In Sichuan Basin, two trillion-cubic-meter level giant gas fields, Anyue and Penglai gas fields, have been discovered in deep to
ultra-deep formations in the Sinian–Cambrian. It is urgent to clarify the main controlling factors for hydrocarbon accumulation to determine the replacement fields. Based on previous studies and combined with geological and geochemical data, a systematic study has been conducted on gas genesis and the main controlling factors for hydrocarbon accumulation in Anyue and Penglai gas fields. The study results conclude that the formation of deep to ultra-deep giant gas fields was mainly controlled by four factors: (1) Two intracontinental rifts were developed during the Late Sinian–Early Cambrian, namely Deyang–Anyue and Wanyuan–Dazhou rifts, developing high-quality source rocks in the Lower Cambrian, with a source rock thickness in the rifts 3–4 times and a gas generation intensity 2–3 times those in neighboring areas. (2) Anyue–Fengjie shelf-rimmed platform was formed during the Late Sinian–Early Cambrian, which controlled the formation of four sets of high-quality platform marginal mound beach or granular beach facies reservoirs, i.e., the second and fourth members of the Sinian Dengying Formation, lower member of the Cambrian Canglangpu Formation, and the Cambrian Longwangmiao Formation, with large thickness and good storage performance of platform marginal reservoirs. (3) The long-term inheritance and development of Gaoshiti–Moxi huge paleo-uplift formed during Tongwan movement stage controlled the formation of large structural–lithologic traps in the core of the paleo-uplift and lithologic traps in the slope zone, with areas of 7500 km² and 5720 km² of Dengying Formation trap groups. (4) The large ancient oil reservoirs cracked and insitu accumulation occurred, with high accumulation efficiency. A large number of liquid hydrocarbon inclusions were observed in various host minerals, with abundance of 10%–80%. The content of reservoir bitumen in gas-rich zone is high, mainly ranging in 1%–8%. Based on factors such as the development degrees of platform marginal mound beach body and intraplatform beach body, large-scale source rocks, structure of the target layer, and exploration degree, as well as geological conditions for stereoscopic hydrocarbon accumulation and composite enrichment, three types of favorable zones for the formation of large and giant gas fields have been proposed. Type Ⅰ zone includes platform marginal beach/intraplatform beach zone where Anyue and Penglai gas fields have been discovered. Type Ⅱ zone is the platform marginal beach/intraplatform beach development zone where breakthroughs have been made or has the potential for large-scale exploration, such as Hongya–
Leshan, Yilong–Guangyuan, Rongchang–Gulin, Bazhong–Dazhou, and Shizhu–Lichuan. Type Ⅲ zone is the intraplatform beach development zone, such as Chongqing–Liangping and Weiyuan–Luzhou. The study results have important reference significance for oil and gas exploration in deep to ultra-deep formations.

The high-efficiency exploration is key initiatives to establish independent and controllable overseas core oil and gas production
zone, and promote the high-quality development of overseas oil and gas business. In response to the four major characteristics of overseas exploration, such as restrictions, risks, economy, and internationalization, CNPC has researched and formed the “1435” high-efficiency exploration system in recent years. “1” represents the goal of creating world-class value creation of overseas exploration. “4” refers to four sets of management systems characterized by “unified management and centralized decision-making”, including business management, risk exploration, whole lifecycle engineering, and collaborative innovation. “3” refers to the evaluation of three aspects: exploration targets, single exploration project, and full cycle exploration. “5” refers to five major projects including “mountain area exploration” technology, risk exploration, centralized exploration, fine exploration, and rolling exploration. As a result, significant achievements have been made in four aspects, i.e., fields without breakthrough for long periods, exploration in new areas and new fields, expanding reserve scale in mature areas, and increasing reserves and production in old areas, discovering two world-class one-billion-ton level oil fields, a trillion-cubic-meter level large natural gas zone, one 500-million-ton level oil field group, and 10 high-quality reserve areas. This effectively supports the beneficial and steady production of 100-million-ton equity production and further consolidates the resource foundation for high-quality development of overseas oil and gas business. The “1435” high-efficiency exploration system will provide solid support for China’s oil companies to continuously enhance their global resource control and ensure the national energy security.

