To clarify the dynamic evolution characteristics of reservoir flow units during water injection development, the upper member of the Neogene Guantao formation in Block M of Gudao Oilfield is taken as a case study. Based on logging data, water injection profile test data, subwell data, and production performance data, among others, the flow zone index (FZI static) was proposed as the static evaluation parameter of the flow unit. The relationship between cumulative water injection (W_T) and FZI change () was fitted. Hence, the caused by water injection is combined with the static parameter of flow unit evaluation (FZI static) as the dynamic parameter of flow unit evaluation (FZI dynamic). The comprehensively evaluated reality of flow units in different periods is characterized by formula . The study shows that as the division standard of the flow unit, the FZI has a good correlation with the test results of the water absorption profile, such as water absorption intensity and relative water injection volume. Using the FZI as the static parameter of flow unit evaluation, four types of flow units were divided as follows: type I flow unit, ; type II flow unit, ; type III flow unit, ; type IV flow unit, . The reservoir porosity and permeability characteristics of different flow units are highly correlated. Moreover, the relative permeability curve confirms that different flow units have different seepage capacities. Though, comparing characteristic reservoir parameters in different periods, the reservoir’s physical parameters became more conducive to fluid flow with the water injection development. The increase in the same water injection rate for type I and II flow units was greater than that for type III and IV flow units. Furthermore, when type I flow units were continuously distributed in a large area, high water consumption bands were formed, absorbing most of the water injection in the water injection wells. Hence, the waterflood efficiency was low. The change in different flow units was mainly controlled by the injection production well pattern and W_T. Combined with the relative change characteristics of interlayer flow units, the changes can be divided into increasing type change and decreasing type changes. Finally, according to the distribution characteristics of different flow units, oil saturation, and water flooding results, strategies for tapping the potential of the remaining oil from three aspects (plane, interlayer, and inner layer) were formulated.

Machine learning is the main technical means for lithofacies logging identification. As the main target of shale oil spatial distribution prediction, mud shale petrography is subjected to the constraints of stratigraphic inhomogeneity and logging information redundancy. Therefore, choosing the most applicable machine learning method for different geological characteristics and data situations is one of the key aspects of high-precision lithofacies identification. However, only a few studies have been conducted on the applicability of machine learning methods for mud shale petrography. This paper aims to identify lithofacies using commonly used machine learning methods. The study employs five supervised learning algorithms, namely Random Forest Algorithm (RF), BP Neural Network Algorithm (BPANN), Gradient Boosting Decision Tree Method (GBDT), Nearest Neighbor Method (KNN), and Vector Machine Method (SVM), as well as four unsupervised learning algorithms, namely K-means, DBSCAN, SOM, and MRGC. The results are evaluated using the confusion matrix, which provides the accuracy of each algorithm. The GBDT algorithm has better accuracy in supervised learning, while the K-means and DBSCAN algorithms have higher accuracy in unsupervised learning. Based on the comparison of different algorithms, it can be concluded that shale lithofacies identification poses challenges due to limited sample data and high overlapping degree of type distribution areas. Therefore, selecting the appropriate algorithm is crucial. Although supervised machine learning algorithms are generally accurate, they are limited by the data volume of lithofacies samples. Future research should focus on how to make the most of limited samples for supervised learning and combine unsupervised learning algorithms to explore lithofacies types of non-coring wells.

The current seismic prediction methods of the shale brittleness index are all based on the pre-stack seismic inversion of elastic parameters, and the elastic parameters are transformed by Rickman and other simple linear mathematical relationship formulas. In order to address the low accuracy of the seismic prediction results for the brittleness index, this study proposes a method for predicting the brittleness index of shale reservoirs based on an error backpropagation neural network (BP neural network). The continuous static rock elastic parameters were calculated by fitting the triaxial test data with well logging data, and the static elastic parameters with good correlation with the brittleness index of shale minerals were selected as the sample data of the BP neural network model. A dataset of 1970 data points, characterized by Young’s modulus, Poisson’s ratio, shear modulus, and the mineral brittleness index, was constructed. A total of 367 sets of data points from well Z4 were randomly retained as model validation data, and 1603 sets of data points from the other three wells were divided into model training data and test data at a ratio of 7:3. The calculation accuracy of the model with different numbers of nodes was analyzed and the key parameters of the BP neural network structure such as the number of input layers, the number of output layers, the number of hidden layers, and the number of neurons were determined. The gradient descent method was used to determine the weight and bias of the model parameters with the smallest error, the BP neural network model was trained, and the stability of the brittleness index prediction model of the BP neural network was verified by posterior data. After obtaining Young’s modulus, Poisson’s ratio, and shear modulus through pre-stack seismic inversion, the BP neural network model established in this study was used to predict the brittleness index distribution of the target layer in the study area. Compared with the conventional Rickman method, the prediction coincidence rate is 69.16%, and the prediction coincidence rate between the prediction results and the real value is 95.79%, which is 26.63% higher. The BP neural network method proposed in this paper provides a reliable new method for seismic prediction of the shale reservoir brittleness index, which has important practical significance for clarifying the shale gas development scheme and improving shale gas exploitation efficiency.

