A method for constructing a heterogeneous compact reservoir differential evolution model

Abstract

The application discloses a kind of heterogeneous compact reservoir difference evolution model construction methods, belong to oil and gas geological exploration technical field, this method includes five core steps: first, core, experiment, well logging and seismic multi-source data are standardized and preprocessed;Then quantitative analysis reservoir porosity, permeability temporal and spatial variation and pore throat characteristics;Then dynamic source rock diagenesis process and quantize its intensity;Further, key factors such as sedimentary microfacies, diagenetic transformation intensity, tectonic activity frequency and burial depth are coupled, and the correlation function between them and reservoir densification degree is established;Finally, a visual model is constructed and optimized by validation data and regional geological background factors.The application solves the problems of multi-source data integration, evolution law quantization and poor model applicability, realizes the integrated dynamic characterization from micro-pore throat to macro-reservoir, significantly improves the prediction accuracy of high-quality reservoir, and provides reliable technical support for compact oil and gas reservoir exploration and development.

Description

Technical Field

[0001] This invention relates to the field of oil and gas geological exploration technology, and in particular to a method for constructing a differential evolution model of heterogeneous tight reservoirs based on multi-source data fusion and multi-factor coupling. Background Technology

[0002] Oil and gas exploration and development is increasingly shifting towards unconventional resources, with heterogeneous tight reservoirs emerging as a crucial replacement area. These reservoirs have undergone complex sedimentary, diagenetic, and tectonic alterations, resulting in complex pore structures, poor physical properties, and extremely uneven spatial distribution. Traditional reservoir modeling methods are largely based on the assumption of homogeneity or weak heterogeneity, relying on static physical parameters to characterize the reservoir's current state. However, these static models cannot reveal how the reservoir evolved from its original sedimentary state through a series of geological processes into its current tight configuration. Consequently, they struggle to predict the spatiotemporal distribution patterns of relatively high-quality areas and are no longer sufficient to meet the demands of efficient tight oil and gas exploration and development.

[0003] To reveal the laws governing reservoir evolution, existing technologies have conducted some exploratory studies, but they generally suffer from a lack of systematic approach. Most methods focus only on the influence of a single geological process, such as simulating the decrease in porosity due to compaction or analyzing the blockage of pore throats by cementation, without coupling multiple geological processes such as sedimentation, diagenesis, tectonics, and burial within a unified dynamic framework. This "seeing the trees but not the forest" approach leads to inaccurate judgments of the main controlling factors of reservoir compaction, and the constructed models have low fit with actual geological conditions. Furthermore, these methods often rely on macroscopic data such as well logging and seismic data, or are limited to microscopic analyses such as core and thin section analysis, lacking effective technical means to organically link the evolution of the microscopic pore throat system with the changes in macroscopic reservoir properties, resulting in a disconnect between the "micro" and "macro" scales.

[0004] In summary, current research on heterogeneous tight reservoirs faces three core challenges: First, at the data level, integrating multi-source heterogeneous data such as core, experimental, well logging, and seismic data is difficult, resulting in severe information silos and hindering a comprehensive and systematic understanding of the reservoirs. Second, at the mechanistic level, there is a lack of a dynamic model capable of quantifying the spatiotemporal evolution of reservoirs driven by the coupling of multiple geological factors, leading to a lack of reliable theoretical basis for predicting high-quality reservoirs. Third, at the application level, existing models are mostly constructed for specific blocks, and due to the failure to consider the differences in regional geological backgrounds, the models have poor universality and are difficult to extend to other basins or strata. Therefore, there is an urgent need in this field for a new modeling method that can deeply integrate multi-source data, systematically quantify the dynamic evolution process of reservoirs, and has good generalization capabilities. Summary of the Invention

[0005] The purpose of this invention is to provide a method for constructing a differential evolution model of heterogeneous tight reservoirs based on multi-source data fusion and multi-factor coupling for oil and gas exploration. By constructing a scientific differential evolution model, the spatiotemporal distribution of high-quality reservoirs can be accurately predicted, thereby reducing exploration risks and improving development efficiency.

[0006] The present invention provides a method for constructing a differential evolution model of heterogeneous tight reservoirs, comprising the following steps performed sequentially: S1. Multi-source data preprocessing: Standardization algorithms are used to normalize core observation data, experimental analysis data, well logging data, and seismic data to eliminate differences in data dimensions. The core observation data includes rock type and pore and fracture development morphology. The experimental analysis data includes mineral composition identified under a thin section microscope, pore throat parameters from mercury intrusion porosimetry, clay mineral content from X-ray diffraction, and microscopic pore structure from scanning electron microscopy. The well logging and seismic data include reservoir thickness, burial depth, and structural development characteristics. S2. Quantitative analysis of reservoir characteristics: Based on the preprocessed data, through pore evolution analysis and diagenetic co-occurrence sequence mapping, the spatiotemporal variation of reservoir porosity and permeability and the development characteristics of the pore throat system are quantitatively characterized. S3. Dynamic tracing of diagenesis: Based on quantitative analysis results, trace the diagenetic process, quantify the intensity of key diagenetic processes, and clarify their impact mechanism on reservoir compaction. S4. Coupling of Differential Evolution Laws: Introducing key influencing factors such as sedimentary microfacies, diagenetic alteration intensity, tectonic activity frequency, and burial depth, and establishing correlation functions between each factor and the degree of reservoir compaction; S5. Model Validation and Optimization: Construct a visual model that can dynamically reflect the differences in reservoir evolution during different geological periods, and use validation data and regional geological background correction factors to validate and iteratively optimize the model, ultimately forming a differential evolution model for predicting high-quality reservoirs in tight oil and gas reservoirs and formulating exploration and development plans.

