The traditional temperature-pressure field analysis method relies heavily on the interpolation of existing borehole data, failing to accurately characterize the coupled seepage-heat transfer processes in geothermal systems. This limitation hinders a comprehensive understanding of genetic mechanisms of geothermal resources.

To address these issues, we constructed a high-precision three-dimensional geological model of the Zhangye Basin by integrating multi-source datasets (borehole information, geophysical data, and digital elevation models). Our integrated approach improved the stratigraphic accuracy between boreholes by 50–300 m compared to conventional interpolation models.Numerical simulations of coupled seepage-heat transfer processes based on the 3D geological model indicated that, compared with the traditional key-node spatial interpolation method, the multi-field coupling analysis more reasonably revealed the temperature and pressure characteristics of the reservoir in the study area. The results showed that in the study area, the geothermal hydraulic head is relatively high in the southeastern part of the basin center and gradually decreases toward the northwest, indicating an overall southeast-to-northwest seepage pattern. Groundwater was recharged into the reservoir through faults and became fully heated by the geothermal field during seepage. As the reservoir depth decreased and the caprock thinned, heat was gradually lost, resulting in a temperature field characterized by higher values in the middle and lower values around the margins, with a central temperature reaching up to 78 °C. A three-dimensional geothermal conceptual model was subsequently developed, which comprehensively interpreted the genetic mechanism of geothermal resources in the basin by integrating structural, hydrogeological, and geothermal geological conditions. Compared with previous two-dimensional models, the three-dimensional conceptual model and the coupled heat–fluid analysis method developed in this study more accurately describe the spatial distribution characteristics of geothermal resources and provide a more reasonable explanation for their genetic mechanisms.

The results reveal the southeast-to-northwest groundwater flow within the reservoir and the genetic mechanism underlying the rhombic-lobate distribution of geothermal resources in the basin. These findings provide a theoretical framework for targeting high-enthalpy geothermal reservoirs and optimizing sustainable exploitation strategies.

The Fuman Oilfield has significant potential for oil and gas exploration, and is a key area for increased oil reserves and production in the Tarim Oilfield. Analyzing the source, thermal maturity, and charging process of crude oils in this area is essential for understanding the hydrocarbon accumulation patterns, enrichment laws, and guiding future exploration activities.

Based on systematic organic geochemical analysis of 15 crude oil samples and fluid inclusion samples from three wells, as well as the orientation segmentation and active phases of the strike-slip fault, this study examines the geochemical characteristics of crude oils, hydrocarbon charging processes, and their main controlling factors in the F_I17 strike-slip fault zone.

The results show that the crude oil in this fault zone is derived from the Lower Cambrian Yurtus source rocks. The equivalent vitrinite reflectance value of the crude oil, calculated from aromatic parameters, ranges from 0.80% to 1.20%. The crude oil has undergone relatively weak thermal chemical sulfate reduction (TSR), gas stripping, and oil cracking. The fault zone and its periphery have experienced three stages of oil charging with different maturities: Late Caledonian, Late Hercynian, and Himalayan periods, and the gas charging occurred during the Himalayan period. However, there are differences in oil maturity within the same period between the northern and southern segments. The northern and middle segments are mainly dominated by oil contributions from the second stage (blue-green fluorescence) and third stage (bright-blue fluorescence), however, the southern segment is predominantly charged by the second-stage oil with bright-blue fluorescence and the Himalayan-period gas invasion altering pro-existing oil. The F_I17 strike-slip fault zone exhibits a north-to-south increasing gradient in the oil thermal maturity, gas stripping intensity, oil cracking degree and contribution of late-stage hydrocarbons.

This spatial variation is governed by fault activity intensity that intensifies southward, which is stronger in the south and weaker in the north. This study suggests the likely presence of deeper, late-stage high maturity hydrocarbons in the northern and middle sections of the fault zone, and the current exploration depth has not yet reached the lower limit of liquid hydrocarbon occurrence. Therefore, deeper reservoirs still have significant potential for oil and gas exploration.

In the Junggar Basin, the exploration targets are deeply buried beneath the earth surface, posing significant challenges for seismic exploration. Specifically, this result in severe high-frequency absorption and attenuation of seismic wave energy. The recorded seismic data are characterized by a limited bandwidth with low dominant frequency, which significantly compromises the accuracy of the sandstone-shale thin interbed identification. Thus, the poor quality of the seismic data makes the reservoir distribution and potential hydrocarbon prediction tasks extremely challenging. Deconvolution is a key technique to improve the seismic data resolution. It aims to reverse the effects of wavelet convolution in the recorded seismic data, and can be performed in either the time domain or the frequency domain. Frequency domain deconvolution usually uses the estimation of seismic spectrum to construct a spectral-broadening operator, thereby expanding the frequency band and increasing the dominant frequency of the seismic data. We develop a frequency-domain deconvolution method to enhance the vertical seismic resolution. The key to this method lies in the use of an optimized target spectrum, and the final goal is to expand the seismic bandwidth effectively.

We propose a generalized cosine broadband spectrum and incorporate it into the spectral inversion. Based on the relationship between the expected spectrum and diagonal matrix of the seismic spectrum, a spectral-broadening forward model is established. Subsequently, a shaping regularization inversion method is used to invert for the spectral-broadening operator. Ultimately, by applying the obtained spectral-broadening operator to the seismic data, the vertical seismic resolution can be enhanced. This is beneficial for more accurate geological interpretation and hydrocarbon exploration.

We use a theoretical model and field seismic data in the Junggar Basin to validate of the proposed method. The theoretical model has verified the effectiveness of the proposed method using the generalized cosine broadband spectrum. Subsequently, we apply the method to real seismic data from the Junggar Basin, where an optimized generalized cosine broadband spectrum is constructed based on the original seismic spectrum. The real seismic data processing results show that the frequency bandwidth of the seismic data has been effectively increased, leading to improved resolution of thin interbeds. The theoretical and field data results confirm the robustness and practical applicability of the proposed method.

The generalized cosine broadband spectrum has a wide frequency band range, and can be flexibly designed. Additionally, the corresponding wavelet exhibits low side-lobe amplitude and narrow side-lobe width, which minimizes interference from reflected waves. The proposed spectral inversion using the generalized cosine broadband spectrum can be used to improve the seismic resolution and provides reliable seismic data for unconventional hydrocarbon exploration.