In order to improve the combined exploitation efficiency of horizontal and vertical wells, and given the fact that the complex and varied spatial structure of sand bodies in the Fuyu oil layer in the Qian’an area of Songliao Basin leads to significant differences in production characteristics of horizontal wells, the sand body types and internal spatial structure are finely dissected according to the theory of configuration analysis, and the internal spatial structure is divided into three configuration styles: spatial clipping type, overlapping type, and separation type. Then, by comparing the productivity characteristics of horizontal wells with different configurations of sand bodies, combined with the analysis of fluid flow law under horizontal well volume fracturing, a main fracture–fracture network–matrix coupled fluid flow model in a tight reservoir based on composite sand body configuration is established. Combined with the actual volume fracturing the horizontal well area, the productivity curves of each cluster in the horizontal section after numerical simulation of volume fracturing of typical horizontal well groups are extracted, which are divided into four types: high-yield stable type, high-yield two-stage type, high-yield rapid-decline type, and low-yield rapid-decline type, and the coupling relationship between the productivity characteristics of each cluster in the horizontal well volume fracturing and sand body configuration style is established, which provides a theoretical basis for the adjustment of different sand body development methods in subsequent oilfields.

Due to the influence of terrigenous debris, the internal pore structure of continental shale is highly heterogeneous, and the controlling factors are complex. This paper studies the structure and controlling factors of shale reservoirs in the first member of the Qingshankou Formation in the Southern Songliao Basin using core data and various analytical test data. The results show that the original deposition and subsequent diagenesis comprehensively determine the shale reservoirs’ pore structure characteristics and evolution law. According to the severity of terrigenous debris, the shale reservoirs in the study area are divided into four categories and six subcategories of lithofacies. By comparing the characteristics of different shale lithofacies reservoirs, the results show that the lithofacies with a high brittle mineral content have more substantial anti-compaction effects, more primary pores to promote retention and a relatively high proportion of mesopores/macropores. Controlling the organic matter content when forming high-quality reservoirs leads to two possibilities. An excessive organic matter content will fill pores and reduce the pore pressure resistance. A moderate organic matter content will make the inorganic diagenesis and organic hydrocarbon generation processes interact, and the development of organic matter mainly affects the development of dissolution pores. The comprehensive results show that A3 (silty laminated felsic shale) reservoirs underwent the pore evolution process of “two drops and two rises” of compaction, cementation and pore reduction, dissolution and pore increase, and organic matter cracking and pore increase, and they are the most favourable lithofacies of the shale reservoirs in the study area.

To deepen our understanding of reservoir heterogeneity, seepage barriers in the study area were divided into interlayer, intralayer, and planar levels based on their spatial distribution characteristics and then investigated by taking the third member of the Neogene Guantao Formation in the block M of the Gudao oilfield as an example. Based on their genesis characteristics, the interlayer seepage barriers were divided into the overbank sand-overbank sand type, overbank sand-channel type, isolated channel superposition type, and channel shallow-cut type. The planar seepage barriers were categorized into the channel boundary type, abandoned channel type, channel-overbank sand type, and floodplain mudstone type. The intralayer seepage barriers were classified into the mudstone type and physical property type. Classifying multiple levels of different types of seepage barriers led to the refinement of their spatial characteristics. The strength of seepage barriers was characterized using the “seepage barrier coefficient” and “reservoir quality coefficient,” and the small-valued coefficients of the interlayer and intralayer seepage barriers indicated that these seepage barriers are characterized by large thickness, high mudstone content, poor physical properties, and weak seepage ability. The strength of the planar seepage barriers was dependent on the sedimentary facies types and channel stages, and the difference between planar seepage barriers was characterized using the reservoir quality coefficient. The seepage barriers were described based on the multilevel classification of seepage barrier categories, the intralayer seepage barriers with a lateral accretion pattern in the meandering river point bars were identified and described, and the distribution characteristics of seepage barriers were summarized at multiple levels. The method of “hierarchical analysis, skeleton construction, and category fitting” was used to establish 3D models of different levels of seepage barriers, and a modeling method based on the characteristics of intralayer seepage barriers under the constraint of architecture pattern was used to model intralayer seepage barriers with a lateral accretion pattern. In the monitoring of water injection profiles, it was found that the extent of blockage achieved by seepage barriers affects the water injection volume and thus controls the fluid transport pattern. Due to the development characteristics of the seepage barrier in the formation, there are some differences in injection production efficiency under different well pattern matching modes. The research on the distribution characteristics of different levels of seepage barrier categories provides a reliable geological basis for improving the injection-production relationship.