[0007] This invention provides a complete and logically rigorous technical process that systematically solves the problems of isolated data, one-sided mechanism, and single scale in traditional methods. Through the sequential execution of these five steps, it realizes for the first time a closed loop of the entire chain from multi-source data fusion, feature analysis, cause tracing, law coupling to model optimization, ensuring that the final model can dynamically, quantitatively and with high precision reflect the spatiotemporal evolution differences of the reservoir.

[0008] Optimally, the multi-source data preprocessing process is executed by a data preprocessing module, which includes a data input unit, a standardization processing unit, a data quality detection unit, and a data storage unit. The data input unit supports importing data in multiple formats, and the standardization processing unit transforms multi-source data into a standardized dataset through standardization algorithms and quality control measures. This solution constructs an efficient and reliable data processing front-end by defining the data preprocessing module and its internal units. This modular design ensures that multi-source heterogeneous data can be standardized and integrated with high quality, laying a unified and clean data foundation for all subsequent analyses and improving data utilization and the input quality of subsequent models from the source.

[0009] Optimally, the quantitative analysis process of reservoir characteristics is executed by the quantitative analysis module, which consists of a mineral composition analysis unit, a pore throat parameter calculation unit, a pore evolution simulation unit, and a sedimentary facies division unit. The mineral composition analysis unit and the pore throat parameter calculation unit rely on experimental analysis data to quantitatively characterize the reservoir's petrological and pore structure characteristics. The pore evolution simulation unit and the sedimentary facies division unit rely on well logging and seismic data to quantitatively characterize the spatiotemporal variation patterns of reservoir properties and sedimentary characteristics. By defining the quantitative analysis module and its functional units, a comprehensive and refined quantitative description of the reservoir's static characteristics is achieved. It correlates and quantifies macroscopic properties (porosity, permeability) with microscopic structures (pore throats), petrology, and sedimentary facies, providing solid and multi-dimensional data support for revealing evolutionary patterns.

[0010] Optimally, the dynamic diagenetic tracing process is executed by the dynamic diagenetic tracing module, which includes an authigenic mineral phase identification unit, a diagenetic intensity quantification unit, and a diagenetic co-occurrence sequence mapping unit. The authigenic mineral phase identification unit identifies diagenetic event sequences based on microscopic analysis and age data. The diagenetic intensity quantification unit and the diagenetic co-occurrence sequence mapping unit work together to trace the diagenetic process and quantify its impact. This scheme, by limiting the dynamic diagenetic tracing module, transforms qualitative diagenetic analysis into quantitative, dynamic evolutionary process tracing. By identifying diagenetic phases and quantifying intensity, it can accurately reveal the contribution of different diagenetic events to reservoir compaction and profoundly elucidate the intrinsic genetic mechanism of reservoir property evolution.

[0011] Ideally, the differential evolution law coupling process is executed by the differential evolution law coupling module, which includes an influencing factor screening unit, a factor correlation analysis unit, and an evolutionary difference quantification unit. The influencing factor screening unit is used to screen sedimentary microfacies, diagenetic alteration intensity, tectonic activity frequency, and burial depth as core influencing factors from multi-source data. The factor correlation analysis unit establishes the correlation function between factors and reservoir properties through grey relational analysis or algorithm modeling. The evolutionary difference quantification unit is used to clarify the contribution of each factor.

[0012] Ideally, the algorithm is modeled using multivariate regression analysis or a random forest machine learning algorithm; the correlation function is used to dynamically characterize the mathematical relationship between the key influencing factors and reservoir porosity and permeability parameters. This scheme provides two optional paths: multivariate statistics and intelligent algorithms, enhancing the adaptability and robustness of the method under different data conditions. By establishing a dynamic mathematical correlation function, it can more accurately simulate the complex nonlinear relationship between multiple factors and reservoir properties.

[0013] Ideally, the model validation and optimization process is executed by the model construction and optimization module, which consists of a model framework building unit, a parameter assignment unit, a model validation unit, and a model optimization unit. The model framework building unit uses 3D visualization technology to construct a multi-dimensional model framework. The parameter assignment unit substitutes correlation functions and quantitative parameters into the multi-dimensional model framework. The model validation unit calculates the model error using cross-validation algorithms with data from validation wells not involved in the modeling. The model optimization unit iteratively optimizes the model based on the model error and regional geological background correction factors. By limiting the model construction and optimization module, the accuracy and reliability of the final model are ensured. This module, through constructing a 3D dynamic framework, conducting rigorous cross-validation, and introducing regional geological background correction, achieves iterative self-optimization of the model, making its prediction results closer to geological reality and significantly reducing model uncertainty.

[0014] Optimized, the method also includes a result output process executed by a result output module. This module comprises a data report generation unit, a 3D model display unit, and a result export unit, used to generate data reports, display 3D models, and export application results to meet the needs of different scenarios in scientific research analysis and exploration reporting. This solution, by adding a result output module, bridges the "last mile" from model calculation to practical application. It provides diverse result output formats (reports, 3D visualizations, etc.), directly transforming complex model research results into intuitive information usable by geologists and decision-makers, thus enhancing the method's practical value.

[0015] This invention also introduces an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method for constructing a differential evolution model of heterogeneous tight reservoirs as described above. This solution protects an electronic device for executing the method, materializing the abstract method flow into a concrete hardware device, clarifying the feasibility of the method, and providing a device-level basis for patent protection, preventing others from copying this method through hardware devices.