Petroleum breakthroughs in the strike-slip fault zones of the northeastern Shunbei area in the last decades reveled the spatial variation in petroleum accumulation types and phase patterns, including the oil accumulation on the north, condensate gas accumulation on the south, and dry gas accumulation on the east. Therefore, to further advance ultra-deep petroleum exploration, it is crucial to reveal the genesis mechanism, source, and thermal maturation of the natural gas from the perspective of overall petroleum distribution.

This study systematically collected gas samples from various fault zones in the Shunbei area in order to analyze the geochemical characteristics of natural gas and reveal its generation mechanism or thermal evolution process.

The results indicate that the natural gas in the Shunbei area was minimally affected by TSR even though some parts of the fault zones showed strong modification of thermal-chemical sulfate reduction (TSR). The natural gases in the No. 1 fault zone and in the northern and middle sections of the No. 5 fault zone are primarily crude oil-associated gases generated by primary kerogen cracking. In contrast, the natural gases in the southern section of the No. 5 and No. 4 fault zones were mainly the mixture of early kerogen cracking gas (oil-associated gas) and late crude oil cracking gas. The natural gas in the Shunbei No. 12 fault zone originated from deeper, high-temperature crude oil cracking gas, which occurred at the wet gas stage. The natural gases in the study area predominantly came from the source rocks of the Lower Cambrian Yurtus Formation, and the parent material of these gases is characterized by benthic algae or a mixture of benthic and planktonic algae. Finally, a regression equation for thermal maturity calculation based on the carbon isotope of methane was established for the hydrocarbon generation process in the source rocks of the Yurtus Formation.

The research findings provide important insights into the origin, source, and thermal maturity of ultra-deep gas, offering valuable references for future studies on hydrocarbon generation in the Shunbei area.

Current research on the Jurassic Kizilnur Formation coal seams in the Kuqa Depression mainly focuses on shallow sections, with limited study of deep coal seams.

To address unknown coal-rock characteristics, reservoir properties, and gas content in the Jurassic Kizilnur Formation coal seams, the geometric distribution characteristics, the characteristics of coal rock and coal quality, reservoir properties, adsorption properties, and gas content of the coal seam in this block were studied based on the following methods, such as drilling data, seismic data processing, wellbore analysis, sampling observation, gas measurement, pore permeability, and isothermal adsorption experiments.

The deep coal seams in the research area are predominantly composed of primary structures, with vitrinite as the major organic component (average 62.2%), indicating type Ⅲ organic matter. The coal is of medium rank, generally characterized by medium-to-high volatile matter, ultra-low sulfur, and low to very low ash content. The average gas porosity is 7.10%, and the average permeability is 0.685×10⁻³ μm². Furthermore, methane adsorption capacity is strongly influenced by macrinite reflectance, water content, and ash yield. In the Kuqa Depression, deep coal seams occur at burial depths of 2 000 –5 000 m, are mainly distributed in the Northern Structural Zone, cover about 2 800 km², with a maximum single-layer thickness of 22 m. The average gas content of the deep coal seam measured is 11.77 m³/t, indicating substantial resource potential.

Compared with shallow coal seams, the deep coal seams of the Kezilnur Formation in the study area are more strongly affected by burial depth, exhibit high maturation, and have achieved large-scale gas production, underscoring their exploration potential. Water saturation emerges as a critical factor controlling gas content in these seams. Notably, the Dibei gentle-slope platform in the Kuqa Depression offers particular advantages for CBM extraction due to favorable hydrological and geological attributes.

The Qikou Sag, a significant hydrocarbon-bearing sub-basin within the Bohai Bay Basin of eastern China, has recently witnessed a critical exploration breakthrough. Well HT1, drilled in the main trough, encountered a sandstone reservoir within the Third Member of the Dongying Formation (Ed₃) and achieved commercial oil flow. However, the precise provenance and spatial distribution of this Ed₃ sandstone remain enigmatic. Identifying the sediment source is fundamental to constructing a reliable "source-to-sink" system, which is crucial for predicting the distribution of high-quality reservoirs. This study aims to definitively determine the provenance of the Ed₃ sandstone through detrital zircon U-Pb geochronology and establish a robust sedimentary model to guide exploration.

To address the provenance question, this study employed laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) U-Pb dating on detrital zircons extracted from the Ed₃ sandstone. Zircon is an ideal mineral for provenance analysis due to its high physical and chemical stability, allowing it to survive multiple cycles of erosion and transport while preserving its primary crystallization age. Thus, the age spectrum of detrital zircons in a sedimentary rock serves as a unique fingerprint that can be compared to the age signatures of potential source areas, enabling a direct link between the sediment and its origin.

The U-Pb dating results from the Ed₃ sandstone revealed a characteristic age distribution pattern. The detrital zircon ages cluster into four primary intervals: 2600−2300 Ma (Neoarchean to Paleoproterozoic), 1680−2000 Ma (Paleoproterozoic), 300−240 Ma (Permian), and a remarkably dominant peak between 160−100 Ma (Jurassic to Cretaceous). The overwhelming abundance of zircons in the 160−100 Ma range immediately pointed towards a significant contribution from a source region with intense Mesozoic magmatic activity. The zircon age signature of the Cangxian uplift was examined. Its age spectrum is characterized by a strong peak at 300−240 Ma and a relative lack of Mesozoic zircons, presenting a clear mismatch with the Ed₃ sandstone signature and effectively ruling it out as the primary source. In contrast, the Yanshan orogenic belt to the north is a complex tectonic zone with a prolonged magmatic history spanning from the Archean to the Mesozoic. Critically, it is renowned for widespread Jurassic-Cretaceous (Yanshanian) magmatism, with a pronounced age peak precisely between 160 and 100 Ma. The striking similarity between the Ed₃ zircon age pattern and the magmatic record of the Yanshan belt provides compelling evidence that the latter served as the principal provenance for the sandstone. Previous studies indicate that the Yanshan orogenic belt experienced significant Cenozoic uplift, coeval with the deposition of the Ed₃, creating a well-defined mountain-basin system. Several major paleo-rivers, such as the ancient Yongding River, ancient Chaobai River, and ancient Luan River, are interpreted to have acted as efficient sediment conduits. These rivers eroded the Mesozoic granitic rocks of the Yanshan belt and transported the detritus southward over long distances into the Qikou Sag.