The fourth member of Xujiahe Formation (Xu 4 member) in Jianyang block of Tianfu Gas Field is an emerging field for increasing
reserves of continental facies near-source tight gas in Sichuan Basin, and the study of its reservoir characteristics and main controlling factors is very crucial. Using casting thin section, grain size, bulk rock, scanning electron microscope and high-pressure mercury injection data, the basic characteristics and classification of reservoirs are discussed in detail, and the main controlling factors for reservoir development and evolution are systematically analyzed in multiple aspects. The delta–lake sedimentary system was mainly developed in Xu 4 member in Jianyang block, and the high-quality reservoirs were mainly concentrated in delta front underwater distributary channel microfacies. The reservoir lithology is mainly composed of lithic feldspar sandstone and feldspar lithic sandstone, with intragranular and intergranular dissolution pores commonly observed. The reservoir porosity generally ranges from 6% to 8%, and the permeability is in the range of 0.05–0.3 mD. Based on the comprehensive study of sedimentary facies zone, reservoir characteristics and fracture development degree, the classification evaluation standard for Class Ⅰ– Ⅳ tight sandstone reservoirs in Xu 4 member is established, and Class Ⅱ fractured-porosity type reservoir is dominant in the study area. The delta front underwater distributary channel and mouth bar microfacies laid the foundation for reservoir development. The intragranular (intergranular) dissolution pores of feldspar and lithics formed by hydrocarbon generation and acid drainage are the key to the development of high-quality reservoirs, and tectonic fractures are important supplement to reservoir space. In the foreland slope zone, source rock–reservoir configuration is superior, and tectonic faults are well developed, in which delta front underwater distributary channel and mouth bar sand bodies improved by late fault activity are the main orientation for tight gas exploration.

In order to identify pore development characteristics and controlling factors for lacustrine carbonate reservoirs in Qianjiang
Formation, Qianjiang Sag, Jianghan Basin, the lithofacies and reservoir physical properties have systematically been analyzed of carbonated rocks in the third member of Qianjiang Formation by means of XRD, thin section, porosity, and high-pressure mercury injection data in typical wells. The study results show that: (1) A total of eight rock types have been identified, i.e., granular carbonate rock, micritic carbonate rock, granular mixed rock, fine-grained mixed rock, felsic-rich clastic rock, clayey shale, sulphate rock, and salt rock. (2) Various rock types are superimposed with each other, forming four major types of lithofacies combinations, namely felsic-rich clastic rock and shale lithofacies combination, granular mixed rock and fine-grained mixed rock intercalated with shale lithofacies combination, granular carbonate rock and micritic carbonate rock intercalated with shale lithofacies combination, and salt rock intercalated with shale lithofacies combination from north to south. (3) The total porosity of various rock types is similar, with an average range of 5.5%–7.3%, but pore sizes vary significantly. Granular carbonate rock and felsic-rich clastic rock have relatively large pore sizes, with average pore volume of micron-sized pores accounting for more than 10% and average throat radius of higher than 140 nm indicated by high-pressure mercury injection data; While the micritic carbonate rock and fine-grained mixed rock have relatively small pore throats, with pore volume of micron-sized pores accounting for less than 5% and average throat radius of lower than 10 nm. It is pointed out that granular carbonate rock and felsic-rich clastic rock have good reservoir physical properties and large pore size, which are conducive to hydrocarbon accumulation, flow, and production, showing predominant lithofacies types in lacustrine carbonate reservoirs and the most favorable exploration targets in the study area.