[0016] This invention also introduces a computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the method for constructing the differential evolution model of heterogeneous tight reservoirs as described above. This solution protects a computer-readable storage medium storing relevant programs, protecting the core algorithms and processes through the medium, ensuring the encapsulation and dissemination security of the method technology, and facilitating the promotion, sale, and licensed use of this software-based method.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. In terms of data integration and mechanism revelation, this invention achieves a leap from "isolated multi-source data" to "deep fusion-driven" approaches. By introducing standardized algorithms and multi-source correlation analysis techniques, it effectively solves the problems of dimensional inconsistencies and information gaps in heterogeneous data such as core, experimental, well logging, and seismic data, constructing a high-quality, integrated data analysis foundation and improving data utilization. More importantly, this method, for the first time, systematically incorporates six core geological factors—sedimentary microfacies, diagenesis, tectonic activity, and burial depth—into a unified analytical framework. By constructing multi-factor coupling correlation functions, it fully reveals the contribution and interaction mechanisms of each factor in the reservoir densification process, overcoming the problem of distorted evolutionary patterns caused by the one-sided consideration of single factors in traditional methods.

[0018] 2. In terms of model accuracy and representation dimensions, this invention achieves an upgrade from "static coarse description" to "dynamic precise simulation." It innovatively introduces multi-well cross-validation and regional geological background correction mechanisms. Through iterative optimization, the model's prediction fitting error for key reservoir properties is stabilized within a small range, significantly improving the accuracy of high-quality reservoir prediction and achieving a qualitative leap in reliability. At the same time, this method breaks through the scale barrier of traditional modeling and successfully realizes the integrated dynamic representation of "microscopic pore throat system - macroscopic reservoir distribution." The model can simultaneously output full-scale information from nanoscale pore throat evolution to kilometer-scale reservoir distribution, providing unprecedented insights into the causes of reservoir heterogeneity.

[0019] 3. In terms of methodological universality and application value, this invention expands from a "specific block model" to a "general technical system." By setting configurable regional geological background correction parameters, this model possesses strong adaptability, enabling it to be adapted to basins with different geological evolution histories, such as the Ordos, Songliao, and Sichuan basins, without structural reconfiguration, thus improving the efficiency of technology promotion and its application scope. Ultimately, this method provides a powerful decision-making tool for oil and gas exploration and development, accurately delineating areas of high-quality reservoirs from different geological periods, fundamentally reducing the exploration risks and development costs of tight oil and gas reservoirs, and contributing to the efficient and sustainable development and utilization of unconventional oil and gas resources in my country. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the method for constructing a differential evolution model of heterogeneous tight reservoirs according to the present invention. Figure 2 This is a schematic diagram of the circuit module structure used in this invention.

[0021] The diagram shows the following modules: 1. Data Preprocessing Module; 11. Data Input Unit; 12. Standardization Processing Unit; 13. Data Quality Inspection Unit; 14. Data Storage Unit; 2. Quantitative Analysis of Reservoir Characteristics Module; 21. Mineral Composition Analysis Unit; 22. Pore Throat Parameter Calculation Unit; 23. Pore Evolution Simulation Unit; 24. Sedimentary Facies Classification Unit; 3. Dynamic Diagenetic Tracing Module; 31. Authigenic Mineral Stage Identification Unit; 32. Diagenetic Intensity Quantification Unit; 33. Diagenetic Symbiotic Sequence Mapping Unit; 4. Differential Evolution Law Coupling Module; 41. Influence Factor Screening Unit; 42. Factor Correlation Analysis Unit; 43. Evolutionary Difference Quantification Unit; 5. Model Construction and Optimization Module; 51. Model Framework Construction Unit; 52. Parameter Assignment Unit; 53. Model Validation Unit; 54. Model Optimization Unit; 6. Result Output Module; 61. Data Report Generation Unit; 62. 3D Model Display Unit; 63. Result Export Unit. Detailed Implementation

[0022] The following detailed description, in conjunction with the accompanying drawings and embodiments, illustrates a method for constructing an intelligent dynamic person-event graph based on multi-source data, according to the present invention. Those skilled in the art will understand that the embodiments described are merely illustrative of the invention and not intended to limit it.

[0023] like Figure 1 As shown, this invention relates to a method for constructing a differential evolution model of heterogeneous tight reservoirs, which realizes the entire process from data acquisition to model application through a systematic workflow. The following is a detailed description of each step: S1. Multi-source data preprocessing: This step aims to establish a unified, standardized data foundation. Specific implementation includes: (1) Data acquisition: The system collects core observation data (detailed records of rock type, degree of pore and fracture development and spatial distribution); experimental analysis data (including mineral composition and structural characteristics obtained from thin section identification, pore throat size distribution and connectivity parameters determined by mercury porosimetry, clay mineral types and contents analyzed by X-ray diffraction, and three-dimensional structure of micropores observed by scanning electron microscopy); and well logging and seismic data (accurately extracting macroscopic features such as reservoir thickness, burial depth curves, and structural fracture systems).

[0024] (2) Data standardization: Min-Max normalization and Z-score standardization are used to unify the dimensions of multi-source heterogeneous data, establish a data quality control system, identify and correct outliers, and finally form a standardized dataset suitable for subsequent analysis.

[0025] S2. Quantitative analysis of reservoir characteristics: Based on the preprocessed data, this step involves multi-dimensional quantitative characterization: (1) Porosity evolution analysis: By combining the compaction trend line and the cement content calculation, the porosity evolution trajectory of the reservoir during geological history is reconstructed, and a porosity-depth evolution model is established.

[0026] (2) Establishment of diagenetic symbiotic sequence: By using authigenic mineral symbiotic assemblage, cutting relationship and isotopic dating data, a diagenetic evolution sequence diagram is drawn to clarify the timing of each diagenetic event.

[0027] (3) Quantitative characterization of the pore throat system: Integrate mercury intrusion curves and scanning electron microscopy data to establish a pore throat radius distribution model, quantitatively calculate topological parameters such as pore throat coordination number and tortuosity, and form a quantitative evaluation system for pore throat structure.

