utilities (3U) within an elastic and elastoplastic geological formation: Employing numerical and analytical techniques

Deep generative networks for multivariate fullstack seismic data inversion using inverse autoregressive flows

The simultaneous prediction of the subsurface distribution of facies and acoustic impedance (IP) from fullstack seismic data requires solving an inverse problem and is fundamental in natural resources exploration, carbon capture and storage, and environmental risk management. In recent years, deep generative models (DGM), such as variational autoencoders (VAE) and generative adversarial networks (GAN), were proposed to reproduce complex facies patterns honoring prior geological information. Variational Bayesian inference using inverse autoregressive flows (IAF) can be performed to infer the solution to a geophysical inverse problem from the encoded latent space of such pre-trained DGM. Successful applications of such approach on crosshole ground-penetrating radar synthetic data inversion demonstrated that the technique's accuracy is comparable to that of Markov chain Monte Carlo (MCMC) inference methods, while significantly reducing the computational cost. Nonetheless, these application examples did not account for the spatial uncertainty affecting the facies-dependent continuous physical property, from which the geophysical data are calculated. This uncertainty can significantly affect the inversion accuracy and its applicability to real data. In this work, specific VAE and GAN architectures are proposed to simultaneously predict facies and co-located IP, while accounting for their spatial uncertainties. The two types of generative networks are used in Bayesian inversion with IAF for the inversion of seismic data. The results are found to reproduce the statistics of the training images and solve the seismic inversion problem accurately, comparably to MCMC inversion. Furthermore, advantages and limitations of the two DGMs are evaluated by comparing the results obtained.

Predicting ground surface deformation induced from CO2 plume movement using machine learning

Carbon capture and storage (CCS), which involves injecting carbon dioxide (CO2) into subsurface, is an increasingly popular process for mitigating human caused greenhouse gas emissions. In order to ensure the safety and efficacy of CCS implementation, it is necessary to possess a comprehensive understanding of the complex behaviour of CO2 plumes within geological formations and their potential impact on ground surface deformation. Therefore, conducting research and analysis on these critical aspects is of vital importance. This research provides a methodology to anticipate ground surface deformations, which result from the motion of CO2 plumes utilising an advanced machine learning (ML) technique. The ML surrogate model has been developed using conditional Generative Adversarial Networks (cGAN). The dataset used for the model training and testing comprises ground surface measurements (tiltmeters), reservoir properties, as well as pressure/volume data. The model has been trained and tested using a set of samples created using a forward finite element model. Results show that the surrogate model is capable of predicting reasonably accurate results while running much faster than the forward model.

Characteristics and Sources of CBM Well-Produced Water in the Shouyang Block, China

The Shouyang Block was selected as the research subject. Comprehensive analysis was conducted using coalbed methane (CBM) well production data, geochemical test data on water produced from the coalbed methane well, and fundamental geological information. The findings reveal the water dynamics in the Shouyang Block are characterized by weak groundwater runoff or retention in most areas. The groundwater head height exhibits a gradual decrease from the north to south, which is closely associated with the monoclinic structure of the Shouyang Block. Overall, water production is relatively high. As the average water production increases, the average gas production gradually decreases. A concentration of high water production wells is observed in the northern part of the Shouyang Block, which gradually increases towards the southeast direction. A comprehensive analysis was conducted on the factors influencing water production, including total water content of coal seams, coal seam porosity, groundwater stability index, groundwater sealing coefficient, D value of the fracture fractal dimension, fault fractal dimension, and sand–mud ratio. The correlation degree was calculated and ranked in order of magnitude through grey correlation analysis. The order of factors that influence water production, from strongest to weakest, is as follows: sand–mud ratio > porosity > fractal dimension of fault > fracture fractal dimension D value > groundwater sealing coefficient > groundwater stability index > total water content of coal seams. The dissolution amounts of carbonate and sulfate are both small, and the water source may mainly come from the sandstone aquifer. Attention should be paid to the distribution and lithological combination of sandstone aquifers in coal-bearing strata in the future exploration and development process of the Shouyang Block. This will help to avoid the potential influence of fault structures and enable the identification of favorable areas for low water and high gas production.

Graph neural network-based topological relationships automatic identification of geological boundaries

Topological relationship identification of geological boundaries greatly affects 3D geological modeling efficiency. However, due to the complex spatial distribution and unstructured data architecture of geological boundaries, the current topological identification still mainly depends on the manual operation and interpretation, which is time-consuming and labor-intensive. And geologists' judgments are based on knowledge and experience that are challenging to replicate through conventional algorithmic approaches. To address these challenges, a graph neural network-based geological boundary topological relationships automatic identification method is proposed. Based on graph theory, we conceptualize geological boundaries and their topological relationships as a graph transforming the identification task into an edge prediction problem within this topological graph. We employ the variational graph auto-encoder (VGAE) model to predict unknown edges in the topological graph, enabling automatic identification of geological boundary topological relationships. A geological information transformation method is proposed to extract and encode the geological boundaries’ features as inputs for VGAE. The proposed method is verified in an engineering case study. The results reveal that this method can efficiently and accurately analyze the geological boundary topological relationships. Moreover, it can greatly reduce the manual workload, requiring only 30% of the manual effort to obtain the geological boundary topological relationships, while achieving an average prediction accuracy of approximately 95%.