Exploration practice indicates that the reservoir sedimentary characteristics are key factors affecting oil and gas production in subsalt lacustrine carbonate rocks in Santos Basin. The high-quality reservoirs are mainly composed of microbial reefs and shoal carbonate rocks deposited in paleo uplifts. Based on seismic structural interpretation and gravity data, the sub-salt structural units and distribution of paleo uplifts in Santos Basin have been redefined. The study results indicate that the basin was divided into five primary structural units, in which the outer uplift zone was originated from the abandoned Ocean Ridge after transition during breakup of the South Atlantic, which was very conducive to the development of microbial reefs. The drilling and seismic data have been used to analyze the core facies, logging facies, seismic facies, and sedimentary facies, identifying the distribution of sedimentary facies of sub-salt strata in the core depression zone. In addition, combined with the structural characteristics and paleogeography during rift period, four types of sedimentary modes of high-quality carbonate reservoirs have been established, including the large gentle slope, isolated high basement, intra-platform highland, and steep slope types. Among them, the large gentle slope and steep slope sedimentary modes were controlled by monoclinal fault blocks, and the intraplatform highland sedimentary mode was controlled by fault-uplift structure, and they are mainly distributed in the outer uplift zone. The isolated high basement sedimentary mode was controlled by fault horst, which was mainly distributed in the western depression zone and outer depression zone.

In Tiaohu and Malang sags of Santanghu Basin, the sediment sources were mainly fallout volcanic ash and endogenous carbonates
during the deposition of second member of the Permian Lucaogou Formation (Lu 2 member). This unique sedimentary setting resulted in a highly complex formation mechanism of high-quality fine-grained reservoirs. Based on petrological and geochemical data, a systematic study has been conducted on sedimentary characteristics and diagenetic process of Lu 2 member fine-grained rocks. The study results indicate that the lithology had a minimal control over physical properties, pore structure, and fluid mobility of the fine-grained rocks, and high-quality reservoirs were formed after diagenesis of fine-grained rocks regardless of lithology. Except for dolomite, organic matter was generally rich in fine-grained rocks with other lithologies. Among them, cyanobacteria were predominant in tuff and dolomitic tuff, while green algae were developed in tuff, dolomitic tuff, and tuffaceous dolomite. The widespread organic matter underwent significant hydrocarbon generation during the middle diagenetic stage A, which not only altered the diagenetic environment, but also effectively and extensively dissolved soluble minerals in the fine-grained reservoirs. Compaction was the primary factor leading to the tightness of fine-grained rocks, while cementation had a relatively minor impact on reservoir quality. Dissolution was the key mechanism for forming high-quality tuff reservoirs, and multi-stage dolomitization significantly enhanced the storage capacity of dolomite reservoirs. Based on the intensity of dissolution and dolomitization, five types of diagenetic facies have been classified for Lu 2 member fine-grained rocks. Among them, the intense dissolution–weak dolomitization facies and the moderate dissolution–moderate dolomitization facies were the most favorable diagenetic facies. Using Matlab random forest method, diagenetic facies in a single well has been predicted, and the distribution area has been determined. The study achievements not only provide valuable insights into identifying favorable facies zones for shale oil exploration in Lu 2 member, but also contribute to understanding the formation mechanism of this unique type of high-quality fine-grained reservoirs.

The carbonate reservoirs in the fourth member of Dengying Formation (Deng 4 member) in Penglai Gas Field, Sichuan Basin,
are one of the key targets for increasing reserves and production in onshore petroleum exploration during the 14^th Five–Year Plan period. Due to the ancient age, great burial depth, and high heterogeneity, the reservoir spatial distribution prediction remains challenging. Based on geological, drilling, logging and seismic characteristics, a well–seismic integrated classification and prediction technology has been established for deep carbonate reservoirs in db1 block, Penglai Gas Field, which enables analyzing formation characteristics and distribution pattern of deep carbonate reservoirs, conducting inversion and prediction of heterogeneous reservoir distribution, and finely characterizing the heterogeneity of reservoir spatial distribution. The study results indicate that: (1) Deng 4 member reservoirs are concentrated in the upper sub-member, with three configuration patterns, i.e., dual upper and lower reservoirs, locally distributed upper reservoir and well-developed lower reservoir, and a single reservoir. (2) The good and medium reservoirs in the study area have a small thickness and sporadic distribution, while the poor reservoir has a great thickness and good continuity in both vertical and lateral directions, which is primarily distributed near wells ps106, db1, ps11, and adjacent areas. (3) The successful classification and prediction of deep carbonate reservoirs not only validate the applicability and reliability of the developed technology, but also provide valuable references for reservoir prediction in Penglai Gas Field and similar gas fields.