[0028] S3. Dynamic Tracing of Diagenesis: This step focuses on revealing the influencing mechanisms of the diagenetic process: (1) Quantification of diagenetic strength: Establish quantitative indicators such as compaction rate, cementation index, and dissolution increment to conduct semi-quantitative-quantitative evaluation of the intensity of key diagenetic processes.

[0029] (2) Diagenesis-burial history coupling: The diagenesis sequence is integrated with the burial history and thermal history models to determine the temperature-pressure window of each diagenesis and establish the genetic relationship between the diagenesis response and the basin dynamic background.

[0030] S4. Coupling of Differential Evolution Laws: This step enables multi-factor collaborative analysis: (1) Screening of influencing factors: Through sensitivity analysis, sedimentary microfacies, diagenetic alteration intensity, tectonic activity frequency, and burial depth were identified as the core control factors.

[0031] (2) Establishment of correlation function: Grey relational analysis and random forest algorithm are used to construct the nonlinear mapping relationship between multiple factors and reservoir properties, and form a quantitative prediction function for reservoir compaction degree.

[0032] (3) Contribution quantification: By ranking the importance of factors, the relative contribution rate of each control factor at different stages of reservoir evolution is clarified.

[0033] S5. Model Validation and Optimization: The final stage ensures the accuracy and usability of the model: (1) Construction of three-dimensional dynamic model: Based on platforms such as Petrel, a four-dimensional geological model containing the time dimension is established, and the aforementioned quantitative relationship is integrated to realize dynamic simulation of the evolution process.

[0034] (2) Verification and optimization: Leave-one-out cross-validation method was adopted to verify the model accuracy using well data that were not involved in the modeling. Background parameters such as regional tectonic settlement rate and heat flow value were introduced to correct the model. The prediction error was stabilized within 8% through iterative calculation.

[0035] (3) Output of results: Generate reservoir evolution data reports, high-quality reservoir distribution prediction maps and three-dimensional dynamic visualization results, which directly support the deployment of exploration wells and optimization of development plans.

[0036] See Figures 1-2 The multi-source data preprocessing process is executed by a specially constructed data preprocessing module 1. As a structured functional system, the data preprocessing module 1 consists of four core sub-modules: a data input unit 11, a standardization processing unit 12, a data quality detection unit 13, and a data storage unit 14. Its core task is to transform the raw, heterogeneous geological data into a clean, unified, high-quality dataset, providing a reliable data foundation for subsequent analysis processes.

[0037] The data input unit 11, serving as the module's data entry point, bears the crucial task of receiving raw data from all sources. It possesses strong multi-format compatibility, seamlessly supporting the import of various professional data formats, including LAS format for well logging data, SEG-Y format for seismic data, and tabular and image formats for experimental analysis data. During the import process, the data input unit 11 automatically parses and extracts key metadata, while simultaneously performing automatic classification and indexing based on data source and type, laying the foundation for subsequent processing.

[0038] The standardization processing unit 12 is the core component of the preprocessing process, specifically responsible for eliminating dimensional differences and systematic biases in the data. The standardization processing unit 12 incorporates multiple standardization algorithms, employing Min-Max normalization or Z-score standardization for continuous data, and one-hot encoding for categorical data. For spatially correlated data, the standardization processing unit 12 performs gridded resampling to ensure all data is unified to the same coordinate system and grid scale. During processing, the standardization processing unit 12 also simultaneously implements quality control, identifying and removing obvious outliers.

[0039] The data quality detection unit 13 performs a comprehensive quality assessment on the standardized data to ensure the reliability of the dataset. The data quality detection unit 13 performs integrity checks to detect missing data and assess missing patterns; performs consistency checks to verify the logical consistency between data from different sources; and uses statistical methods to identify potential outliers to provide a basis for subsequent geological expert analysis.

[0040] The data storage unit 14, serving as the endpoint of the preprocessing process, is responsible for the secure storage and management of the standardized dataset. It employs a structured storage scheme, storing the processed data in a database or specific file format and establishing a comprehensive data indexing system. The data storage unit 14 also supports data version management, records processing logs to ensure data integrity and traceability, and provides standardized data access interfaces for downstream analysis modules, guaranteeing smooth data flow within the system.

[0041] The quantitative analysis process of reservoir characteristics is executed by the reservoir characteristic quantitative analysis module 2. This module is the core analysis layer for model construction, responsible for transforming preprocessed multi-source data into key quantitative parameters characterizing the static features of the reservoir. The reservoir characteristic quantitative analysis module 2 consists of four specialized units: a mineral composition analysis unit 21, a pore throat parameter calculation unit 22, a pore evolution simulation unit 23, and a sedimentary facies division unit 24. These four units systematically and quantitatively characterize the reservoir from four dimensions: rock and mineral composition, pore structure, spatiotemporal evolution, and sedimentary background.

[0042] The mineral composition analysis unit 21 is specifically responsible for the precise quantification of reservoir petrological characteristics. This unit primarily processes experimental analysis data from thin section microscopy and X-ray diffraction. Through intelligent recognition of thin section images and mineral spectrum analysis, the unit can accurately quantify the relative content of framework minerals such as quartz, feldspar, and rock fragments, as well as the type and distribution of clay minerals such as illite and chlorite. Its output includes not only simple mineral percentages but also quantitative statistics on mineral assemblages and diagenetic products (such as siliceous overhangs and iron carbonate cements), providing fundamental data for subsequent analysis of reservoir diagenetic evolution and mechanical properties.