This study successfully resolves the provenance of the Ed₃ sandstone in the main trough of the Qikou Sag. Detrital zircon U-Pb geochronology unequivocally identifies the northern Yanshan orogenic belt as the primary source, excluding the nearby Cangxian uplift. The construction of the "remote source, channelized transport, deep-trough deposition" model has profound implications for exploration. It validates the existence of high-quality sandstone reservoirs in the deep troughs formed by long-distance transport systems. Consequently, future exploration efforts should prioritize mapping the paleo-drainage pathways that funneled sediments from the Yanshan belt into the main trough of the Qikou Sag. This model significantly reduces exploration risk in the deep plays of the depression and provides a valuable analog for searching for similar subtle reservoirs in other continental rift basins worldwide.

Tectonic fractures play a significant role in enhancing the physical properties and hydrocarbon productivity of deep volcanic buried-hill reservoir. Due to their strong heterogeneity, precise characterization of fracture density across different parts of the fault zone is required. In practical production, the acquisition cost of core and imaging logging data is relatively high. Therefore, a comprehensive fracture index (CFI) based on conventional logging curves is proposed to characterize the internal structure of fault zones and quantitatively evaluate fracture density, providing a basis for the exploration of fractured volcanic reservoir.

A CFI model was constructed for Mesozoic volcanic rocks in the Bonan area of the Bohai Bay Basin using seven logging parameters: acoustic travel time (AC), neutron porosity (CN), bulk density (DEN), caliper (CAL), deep lateral resistivity (RD), shallow lateral resistivity (RS), and micro-lateral resistivity (RMLL). The variation of the cumulative CFI curve slope was used to quantitatively define the boundaries of fracture-dense zones, partitioning the fault zone into the fault core, damage zone, and host rock. The damage zone exhibited the highest fracture density, followed by the host rock, while the fault core was nearly fracture-free.

In the Mesozoic volcanic rocks of the study area, the seven conventional logging curves showed distinct responses in fracture-developed zones: AC, CN, and CAL increased, whereas DEN, RD, RS, and RMLL decreased. The CFI curve showed a pronounced rise at fracture-rich intervals. Comparison with imaging logs and core data confirmed that high CFI values correlated with fracture zones. By integrating seismic interpretation to identify the fault core, the distribution of the damage zone and host rock could be delineated, enabling effective fracture prediction in Mesozoic volcanic rocks in the Bonan area.

The CFI model, derived from logging curves, can effectively quantify the intensity of fracture development in volcanic rocks, locate the downhole spatial distribution of fracture-dense zones, and reliably identify of the structure of a fault zone via cumulative CFI curve analysis.

A high-resolution, continuous chronostratigraphic framework was established using the astronomical cycles, and the paleoclimatic characteristic was analyzed for the Oligocene Lingshui Formation of the Beijiao Sag.

The cyclostratigraphy characteristics and the paleoclimatic evolution were determined by time series and multiple paleontological proxies analyses, respectively, in the Well A of the Beijiao Sag.

Cycles of 405 ka long eccentricity, 100 ka short eccentricity, 42.7 ka and 43.2 ka obliquity, and 20.5 ka precession were identified in the Lingshui Formation of the Beijiao Sag. Considering the time anchor of 23.03 at the top of Oligocene, the astronomical tuning revealed a "floating" astronomical chronology scale of 5.59 Ma. Within the high-precision absolute astronomical timeframe of the Lingshui Formation (28.62−23.03 Ma), the bottom boundary ages of Ling Ⅰ, Ling Ⅱ, and Ling Ⅲ members can be determined as 23.97 Ma, 25.43 Ma, and 28.62 Ma, respectively. These ages correspond to the average sedimentation rates of 18.4 cm/ka, 8.6 cm/ka, and 12.8 cm/ka, respectively. The palaeoclimatic indicators, including the foraminifera, organic carbon, organic detrital fractions, and carbon and oxygen isotopes, suggest an overall warm-cool-warm trend during the deposition of Lingshui Formation. The comparison of palaeoclimate evolution with astronomical rotation suggested that the climate evolution during this period was mainly controlled by astronomical orbital forces such as eccentricity and age difference. Astronomical cyclostratigraphic study on the Lingshui Formation of the Beijiao Sag explores the control of astronomical orbital parameters on paleoclimate and establishes a high-resolution and continuous chronostratigraphic framework.

This method provides a refined chronological framework for predicting high-quality reservoir intervals in oil and gas exploration, promoting the development of oil and gas exploration in the Beijiao Sag.

Currently, exploration and development of shale hydrocarbon primarily focus on mid-shallow marine and lacustrine shales. However, research on the sedimentary characteristics of deep lacustrine shales in rift basins is scarce, so their exploration potential is unclear. Consequently, this study holds significant practical implications and can provide valuable insights into the exploration and development of deep lacustrine shales in the rift basins. A typical deep shale series developed in the distal part of the fan front and the deep lake area of the Third Member of the Shahejie Formation in the Banqiao Sag, which is a continental rift basin. This study aims to identify the sedimentary characteristics of the deep lacustrine shale in the Third Member of the Shahejie Formation in the Banqiao Sag, to demarcate the vertical sweet spots of shale oil and gas, and to comprehensively evaluate the exploration potential of the deep lacustrine shale.

Based on data from core analysis, well logging, and various analytical tests, by employing techniques such as geological statistics, sedimentary facies and organic geochemical analysis, rock thin section observation, X-ray diffraction of whole-rock minerals, micro-CT scanning, and geophysical prediction, the macroscopic sedimentary features, rock fabric types, bedding structures, as well as microscopic structures and compositional characteristics of the shale were analyzed to explore its material basis for hydrocarbon generation and the conditions for hydrocarbon formation.

Macroscopically, the shale series of the Third Member of the Shahejie Formation can be classified into three types: mixed sedimentation type, intercalated type, and interbedded type. The mixed sedimentation type and intercalated type have a wide distribution range. The rocks mainly consist of black and dark gray mudstones, which are the dominant rock fabric types. Microscopically, extremely thin-laminated and thin-laminated mud shales exhibit relatively high organic carbon contents and are the dominant lamination types. The analysis of diagenetic evolution indicates that below a depth of 3 350 m, there are zones where interlayer water is rapidly expelled and zones where interlayer water is expelled rapidly along with organic matter. The resulting abnormal high pressure is conducive to the preservation of microscopic pores.