The Oligocene Huagang Formation is mainly composed of a shallow-water delta sedimentary system in the central–southern Xihu
Sag, East China Sea shelf basin. It is difficult to predict the reservoir due to the large burial depth of the main target layer, similar impedance of sandstone and mudstone, complex channel stages and superposition relationship. In response to this problem, the Shuey approximate gradient term of Zoeppritz equation has been studied in detail, and seismic forward analysis has been conducted, which clarify that AVO gradient is related to Poisson’s ratio, and it is the most sensitive and consistent to reservoir response. Using AVO gradient to finely delineate the channel boundary, a total of four distributary channels have been identified in H5 layer. Channels No.1 and No.4 are characterized by NW–SE trending, with the average width, width-to-thickness ratio and curvature of 0.7 km and 0.8 km, 29 and 28, 1.04 and 1.06, respectively. Channels No.2 and No.3 are characterized by NE–SW trending, with average width, width-to-thickness ratio and curvature of 0.85 km and 3.1 km, 26 and 70, 1.40 and 1.09, respectively. Based on sequence stratigraphy and relative isochronous surface flattening method, the stages of channels have been identified, including channels No.1, No.2, No.3 and No.4 from early to late stages. After multi-information cross-validation analysis such as seismic response characteristics of various channels and possible gas bearing flat points, it is regarded that the four channels in H5 layer are disconnected and have the conditions for independent hydrocarbon accumulation. In addition, a “fault-sand coupling” hydrocarbon accumulation mode for lithologic traps in anticlinal wings has been established, which indicates that the anticlinal wings have better hydrocarbon accumulation conditions than the core. Based on the above understanding, one exploration well has been drilled each for channel No.4 in structure A and channel No.3 in structure B, and good exploration results have been obtained in lithologic oil and gas reservoirs for the first time in the study area, confirming the applicability and reliability of the technique and method, and putting forward the favorable exploration zone in the near future.

Currently, there are numerous methods for evaluating rock brittleness index, which are mainly based on mineral composition or
rock mechanical properties, but most evaluation indicators are costly and time-consuming to obtain. By utilizing machine learning technique, a rock brittleness index prediction method based on Stacking ensemble learning concept has been proposed, which involves parallel training of Gradient Boosting Decision Tree model (GBDT), Random Forest model (RF), Naive Decision Tree model (DT), Support Vector Regression model (SVR), and LightGBM model. After hyperparameter optimization using a tree-structured Parzen estimator for each model, the XGBoost model has sequentially been used to fuse training results of the base models to achieve rapid parameter optimization and prediction of rock brittleness index. The study results indicate that the prediction results using two-layer Stacking model optimized by tree-structured Parzen estimator show significant advantages compared to those by the base models, with the explained variance score (EVS) reaching up to 0.97, and the coefficient of determination (R²) reaching a maximum of 0.967. Given the same dataset, this model obtains the lowest Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), indicating that it can effectively fit the variation pattern of rock brittleness index in the technical context of supervision and learning, which verifies its practical value in predicting rock brittleness index.

In Qingcheng area, Ordos Basin, the interlayered type shale oil reservoir is characterized by high heterogeneity in both lateral
and vertical directions, low degree of natural fracture development, and large two-direction horizontal stress difference. The preliminary microseismic data indicates that there is distinct feature of double-wing fractures, and the fracture pattern is relatively simple. In order to further understand the post-fracturing fracture pattern and spatial distribution, core section has systematically been taken in a high-inclination well in Qingcheng shale oil hydraulic fracturing test site. Using multiple methods such as CT scanning, core observation, microscopic imaging, and well logging data, the fine description of fracture occurrence, fracture surface morphology, and filling characteristics is conducted, and the fractures are classified and identified. Meanwhile, the dye proppant, mud tracer, and spatial distance and distribution law are used to perform corresponding regression and comprehensive analysis on hydraulic fractures. The study results show that different types of fractures show different characteristics. There is a low proportion of natural tectonic fractures, but a high proportion of bedding fractures and hydraulic fractures with significant characteristics. The hydraulic fractures are characterized by clustering features, and the fracture number is much higher than the corresponding stages in adjacent hydraulic fracturing wells. There are differences in fracture propagation in local areas, but they generally extend along the principal stress direction under the influence of stress difference and azimuth. There are few distinctly propped fractures, but the mud proppant is widely distributed in fractures with a higher proportion of small particles. The tracer analysis shows that the number of hydraulic fractures in each stage is the key contributor to oil production. The fracture recognition, classification, and regression
method obtained by integrating multiple methods can be used as a reference for post-fracturing coring fracture analysis of other unconventional oil and gas reservoirs, and the fracture understanding has certain guiding significance for further fracturing optimization.