[0043] The pore throat parameter calculation unit 22 focuses on the precise characterization of reservoir pore structure. It deeply integrates mercury intrusion porosimetry (MIP) experimental data and scanning electron microscopy (SEM) images. For MIP data, the unit calculates key parameters such as pore throat radius distribution, displacement pressure, median pressure, and pore throat sorting coefficient using mercury intrusion-extrusion curves. Simultaneously, combined with SEM digital image analysis, the pore throat parameter calculation unit 22 can extract topological parameters such as pore shape factor, pore throat coordination number, and connectivity. By fusing macroscopic MIP curves with microscopic image features, the pore throat parameter calculation unit 22 constructs a quantitative pore throat structure model that comprehensively reflects reservoir permeability and reservoir space.

[0044] The porosity evolution simulation unit 23 is dedicated to reconstructing the dynamic changes of reservoir properties in time and space. It primarily relies on well logging data (such as sonic transit time and density) and seismic attribute volumes. First, using geophysical response equations, the unit inverts the well logging data into a continuous vertical distribution of physical parameters such as porosity and permeability. Then, combining a stratigraphic history model, it simulates and reconstructs the porosity evolution path of the reservoir from deposition to the present through techniques such as compaction trend line fitting and cement content stripping. The final output is a reservoir property evolution model within a spatiotemporal framework, clearly revealing the key periods and intensity of reservoir compaction.

[0045] The sedimentary facies delineation unit 24 aims to clarify the sedimentary background and spatial configuration of the reservoir. Utilizing seismic facies, well logging facies, and core observation data, and through seismic attribute extraction and cluster analysis, the unit first delineates macroscopic facies zone boundaries on a plane. Subsequently, using the morphology, amplitude, and contact relationships of well logging curves, a pattern recognition algorithm is employed to finely delineate sedimentary microfacies vertically, such as braided channels and interdistributary bays. The output of the sedimentary facies delineation unit 24 is a three-dimensional sedimentary facies model of the reservoir, quantitatively characterizing the geometric morphology, spatial distribution, and intrinsic relationship between different facies zones and reservoir properties, providing a sedimentological basis for understanding the causes of reservoir heterogeneity.

[0046] These four units do not operate in isolation, but rather work in close coordination. The mineral composition and pore throat parameter units together define the "material basis" and "structural properties" of the reservoir, while the pore evolution and sedimentary facies division units together reveal the evolution process of these properties in the "time dimension" and the distribution pattern in the "spatial dimension." Together, they provide a complete quantitative input for the next stage of dynamic diagenetic tracing.

[0047] The dynamic diagenetic tracing process is executed by the dynamic diagenetic tracing module 3, which is the core for revealing the reservoir evolution history and the causes of compaction. Its function is to interpret dynamic geological processes from static rock characteristics. The dynamic diagenetic tracing module 3 consists of three logically coherent professional units: authigenic mineral stage identification unit 31, diagenetic intensity quantification unit 32, and diagenetic co-occurrence sequence mapping unit 33. These three units work together to accurately reconstruct the diagenetic environmental changes experienced by the reservoir and quantitatively assess the shaping effect of various diagenetic events on reservoir properties.

[0048] The authigenic mineral epoch identification unit 31 is the starting point for diagenetic history research, undertaking the task of sequencing diagenetic events. The unit comprehensively utilizes microscopic analysis data such as rock thin sections, scanning electron microscopy, and cathodoluminescence, combined with geochronological data such as fluid inclusion homogenization temperature and authigenic mineral isotope dating. Its core workflow is as follows: First, by observing the crystal morphology, occurrence, and intercutting and inclusion relationships of authigenic minerals (such as quartz with large margins, carbonate cements, and clay minerals), the relative chronological order of their formation is determined. Then, fluid inclusion homogenization temperature data is used to calibrate the paleotemperature range of these diagenetic events, or isotope dating techniques (such as illite K-Ar dating) are used to directly obtain the absolute geological age of key diagenetic events. The output of the authigenic mineral epoch identification unit 31 is a preliminary sequence of diagenetic events arranged chronologically and constrained by temperature or age.

[0049] The diagenetic intensity quantification unit 32 is responsible for converting qualitative diagenetic phenomena into quantitative geological parameters, accurately assessing the intensity of reservoir stimulation by various diagenetic processes. The diagenetic intensity quantification unit 32 receives quantitative data from the mineral composition analysis unit 21 and the pore throat parameter calculation unit 22. By establishing a mathematical model, the diagenetic intensity quantification unit 32 calculates the intensity indicators of key diagenetic processes, such as: Compaction strength: The amount of porosity reduction due to compaction is calculated by comparing the original porosity (estimated based on sedimentary models) with the residual porosity under the current rock particle contact relationship (such as line contact, concave-convex contact).

[0050] Cementation strength: The porosity loss caused by cementation is calculated by statistically analyzing the volume content of cement in the thin section and combining it with geochemical data.

[0051] Dissolution intensity: The increase in porosity caused by dissolution is calculated by identifying and quantifying the volume of secondary dissolution pores.

[0052] The output of diagenetic intensity quantification unit 32 is a quantitative assessment of the net change in porosity brought about by each important diagenetic process (compaction, cementation, dissolution, etc.).

[0053] The diagenetic co-existence sequence mapping unit 33 is the module's integration and visualization center, responsible for integrating the results of the first two units into a complete diagenetic evolution map. The diagenetic co-existence sequence mapping unit 33 works closely in collaboration with the diagenetic intensity quantification unit 32. It first uses the diagenetic event sequence established by the authigenic mineral phase identification unit 31 as a time frame, and then accurately assigns the intensity data (porosity increase / decrease) of each phase calculated by the diagenetic intensity quantification unit 32 to the corresponding time nodes. Finally, the diagenetic co-existence sequence mapping unit 33 generates a comprehensive diagenetic evolution model map. This model map uses geological time as the horizontal axis, intuitively displaying the type, sequence, intensity of each diagenetic event and its cumulative effect on reservoir porosity. It can also be coupled with burial history and thermal history curves to clearly indicate the paleotemperature, paleopressure, and other environmental conditions at the time of key diagenetic events.