Comprehensive research shows that the main part of the shale series in the Third Member of the Shahejie Formation in the Banqiao Sag has reached the mature to highly mature stage. It possesses favorable conditions for the enrichment and accumulation of oil and gas on a large scale. Moreover, it is indicated that the C2 and C6 sweet spots are the main zones of oil and gas enrichment, providing theoretical support for the exploration and evaluation of shale oil and gas. The impact of tectonic activities on the generation and accumulation of shale oil and gas has not been adequately considered. Research on shale formations is still in its early stages, and numerous issues remain that require in-depth investigation. It is essential not only to elucidate the material basis for the occurrence of shale oil and gas but also to identify the key factors influencing their production. Therefore, future efforts should focus on achieving technical breakthroughs in the field of integrated geological and engineering approaches.

To identify the reservoir characteristics and influencing factors of Benxi Formation bauxite in the Linxing area,

the paper carried out X-ray diffraction (XRD), cast sheet imaging, scanning electron microscope-energy-dispersive spectrometer (SEM-EDS), mercury intrusion porosimetry, nitrogen and carbon dioxide adsorption, and routine porosity measurements, coupled with seismic logging data, to characterize the mineral composition, pore structure, and physical properties of the bauxite reservoir and to discuss the controls on its development.

The results show that the aluminum-bearing minerals in Benxi Formation bauxite in the Linxing area are dominated by diaspore. Pore types are primarily intra-granular, intergranular, matrix, and microcrack pores, with occasional organic pores. Pore volume is mainly contributed by mesopores and macropores, with pore-size peaks mainly at 30−70, 80−130, 4000−13000 nm. The overall reservoir physical properties are modest, with an average porosity of 3.28% and an average permeability of 1.398×10⁻³ μm². However, the lower part of the section, which has a higher diaspore content, exhibits relatively better properties. Finally, the development of the Linxing bauxite reservoir is controlled by palaeo-geomorphology, palaeo-sedimentary environment, and diagenesis. Specifically, accumulation and distribution of bauxite are controlled by the paleogeomorphology of depressions and troughs and by enclosed to semi-enclosed intermittent swamps and lagoons. Diagenesis determines the reservoir-space types and the resulting physical-property conditions within the sedimentary context.

The research results can provide theoretical guidance for the bauxite gas exploration.

Previous thermal structure analyses in the Songliao Basin have predominantly focused on the sedimentary scale investigations within the north-south zone. The lack of a comprehensive basin-wide thermal structure analysis at the lithospheric scale has hindered the genetic interpretation of its geodynamic setting.

Utilizing published datasets of surface heat flow, geothermal gradient, and thermophysical properties, this study enhances the existing framework by incorporating new thermophysical measurements from the Yaojia Formation, Qingshankou Formation and Quantou Formation. Supplementary geothermal filed datasets were integrated to establish a holistic characterization of the basin's geothermal regime. This multi-scale approach enables systematic analysis of the contemporary lithospheric thermal architecture.

The results reveal that the geothermal temperature gradient in the Songliao Basin ranges from 21.10 to 63.45°C/km, with an average of 41.41±7.97°C/km, significantly exceeding the global average of 30°C/km. The distribution of surface heat flow values ranges from 30.38 to 106.58 mW/m², with an average of 71.85±12.87 mW/m², surpassing the global average of 60 mW/m², confirming the basin as a typical "hot" basin. Under the influence of Pacific plate subduction, delamination and thermal erosion, the lithosperic thickness have thinned to 58.59 km. Radiogenic heat production in the thinned crust contributes 16.40 mW/m² (22.83% of total surface heat flow), while mantle-derived heat flow from upwelling melts triggered by stagnant slab dehydration accounts for 55.45 mW/m² (77.17% of total)

Controlled by lithospheric thinning and mantle upwelling, the Songliao Basin exhibits characteristic of "hot" basin attributues with a "hot mantle and cold crust" lithospheric thermal structure.

The development of karst reservoirs in the Maokou Formation of the Lower Permian in southern Sichuan Province is critical for conventional oil and gas exploration in this region. The development level of supergene karst reservoirs is directly controlled by the rich and diverse karst microgeomorphology, leading to strong lateral heterogeneity within these reservoirs.

In this study, 3D seismic data, combined with strata thickness, gradient structure tensor attributes, and geological body carving techniques, were employed to characterize the microgeomorphic features of the Dongwu karst in the Yunjin area. Moreover, favourable zones for surface karst reservoirs were predicted through model forward modeling and amplitude attributes.

The results suggested that significant differences in karst microgeomorphology existed during the Dongwu period in the Yunjin area, where a series of karst caves developed. These caves exhibit a seismic feature of a "pull-down" in the seismic event axis at the top boundary of the Maokou Formation. ② The distribution pattern of karst collapse bodies in the Yunjin syncline area was effectively characterized using gradient structure tensor attributes combined with geological body carving technology. The karst collapse bodies exhibit three distinct distribution patterns: Isolated, linear, or contiguous distributions. ③ The seismic amplitude on both sides of the karst collapse bodies is weakened, indicating strong karstification and the development of karst caves, which are favorable zones for reservoir development.

This research provides guidance for the subsequent prediction and exploration of reservoirs of the Maokou Formation.

Dry rock, a widely distributed and abundant geothermal resource, is of paramount importance in the reduction of fossil energy consumption, alleviating environmental pollution, and ensuring energy security. Enhanced geothermal systems are currently the main method for developing hot dry rock resources and are generally implemented through several methods, such as engineering site selection, geothermal drilling, thermal reservoir stimulation, flow testing, and power generation engineering. Among these, flow tests are a crucial link in undertaking thermal reservoir stimulation and power generation engineering and are used to form injection and production well groups, evaluate cycle circuits, expand heat exchange capacity, and lay the foundation for ultimately achieving power generation goals safely and stably. The implementation process of flow tests is long-term and complex, which can easily lead to problems such as insufficient connectivity, strong microseismic effects, liquid leakage, scaling of the circulating liquid, and insufficient equipment reliability. Therefore, flow tests at hot dry rock development sites internationally are often intertwined with drilling and reservoir stimulation and accompanied by scheme adjustments to gradually achieve the power generation goal.

This article briefly summarizes the flow test experiences and exploration directions of typical hot dry rock development EGS at home and abroad, elaborates on the influence of various factors on flow tests, and proposes development suggestions based on the actual situation at the Qinghai Gonghe site. In the past, improving the effectiveness of flow tests was achieved through methods such as adjusting the development layer and well groups, long-term circulation, hydraulic fracturing, and chemical stimulation. However, current technicians have focused on accurately obtaining engineering parameters and improving the design of well groups and reservoir stimulation processes.