The deep coal seam reservoirs are characterized by ultra-low permeability, well-developed micropores, but poor connectivity, and
great difficulty in reservoir reconstruction. In order to deepen the understanding of reservoir stimulation mechanism of No.8+9 deep coal seams in the eastern margin of Ordos Basin and to provide guidance for the subsequent fracturing construction, a hybrid model with two-stage synergistic architecture has been proposed, and the characteristics of fracturing construction curves of No.8+9 deep coal seams and sand-coal stacked layers have systematically been analyzed, which reveals the influence mechanism of parameters such as sand addition amount, number of stages, and perforation technology on the fluctuation of construction pressure, and clarifies the production capacity of vertical and directional wells with various fracturing parameters. In addition, the targeted fracturing technology has been put forward and applied to field practice. The study results show that the “TSLANet–Kmeans++ (DTW)” hybrid model has the best performance under various classification conditions, and it has the best discrimination when classifying fracturing curve of deep coal seams into four types, i.e., pressure plateau after high pressure fracturing, steadily pressure rising, steady pressure decline, and difficult sand addition, with the accuracy of the model reaching up to 92.7%. The staged plugging technology and pre-pad fluid volume have a great influence on fracturing pressure. For wells with high rock breaking pressure peaks, multi-stage plugging, high pre-pad fluid ratio and low viscous fluid ratio can be used to reduce risks of fracturing complex caused by high fracturing pressure. The main controlling factors for well production capacity include liquid volume, sand addition amount, displacement, and sand ratio. It is beneficial for enhancing well production capacity by fracturing with less water, controlling liquid volume, increasing sand–liquid ratio, and increasing displacement. A composite fracturing and reconstruction technology for deep coal seams with the core idea of “repeatable low-damage fracturing fluid + multi-stage plugging + high displacement + temporary plugging at the fracture opening end” has been proposed and successfully been applied in wells A-18 and B-4H, with a steady single well gas rate exceeding 10×10⁴ m³/d, which provides theoretical basis and technical support for the high-efficiency development of deep CBM in Ordos Basin.

The evaluation of comprehensive sweet spots of shale oil reservoir involves both geological and engineering factors. A conventional logging curve based fractal characteristic method has been proposed to evaluate comprehensive sweet spot interval of shale oil reservoir in horizontal fracturing wells. Taking Chaiping 2 and Chaiping 4 horizontal fracturing wells in sweet spot zone of Yingxiongling shale oil reservoir in Qinghai Oilfield as case studies, fractal characteristic parameters are extracted from conventional well logging data. By integrating post-fracturing production profile monitoring data, and applying gray relation analysis method, parameters such as weighted multifractal spectrum width (Δα') and weighted fractal dimension (D') are introduced to establish a production capacity evaluation standard based on fractal characterization. The results show that the weighted fractal evaluation index for Class Ⅰ comprehensive sweet spot is 0.75<Δα'≤1, 0<D'≤0.25, that for Class Ⅱ comprehensive sweet spot is 0.35<Δα'≤0.75, 0.25<D'≤0.8, and that for Class Ⅲ comprehensive sweet spot is 0<Δα'≤0.35, 0.8<D'≤1. After validated by Well Chaiping 14-2, it is indicated that an overall accuracy reaches up 72.7%, which is significantly higher than 30.1% by conventional method. This integrated fractal evaluation approach provides technical support for precise identification and high-efficiency development of comprehensive sweet spot intervals of shale oil reservoirs.