[0054] Through iterative cycles and mutual verification of these three units, the dynamic diagenetic tracing module 3 ultimately achieved a "recap" of the entire diagenetic life of the reservoir. It not only answered the question of "what the reservoir experienced," but also revealed more precisely "when these experiences occurred" and "how much impact they had," providing crucial process dynamics evidence for establishing differential evolution laws in the next stage.

[0055] The differential evolution law coupling process is executed by differential evolution law coupling module 4. Differential evolution law coupling module 4 is the core decision engine of the entire methodology. Its core task is to reveal the key driving mechanisms controlling reservoir heterogeneity and evolution paths, and to transform geological knowledge into quantifiable predictive models. Differential evolution law coupling module 4 accomplishes this task collaboratively through three sequentially connected and functionally integrated units: influence factor screening unit 41, factor correlation analysis unit 42, and evolutionary difference quantification unit 43.

[0056] The influencing factor screening unit 41 is the foundation and starting point of the regular coupling. Its function is to accurately identify the core control factors that play a dominant role in the differential evolution of reservoirs from massive multi-source geological data. The influencing factor screening unit 41 receives preprocessed and quantitative data from all the aforementioned modules and constructs an initial factor pool that includes multiple dimensions such as sedimentation, diagenesis, structure, and burial. The influencing factor screening unit 41 first conducts preliminary screening through geological mechanism analysis. For example, based on sedimentological principles, it retains "sedimentary microfacies" because they control the original material basis of the reservoir. Subsequently, the unit uses statistical methods such as principal component analysis and Pearson correlation analysis to conduct sensitivity analysis on the remaining factors and key reservoir properties (such as porosity and permeability), calculates the variance contribution rate of each factor and its correlation coefficient with the target variable, thereby objectively eliminating redundant and secondary factors. Through this series of qualitative and quantitative screening processes, the influencing factor screening unit 41 finally selected the most representative set of core influencing factors, which typically include: sedimentary microfacies (controlling the original composition and structure), diagenetic alteration intensity (characterizing the degree of later alteration), tectonic activity frequency (affecting fracture development and fluid migration), and burial depth (comprehensively reflecting the temperature and pressure environment).

[0057] The factor correlation analysis unit 42 follows closely behind, responsible for establishing a precise mathematical mapping relationship between core influencing factors and reservoir properties. Addressing the complexity and nonlinearity of geological systems, the factor correlation analysis unit 42 provides various advanced algorithmic tools. When the relationship between factors and properties is unclear, the unit prioritizes grey relational analysis, calculating the grey relational degree between each factor sequence and the reservoir property sequence to determine the order of their influence. This method has low data sample size requirements and strong adaptability. When sufficient data samples are available, the factor correlation analysis unit 42 activates more powerful algorithmic modeling functions, including multivariate nonlinear regression to fit explicit mathematical equations, or using machine learning algorithms such as random forests and support vector machines to capture the complex interactions between factors and their deep nonlinear correlations with reservoir properties. Ultimately, the output of the factor correlation analysis unit 42 is one or more high-precision correlation functions, whose mathematical form can be an explicit equation or an implicit machine learning model, accurately describing the reservoir property values ​​corresponding to different factor combinations.

[0058] The evolutionary difference quantification unit 43 is the final step in the process, aiming to accurately interpret the geological meaning behind the correlation function, that is, to clarify the specific contribution weight of each core influencing factor in driving the differential evolution of the reservoir. The evolutionary difference quantification unit 43 deeply analyzes the mathematical model established by the factor correlation analysis unit 42. For multiple regression models, the evolutionary difference quantification unit 43 directly compares the contribution of each factor by calculating standardized regression coefficients. For machine learning models such as random forests, the evolutionary difference quantification unit 43 calls the built-in feature importance ranking function, and quantitatively assesses the degree of determination of each factor in the model prediction by calculating indicators such as "Gini impurity reduction" or "ranking importance". The evolutionary difference quantification unit 43 can not only output a global ranking of factor contributions, but also further analyze the dynamic changes in the contribution of each factor in different geological periods or sedimentary facies zones. For example, its analysis results may clearly indicate that in shallow layers, sedimentary microfacies are the main controlling factor of reservoir properties; while with increasing burial depth, the contribution weight of diagenetic alteration intensity increases significantly.

[0059] Through the precise collaboration of these three units, the differential evolution law coupling module 4 successfully transformed qualitative geological concepts into quantitative mathematical relationships, and ultimately clarified the relative importance of various geological processes in reservoir differential evolution, providing the most critical mathematical and geological basis for the final construction of an accurate and well-defined dynamic evolution model.

[0060] In the aforementioned coupling process of differential evolution laws, algorithmic modeling serves as the core technical means, primarily employing two methods: multiple regression analysis and random forest machine learning algorithms. The multiple regression analysis approach establishes explicit mathematical equations to characterize the linear or nonlinear relationships between key influencing factors and reservoir properties, clearly expressing the independent influence trends and directions of each factor. The random forest machine learning algorithm, on the other hand, is suitable for handling more complex nonlinear relationships and interaction effects between factors. This algorithm, through ensemble learning by constructing a large number of decision trees, does not require pre-defined mathematical relationships and can automatically mine deep patterns from the data, possessing stronger pattern recognition capabilities and predictive robustness.