In summary, the formulation of a flow test plan needs to fully consider geological factors and adapt to local conditions. Moreover, key technologies in flow tests, such as reservoir evaluation, reservoir stimulation, and engineering implementation, are worthy of in-depth research. The progress of these key technologies requires the establishment of numerical models with greater accuracy, improvements in the accuracy and stability of various monitoring techniques, the application of more diverse hydraulic fracturing and chemical stimulation processes, and the establishment of a more comprehensive risk prevention and control system for induced earthquakes. In addition, the application of new technologies is also a possible breakthrough. Supercritical carbon dioxide and liquid nitrogen fracturing technologies have advantages in thermal reservoir fracturing and energy enhancement in hot dry rock stimulation. Explosive fracturing technology has a certain effect on increasing the complexity of fractures near wellbores and enhancing the injection capacity of well groups. Finally, with the goal of experimental power generation and focusing on key issues in multi-well group flow tests, improving the construction process of flow tests is also an effective means to improve efficiency and reduce costs. With increasingly mature development technology, hot dry rock geothermal resources will become an important part of China's energy structure, playing a significant role in economic development and environmental protection.

To address this challenge, it is imperative to strengthen the research and geological exploration of new types of niobium deposits.

This study focuses on the clay rocks in the lower part of the Upper Permian Longtan Formation (P₃l) in Xingwen area of southern Sichuan. Based on the analysis of Nb contents in the collected samples, we combine various analytical techniques such as X-ray powder diffraction (XRD), scanning electron microscopy with energy dispersive spectrometer (SEM-EDS), and electron probe micro-analyzer (EPMA) to conduct mineral identification and quantitative analysis on Nb-enriched samples.

The results show that the content of Nb₂O₅ in the clay rocks ranges from 41×10⁻⁶ to 437×10⁻⁶, with an average of 187.2×10⁻⁶, reaching the lowest industrial indicator for weathering crust-type deposits. The enrichment of elements like Li, Ga, and others is also significant, making the clay layer rich in multiple critical metals and possessing considerable ore-forming potential. XRD analysis reveals a substantial presence of anatase in the Nb-rich clay rock. EPMA analysis indicates that the content of Nb₂O₅ in anatase ranges from 0.09% to 3.40%, with an average of 1.17%. Based on the Nb content in anatase, SEM-EDS scanning, and the Nb₂O₅ content of whole-rock samples, we conclude that Nb primarily exists in anatase as an isomorphous substitution, and some are adsorbed by clay minerals. Nb is predominantly inherited from the weathering products of minerals such as sphene in the Emeishan basalts, and the weathering degree has a significant impact on the enrichment and mineralization of Nb, with characteristic of weathering-sedimentary deposit.

The research results provide a scientific basis for the geological prospecting, resource evaluation, and comprehensive development and utilization of niobium resources.

To improve the calculation efficiency of slope stabilities for multiple profiles, this study prop oses a method based on spatial slope stability coefficient calculations.

The two-dimensional residual thrust method is incorporated with spatial profiles, which forms a residual thrust method for spatial profiles.

The calculation process of the residual thrust method for spatial profiles is optimized by the automatic generation and management of profiles within a three-dimensional engineering geological mode of a slope. In order to achieve rapid and efficient calculation of slope stability coefficients, a method based on multi-threaded parallel computation is introduced. Combining this approach with the residual thrust method for spatial profiles allows for more accurate and efficient computation of stability coefficients for multiple spatial profiles.

Using an open-pit mine in Xilinhot, Inner Mongolia, as a case study, we establish a three-dimensional slope engineering geological model for the internal dump site slope. On this basis, seven spatial profiles are automatically generated, and stability coefficients are calculated using multi-threaded parallel computation. Finally, the results are presented visually.

The research findings indicate that adopting multi-threaded parallel computation for calculating slope stability coefficients for multiple profiles can fully utilize computing resources and significantly improve computational efficiency. Furthermore, application in engineering projects confirms the feasibility and applicability of the proposed method.

The seepage in fractured rock masses is non-uniform and anisotropic, with its complexity reflected in the density, orientation, and trace length of individual fractures, as well as the connectivity of fracture networks. The connectivity of fracture networks poses a significant challenge in calculating the seepage parameters of three-dimensional (3D) fractured rock masses. Currently, methods for calculating seepage parameters of 3D fractured rock masses have their own advantages and disadvantages, depending on the model assumptions they used. To analyse the hydraulic anisotropy and permeability coefficients of fractured rock masses, a new method for solving the permeability coefficient of a 3D fractured rock mass based on dense sections is proposed using the dimensionality reduction.

This method involves simulating 3D fracture networks and approximating fractured rock masses using dense sections in different directions. The 3D fracture network is decomposed into multiple continuous two-dimensional (2D) sectional fracture networks, and the graph theory is applied to analysing the hydraulic connectivity and permeability paths. Water head boundary conditions are then set to calculate the permeability coefficient. By leveraging the geometric relationships among lines, surfaces, and volumes in 3D space, the calculated permeability coefficient is expressed as the directional permeability in 3D space, and a permeability tensor for the 3D fractured rock mass is constructed.

By treating the section permeability coefficient as a permeability ellipse, the new method calculates the equivalent permeability coefficient of the section and provides a formula for the equivalent permeability tensor of the rock mass. The scale effect and representation of anisotropy in the rock mass are also discussed. The 3D and 2D fracture networks are constructed, and different sizes of unit cells are intercepted to calculate the permeability coefficient, with a typical unit size determined to be 20 m. The method's feasibility is verified through field water pressure experiments.

This method provides a important reference for solving the permeability coefficients in different directions of heterogeneous fractured rock masses at various scales.

The spatial distribution analysis of earthquake-induced landslides offers critical insights for identifying potential geological disasters in affected regions, which is fundamental for informing post-disaster reconstruction planning, relocation strategies, site selection processes, and the development of effective geological disaster mitigation measures.

Taking the earthquake that struck Luding County, Ganzi, Sichuan Province, on September 5, 2022, as a case study, earthquake-induced landslides were initially identified using artificial visual 3D remote sensing data. This analysis was based on an optical image (DOM) with a resolution of 0.2 meter and a digital elevation model (DEM) with a resolution of 0.5 meter, both acquired post-earthquake. Field investigations and subsequent corrections were performed to validate and finalize the landslide inventory. Subsequently, the relationships between the distribution of earthquake-induced landslides and various geological factors, including topography, geological structures, and earthquake parameters, were systematically analyzed.

①The Luding earthquake triggered 9248 landslides, covering an area of approximately 680 km², with the majority classified as small- to medium-sized. The highest landslide density was observed at the tectonic intersection of the Xianshuihe fault, Daduhe fault, and Jinping mountain fault. The cumulative landslide area reached approximately 45.57 km², with an average landslide area of 4941 m². ② The spatial distribution of landslide in the earthquake-affected region is predominantly controlled by PGA and fault structures, with the majority of landslides concentrated in areas where PGA exceeds 0.6 g and within 1 km on either side of the seismogenic fault. Furthermore, landslide occurrence exhibits a negative correlation with proximity to water systems and roads. At a local scale, topographic factors significantly influence landslide distribution, with the highest frequency observed at elevations ranging from 1200 to 2400 m, on slopes inclined between 30° and 60°, and predominantly facing east or southeast. Additionally, landslides are more prevalent in areas underlain by hard rock strata. ③ The number and area of landslides exhibit an exponential relationship with the magnitude of the Luding earthquake. Leveraging high-precision data, our analysis identified a significantly higher number of earthquake-induced landslides compared to previous studies, with a reduced minimum landslide area and an increased total affected area.

The findings of this study have been implemented to support post-disaster recovery and reconstruction efforts in the Luding earthquake-affected region, providing a scientific basis for enhancing resilience and reducing future earthquake risks.

The bonding performance of the anchor-mortar interface is influenced by multi factors, yet existing studies predominantly focus on single-factor effects, leaving a gap in understanding the combined impact of multiple factors. This study aims to address this gap by systematically investigating the interfacial bonding performance under multifactorial conditions.

The anchor-mortar interface is taken as the research object. Electrochemical impedance spectroscopy is used to characterize the interfacial state and derive electrochemical parameters under different varying conditions. Pullout tests are conducted to quantify the bond strength of the anchor-mortar interface. By correlating electrochemical parameters with pullout load data at the completion of sample curing, the relationship between these parameters and interfacial performance is established. Additionally, the influence of the three factors on the bonding performance of the anchor-mortar interface is analyzed.

Sensitivity analysis of the orthogonal test results reveals that the pullout load is predominantly governed by the anchor rod diameter, whereas the pore solution resistance (R_s) is primarily influenced by the water-cement ratio. No dominant factor is identified for the charge transfer resistance (R_ct). During the early curing stage, two distinct interfacial states are observed under the combined influence of the three factors: A fully developed passivation film and partially developed passivation film. Within the tested parameter range, the pullout load increases with larger fine sand particle sizes and lower water-cement ratios. A positive correlation is observed between the pullout load and both the pore solution resistance (R_s) and charge transfer resistance (R_ct).

These findings provide critical insights for optimizing the formulation and application of anchored structural mortar.

This study investigates the dissolution and mechanical degradation of carbonate rocks under the influence of acidic leachate.

Limestone samples are subjected to acidic leachate under varying flow conditions and durations. The dissolution effects are evaluated through comprehensive analyses of apparent characteristics, mass loss, porosity, uniaxial compressive strength, acoustic emission activity, and other key indicators. These measurements elucidate the impact of the acidic leachate on the physical and mechanical properties of carbonate rocks.

Experimental results demonstrate that the dissolution rate and porosity of the rock samples increase with prolonged exposure time and higher flow rates, whereas the mechanical strength exhibits a corresponding decline. As dissolution progresses, the formation of increasingly thicker fluorite mineral layers on the rock surfaces is observed, which subsequently attenuates the dissolution rate. Furthermore, the failure mode of uniaxial samples shifts from shear-dominated to tensile-dominated damage. The acidic environment of phosphogypsum leachate induces the dissolution of internal minerals within limestone, resulting in significant changes to macromechanical properties.

These findings provide critical theoretical and experimental insights for assessing the stability of karst media under acidic wastewater conditions, as well as for optimizing the safety design of acidic wastewater treatment and tailings management systems.

A groundwater flow and solute transport model incorporating fault permeability was developed using GMS software.

The model simulated the impact of characteristic landfill pollutant seepage on the groundwater environment and a water delivery tunnel under various conditions, considering enhanced fault permeability.

The simulation results show that: Under normal conditions, due to the anti-seepage measures at the bottom of the plant, the polluant diffusion range is largely confiend within the plant site, and the maximum pollutants concentration at the center iremains below 0.5 mg/L. Under abnormal conditions (leakage scenarios), pollutants from leakage source No.1, No.2 and No.3 reached the water delivery tunnel location on the 800th day, 4015th day and 1095th day, respectively. The vertical diffusion range of pollutants progressively increased across the three leakage scenarios, whereas difference in the horizontal(area) diffusion range were minimal. Enhanced permeability of the F4 fault poses a greater environmental risk to pollutant migration in leakage source scenarios No.1 and No.3. In leakage source scenario No.2, however, the enhanced permeability of the F3 fault does not significantly increase environmental risk levels in the tunnel area. To safeguard the local soil and water environment, anti-seepage reinforcement of potentially affected tunnel areas is necessary.

The findings of this research provide scientific support for groundwater pollution prevention and risk assessment in the study area.

The complex structure of river-lake-groundwater interaction zone governs the migration and transformation and the fate of nutrients. Investigating the spatial-temporal distribution characteristic and controlling factors of nitrogen in groundwater holds significant implications for water quality and ecosystem protection.

This study focus on the river-lake-groundwater interaction zone along the middle reaches of the Yangtze River. Groundwater samples were collected from two confined aquifer profile and one phreatic aquifer profile during both monsoon season and non-monsoon season. Hydrogeochemical analysis was conducted, and the controlling factors of nitrogen enrichment were identified using random forest regression.

The results indicate that the main form of nitrogen in groundwater is ammonium (NH₄-N). The content of NH₄-N in confined groundwater during monsoon season (0.37−4.96 mg/L, average 2.07 mg/L) was significantly higher than that in the pore phreatic water (0−0.096 mg/L, average 0.012 mg/L).Low concentration of Cl⁻ / SO₄²⁻ and high concentration of CO₂ / HCO₃⁻ indicate that NH₄-N enrichment is closely related to the mineralization of nitrogen-bearing organic matter under natural condition. In contrast, in pore phreatic groundwater during the non-monsoon season, the NH₄-N concentration (average 1.25 mg/L) was significantly higher than during monsoon season(average 0.032 mg/L), displaying prominent seasonal variation. The concentration of Fe / Mn was also relatively high in this aquifer during the non-monsoon season. Combined with water-level fluctuation data, the pore phreatic water is significantly influenced by human activities. Following NH₄-N enrichment within the aquifer, feammox can occur.