[0061] The final output of the above modeling process is the correlation function, which is the core carrier for dynamically representing the mathematical relationship between multiple factors and reservoir properties. When using multiple regression analysis, the correlation function manifests as a specific regression equation, and the prediction results can be directly calculated by substituting the factor values. When using the random forest algorithm, the correlation function represents the trained ensemble model as a whole, achieving accurate mapping of physical property parameters through complex internal decision-making logic. These two forms of correlation functions together constitute a quantitative expression of the reservoir differential evolution law, providing key mathematical model support for the transition from qualitative analysis to quantitative prediction.

[0062] The model verification and optimization process is executed by the model building and optimization module 5, which is a key step in ensuring the accuracy and practicality of the final model. The model building and optimization module 5 consists of four parts: a model framework building unit 51, a parameter assignment unit 52, a model verification unit 53, and a model optimization unit 54, forming a complete quality control system from framework building to iterative optimization.

[0063] The model framework construction unit 51 is responsible for creating the three-dimensional digital carrier of the model. Employing advanced three-dimensional visualization and geological modeling technologies, the model framework construction unit 51 constructs a multi-dimensional model framework comprising spatial three-dimensionality and temporal one-dimensionality. Based on seismic interpretation layers, fault models, and drilling trajectories, this framework accurately depicts the macroscopic geological framework of the target reservoir and delineates a high-resolution three-dimensional mesh system. Simultaneously, the framework embeds a timeline, enabling dynamic playback and simulation of different geological periods, providing a fundamental platform for showcasing the reservoir's evolutionary history.

[0064] The parameter assignment unit 52 is the process of giving the model geological life. The parameter assignment unit 52 receives correlation functions from the differential evolution law coupling module and various quantitative parameters from the reservoir characteristic quantitative analysis module. Its core task is to dynamically assign these mathematical relationships and parameters to each grid cell in the three-dimensional model framework. Specifically, the parameter assignment unit 52 uses a spatial interpolation algorithm to transform point-based well data into an initial parameter field on the surface, and then drives the correlation function to run at each grid point. For example, for a specific grid, the cell reads its corresponding sedimentary microfacies, diagenetic strength, and other factor values, and calculates the porosity and permeability values ​​of the grid in different geological periods through the established correlation function, thereby realizing the digital reconstruction of the spatiotemporal changes in reservoir properties.

[0065] The model validation unit 53 is responsible for objectively evaluating the preliminary model. The model validation unit 53 employs a rigorous cross-validation algorithm to quantify the model's accuracy. Specifically, it reserves data from several validation wells within the study area that did not participate in the initial modeling process as "blind well" test samples. The model validation unit 53 compares the model's prediction results (such as porosity curves) at these well locations point by point with the actual data obtained from well logging interpretation or core analysis. By calculating indicators such as mean absolute error and root mean square error, it accurately calculates the model's prediction error, thereby objectively evaluating the model's reliability in practical applications.

[0066] The model optimization unit 54 is an intelligent step that self-corrects and improves the model based on the verification results. The model optimization unit 54 receives the error analysis report calculated by the model verification unit. When the error exceeds the preset tolerance, the optimization unit initiates an iterative optimization program. It not only adjusts the key coefficients in the correlation function in reverse but also introduces regional geological background correction factors, such as basin tectonic subsidence rates and regional heat flow values, to perform regional calibration of the model. Through multiple iterative cycles of "verification-correction," the model error is continuously reduced until its prediction accuracy stabilizes within an acceptable threshold, ultimately forming a heterogeneous tight reservoir differential evolution model that highly matches regional geological patterns and provides reliable prediction results.

[0067] In this invention, after model construction and optimization, the method also includes a crucial result output process, which is dedicated to the result output module 6. Its core function is to transform complex model data and research results into user-friendly and diverse result formats suitable for different application scenarios. The result output module 6 includes a data report generation unit 61, a 3D model display unit 62, and a result export unit 63, which together ensure the effective transmission and efficient utilization of research results.

[0068] The data report generation unit 61 is responsible for systematically organizing and automatically outputting the model's core data, key parameters, and statistical analysis results. Based on preset templates or custom requirements, the data report generation unit 61 can quickly generate well-structured and detailed professional reports. These reports include, but are not limited to: statistical tables of physical property parameters for each well point and region (such as maximum, minimum, average, and distribution frequency of porosity and permeability), diagenetic intensity zoning data, planar distribution data of high-quality reservoir thickness, and model accuracy verification reports. The generated reports support multiple formats, such as PDF for formal document distribution and Excel for subsequent in-depth data mining and custom analysis, meeting the needs of researchers for quantitative data analysis and archiving management.

[0069] The 3D model display unit 62 is the core of the overall output visualization, dedicated to transforming abstract model data into intuitive and dynamic 3D images. Based on a powerful 3D graphics engine, the 3D model display unit 62 can perform high-performance rendering and interactive operation of the constructed differential evolution model. Users can not only browse the current physical property distribution of the reservoir from multiple angles (horizontal, cross-sectional, arbitrary cross-section) and dimensions, but also dynamically replay or simulate the porosity evolution, diagenetic changes, and high-quality reservoir development process from deposition to the present through the built-in time slider. The 3D model display unit 62 supports attribute filtering and highlighting functions. For example, it can quickly delineate and highlight "sweet spots" where porosity exceeds a specific threshold, making the heterogeneous characteristics and evolutionary patterns of the reservoir immediately apparent. This is suitable for comprehensive assessment within the team and for reporting exploration decisions to management.

[0070] The output export unit 63 ensures seamless integration of research results with other professional software platforms and workflows. It provides a flexible export interface, supporting the output of the final model and its data in various industry-standard formats. For example, it can export the 3D mesh model and its attribute parameters to the input formats required by numerical simulation software such as CMG and Eclipse, for direct use in development simulations; it can also export to common geological formats such as SEGY and DSG for further use by geological modeling software or other seismic interpretation platforms. Furthermore, the output export unit 63 supports exporting key maps and animated videos as high-resolution images and multimedia files, facilitating the creation of presentations and technical posters. This function fundamentally solves the "last mile" problem of research results moving from research platforms to production applications, meeting the application needs of the entire chain from scientific research to exploration and production.