This study enhances understanding of the occurrence patterns of nitrogen within river-lake-groundwater interaction zone and provides a theoretical basis for studying biogeochemical processes associated with the nitrogen cycle.

Marine ecological observation has become one of the essential approaches to supporting ocean exploration, ecological protection and restoration, disaster prevention and reduction, and early warning of ecological hazards. With the intensification of human activities and the accelerating impacts of global climate change, typical marine ecosystems are increasingly facing severe degradation and fragility. These threats include frequent red tides, eutrophication, and biodiversity loss in coastal waters, all of which impose significant economic and ecological costs. Against this backdrop, there is an urgent demand to establish advanced real-time, automated, and intelligent ecological observing systems. Such systems are expected to capture dynamic variations of marine environmental parameters and provide a scientific basis for sustainable marine management and the construction of ecological civilization.

This paper provides an overview of international progress in marine ecological observation technologies, focusing on nearshore, deep-sea, and seafloor monitoring systems. Developed marine countries have built integrated observation networks composed of buoys, coastal stations, underwater platforms, and satellite remote sensing, which together achieve long-term, continuous, and real-time monitoring of multiple environmental and ecological parameters. Typical examples include the United States Integrated Ocean Observing System (IOOS), the Ocean Observatories Initiative (OOI), Canada's NEPTUNE seafloor network, and Europe's EMSO project. These networks have demonstrated the capability to monitor not only hydrological and meteorological variables but also ecological indicators critical for predicting and preventing ecological disasters. By comparison, China's ocean ecological observation systems started relatively late and have mainly relied on coastal eco-buoys since the reform and opening-up period. Although more than 100 buoy-based monitoring stations have been deployed across major coastal provinces, these systems are still limited in scope, parameter diversity, and integration capacity. Current observation remains dominated by localized, single-point monitoring, with insufficient network connectivity and poor spatial resolution. Moreover, the existing systems are constrained by limited endurance, high maintenance costs, and weak data-sharing mechanisms, making it difficult to fully meet the requirements of ecological disaster early warning and ecosystem protection in the new era.

To overcome these limitations, this paper proposes that China's ecological observation should shift from simple buoy-based approaches toward the construction of collaborative, stereoscopic observation networks. Such networks should integrate multiple platforms: Shore-based towers, eco-buoys, underwater cabled observatories, shipborne systems, unmanned aerial vehicles, and satellite remote sensing: Forming a comprehensive "shore-sea-air-space" observation framework. With enhanced accuracy, extended parameters, and improved spatial and temporal coverage, these networks would enable dynamic, multi-dimensional monitoring of marine ecological processes. The establishment of such integrated networks is expected to play a crucial role in early warning of ecological disasters such as harmful algal blooms, coral bleaching, and ecosystem degradation. Furthermore, the accumulation of long-term, continuous datasets will provide essential support for assessing ecosystem health, guiding ecological restoration projects, and improving the resilience of coastal socio-economic systems. On a broader scale, advancing China's marine ecological observation capacity will also contribute to international collaboration in global ocean observing, strengthen data-sharing, and enhance China's role in addressing global marine environmental challenges. In summary, the paper highlights both the progress and the limitations of China's current ecological observation efforts and argues for a transformation toward advanced, stereoscopic, and multi-platform ecological observing systems. This transition is vital for improving marine ecological protection, supporting disaster prevention and mitigation, and achieving the long-term goal of sustainable ocean development under ecological civilization.

Land resource assessment and specialty industry cultivation are pivotal for China's rural revitalization and poverty alleviation initiaives. To optimize regional agricultural planning and specialty sector development, this study implemented a selenium-rich land resource survey in Tunliu District, Changzhi City, Shanxi Province.

A comprehensive land quality evaluation framework was established, integrating three dimensions: Selenium-rich industrial potential, ecological integrity, and arable land productivity. The fuzzy mathematical method, entropy weighting, and composite index modeling were synergistically applied for land classification, with spatial patterns visualized through ArcGIS.

The results show that: ①54.47% of the study area contains medium-to-high selenium concentrations, predominantly clustered in eastern regions. Arable land productivity demonstrates spatial heterogeneity, with lower values in western zones versus higher eastern values. ②299.33 km² (26.21% of total) of first-class quality land was identified in eastern plains. ③Wheat, chili peppers, and bell peppers exhibit 100% selenium enrichment compliance with crops in good condition. ④Based on land and crop evaluation results, and considering spatial planning, the land is classified into three types: A, B, and C. Zone A (Premium): Supports selenium-rich wheat/vegetable cultivation and agro-tourism (eastern plains). Zone B (Transitional): Suitable for diversified cropping with selenium supplementation. Zone C (Marginal): Recommended for agricultural product processing and trade hubs due to suboptimal selenium (<0.30 mg/kg) and nutrient levels.

This study provides theoretical support and scientific recommendations for planning the selenium-rich agricultural industry in Tunliu District and for its integrated regional development.

Seabed carbon dioxide (CO₂) geological sequestration technology has gained popularity in the fields of carbon sequestration and carbon neutralization. In the northern South China Sea, favorable conditions for the formation of CO₂ hydrates exist in seabed sediments, where hydrates forming in cracks and pores may block the further upward migration of CO₂ and generate self-sealing capacities. However, the effects of CO₂ leakage in fractures and the instability of hydrate plugging remain unclear.

A visualization experimental platform for hydrate growth and the instability process under water injection supercharging at high pressure and low temperature was used to observe CO₂ hydrate formation. The platform simulated the conditions of seafloor sedimentary cover at 2℃ and 3-4 MPa, and evaluated hydrate instability and plugging effects using parameters of breakthrough pressure, breakthrough pressure difference, duration, permeability coefficient at the initial instability stage, and plugging rate.

The experimental results show that hydrate formation proceeds through four stages: Nucleation, expansion, formation, and aggregation. Hydrates formed in cracks effectively block the migration of fluids such as water and CO₂; however, instability begins when fluid pressure increases and reaches the critical breakthrough pressure. The instability process can be divided into two stages: Particle size degradation and surface friction failure. The hydrate mass core becomes unstable first, while the sealing state persists until the surface of the hydrate fails due to friction. The breakthrough pressure ranges from 6.414 to 6.966 MPa, and the breakthrough pressure difference is between 2.403 and 3.203 MPa. Hydrate instability is primarily influenced by flow rate, with the flow rate affecting the instability rate through interface effects. The breakthrough pressure of hydrates is mainly determined by temperature and pressure, while flow rate influences the exact time when hydrates enter instability. At 3-4 MPa seafloor sedimentary cover, the sealing rate is 99.0%-99.6%, and the permeability coefficient at the initial instability stage ranges from 0.555×10⁻³ to 1.260×10⁻³ μm².

The experimental results offer a reference for risk assessments of overlying layers in similar conditions in the South China Sea for CO₂ seabed geological sequestration. To maintain the sealing effect of CO₂ hydrates, the pressure difference between the capped hydrate layer and the actual seabed pressure should be less than 2 MPa.

Geophysical methods can effectively monitor the dynamics of water flow and material transport in 4D hydrogeological processes, and its imaging accuracy is often closely related to the monitoring scheme. Taking the commonly used electrical resistivity tomography (ERT) as an example, obtaining good imaging accuracy often requires a large number of electrode arrays, leading to a long data acquisition time and inability to respond in a timely manner to 4D hydrogeological dynamics. Previous optimization studies of ERT monitoring schemes have mainly focused on surface ERT, whereas cross-hole ERT (CHERT) has received far less attention.

Due to the advantages of CHERT in high-precision characterization, this study proposes using Bayesian experimental design to optimize CHERT monitoring scheme. Through laboratory static and dynamic tests and a field application, data acquisition time and imaging accuracy between the optimized and traditional electrode configurations to evaluate the effectiveness of the Bayesian-based optimization.

Laboratory tests demonstrate that the optimized monitoring scheme can reduce acquisition time by approximately 75%, and the inversion results more accurately delineate the evolving resistivity anomaly zones, thereby mitigating the lag effect observed in traditional schemes. The field application shows that the optimized scheme can reduce monitoring time by approximately 95%.

The optimization of CHERT electrode configurations based on Bayesian experimental design provides a technical basis for efficient monitoring of 4D hydrogeological processes.

To reduce the risk of landslide disasters and further ensure regional sustainable development, conducting research effectively on landslide monitoring and prediction is of great practical significance. This paper aims to continuously improve the prevention and control level of landslide disasters by studying the key technologies and methods in landslide monitoring and prediction, and will analyze the efficiency and accuracy of various algorithms in scenarios of landslide monitoring and prediction.

In terms of feature extraction technology, the performance of three image feature matching algorithms, namely scale-invariant feature transform (SIFT), speeded up robust features (SURF) and assessment of scale-invariant features transform (ASIFT), were compared and analyzed. Comparative results demonstrate that ASIFT has significant advantages in the number of matches, precision rate and recall rate, and is especially suitable for complex environmental scenarios with high-accuracy requirements. In terms of optical flow analysis technology, the application effects based on the Lucas-Kanade sparse optical flow method and the Horn-Schunck dense optical flow method were discussed. Among them, the Lucas-Kanade sparse optical flow method has high computational efficiency and is suitable for real-time application scenarios, but there is a risk of missing important motion information. The Horn-Schunck dense optical flow method can provide comprehensive optical flow field information and is suitable for complex environmental scenarios. However, it has the drawback of high computational complexity and thus is difficult to be used in real-time processing. In terms of landslide susceptibility prediction, the advantages and disadvantages of the application of classic machine learning methods such as support vector machine (SVM), decision tree (DT), and random forest (RF) in landslide prediction are introduced in detail. The model performance based on particle swarm optimization-support vector machine (PSO-SVM) is mainly studied. This model optimizes hyperparameters, the classification accuracy, generalization ability and prediction accuracy of the model have been significantly improved. Furthermore, by introducing the Faster R-CNN model and leveraging its advanced convolutional neural network architecture, this paper realizes the automatic identification and classification of landslide events in complex scenarios, further enhancing the efficiency and accuracy of landslide monitoring and early warning.

The research shows that the accuracy rate of local feature extraction by ASIFT is 0.84, the tracking error of the Lucas-Kanade sparse optical flow method is 0.12, the root mean square error of the PSO-SVM model is 0.52, and the confidence level of the Faster R-CNN model in the automatic recognition and classification of landslide images can reach 0.98. The comprehensive performance is significantly improved compared with other algorithms in this paper.

In summary, by introducing artificial intelligence algorithms and integrating multi-disciplinary technical means, this paper has comprehensively enhanced the efficiency and accuracy of landslide monitoring and prediction technology. The research results provide more powerful technical support for the prevention and control of landslide geological disasters.

In tunnel engineering, the prerequisite for tunnel design and construction safety is to accurately assess the amount of deformation of the rock surrounding a tunnel.

In this work, the firefly algorithm (FA), whale optimization algorithm (WOA), and gray wolf optimization algorithm (GWO) are combined with optimized support vector regression (SVR), and three intelligent hybrid swarm optimization prediction models are constructed to predict the deformation of extruding surrounding rock tunnels. A database containing 62 samples was constructed, and seven initial parameters of tunnels and surrounding rock were selected as the input parameters of the prediction models, with the radial deformation of tunnels as the output quantity. The coefficient of determination (R²), root-mean-square error (RMSE), and mean absolute error (MAE) were selected as the evaluation indices of the model's prediction performance. Finally, the effects of different input parameters on the prediction results of tunnel rock deformation were evaluated via normalized mutual information values.

Compared with the GWO-SVR and WOA-SVR models, the FA-SVR model demonstrated superior predictive performance during both the training and testing phases. The corresponding R² values were 0.9634 and 0.9648, the RMSE values were 18.786 and 14.699, and the MAE values were 9.460 and 11.170 for the training and testing sets, respectively. The ranking of predictive capability was as follows: FA-SVR>WOA-SVR>GWO-SVR. The firefly algorithm, whale optimization algorithm, and gray wolf optimization algorithm can improve the prediction performance of the support vector regression model. The FA-SVR model has the best prediction effect, and the optimized hybrid prediction model performs significantly better than the classical models. The sensitivity analysis reveals that the joint frequency is the most important parameter affecting the predicted value of deformation of the rock surrounding a tunnel.

The research results can provide an important reference for the safety control of tunnel engineering.