[0071] This invention also provides an electronic device for implementing the aforementioned method. The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can fully implement the aforementioned method for constructing a differential evolution model of heterogeneous tight reservoirs. In practical applications, this electronic device can be a high-performance workstation or server, where the processor is responsible for executing all core algorithms such as data preprocessing, parameter calculation, and model construction; the memory is used to store massive amounts of multi-source geological data, intermediate processing results, and the final three-dimensional evolution model. This hardware implementation allows this innovative method to move beyond purely theoretical research and become a practically deployable and operational technical system, providing geological researchers with a powerful computational analysis platform.

[0072] Furthermore, this invention also provides a computer-readable storage medium storing a computer program. When this program is executed by a processor, it can also implement the aforementioned method for constructing a differential evolution model of heterogeneous tight reservoirs. This storage medium can be a physical carrier of various forms, such as a solid-state drive, optical disc, or USB flash drive, or it can be virtual storage space in a server. Through this implementation, the method of this invention can be solidified into an independent software product or software module, facilitating copying, distribution, and installation across different computing devices, greatly promoting the widespread application of this technology. Researchers only need to install the program on a computer equipped with the appropriate hardware to obtain professional reservoir differential evolution analysis capabilities, effectively lowering the barrier to entry for the technology.

[0073] In summary, this invention provides a method and related equipment for constructing a differential evolution model of heterogeneous tight reservoirs. This method integrates multi-source geological data to construct a complete technical process from data preprocessing, feature analysis, diagenetic tracing, to regular coupling and model optimization. Its core lies in using standardized algorithms to eliminate data discrepancies, characterizing reservoir features through quantitative analysis, and introducing multi-factor coupling analysis to reveal evolutionary patterns. Finally, a dynamic evolution model is formed through three-dimensional visualization technology and iterative optimization. This method effectively solves the problems of difficult data integration, insufficient quantification of evolutionary patterns, and limited model applicability in traditional research, achieving multi-scale integrated characterization from microscopic pore throats to macroscopic reservoirs. Furthermore, this invention also provides electronic equipment and storage media for implementing this method, enabling the technical solution to be transformed into a practical technical tool, providing reliable technical support for the prediction and exploration and development decisions of high-quality tight oil and gas reservoirs.

[0074] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the technical solutions. Although the applicant has described the present invention in detail with reference to preferred embodiments, those skilled in the art should understand that any modifications or equivalent substitutions made to the technical solutions of the present invention cannot depart from the spirit and scope of the present invention and should be covered within the scope of the claims of the present invention.

Claims

1. A method for constructing a differential evolution model of heterogeneous tight reservoirs, characterized in that, This includes the following steps performed sequentially: S1. Multi-source data preprocessing: Standardization algorithms are used to normalize core observation data, experimental analysis data, well logging data, and seismic data to eliminate differences in data dimensions. The core observation data includes rock type and pore and fracture development morphology. The experimental analysis data includes mineral composition identified under a thin section microscope, pore throat parameters from mercury intrusion porosimetry, clay mineral content from X-ray diffraction, and microscopic pore structure from scanning electron microscopy. The well logging and seismic data include reservoir thickness, burial depth, and structural development characteristics. S2. Quantitative analysis of reservoir characteristics: Based on the preprocessed data, through pore evolution analysis and diagenetic co-occurrence sequence mapping, the spatiotemporal variation of reservoir porosity and permeability and the development characteristics of the pore throat system are quantitatively characterized. S3. Dynamic tracing of diagenesis: Based on quantitative analysis results, trace the diagenetic process, quantify the intensity of key diagenetic processes, and clarify their impact mechanism on reservoir compaction. S4. Coupling of Differential Evolution Laws: Introducing key influencing factors such as sedimentary microfacies, diagenetic alteration intensity, tectonic activity frequency, and burial depth, and establishing correlation functions between each factor and the degree of reservoir compaction; S5. Model Validation and Optimization: Construct a visual model that can dynamically reflect the differences in reservoir evolution during different geological periods, and use validation data and regional geological background correction factors to validate and iteratively optimize the model, ultimately forming a differential evolution model for predicting high-quality reservoirs in tight oil and gas reservoirs and formulating exploration and development plans.

2. The method for constructing a differential evolution model of heterogeneous tight reservoirs according to claim 1, characterized in that, The multi-source data preprocessing process is executed by the data preprocessing module (1), which includes a data input unit (11), a standardization processing unit (12), a data quality detection unit (13), and a data storage unit (14). The data input unit (11) supports the import of multi-format data, and the standardization processing unit (12) transforms multi-source data into a standardized dataset through standardization algorithms and quality control methods.

3. The method for constructing a differential evolution model of heterogeneous tight reservoirs according to claim 1, characterized in that, The quantitative analysis process of reservoir characteristics is executed by the quantitative analysis module (2), which consists of a mineral composition analysis unit (21), a pore throat parameter calculation unit (22), a pore evolution simulation unit (23), and a sedimentary facies division unit (24). The mineral composition analysis unit (21) and the pore throat parameter calculation unit (22) rely on experimental analysis data to quantitatively characterize the petrological characteristics and pore structure characteristics of the reservoir. The pore evolution simulation unit (23) and the sedimentary facies division unit (24) rely on well logging and seismic data to quantitatively characterize the spatiotemporal variation law of reservoir properties and sedimentary characteristics.

4. The method for constructing a differential evolution model of heterogeneous tight reservoirs according to claim 1, characterized in that, The dynamic diagenetic tracing process is executed by the dynamic diagenetic tracing module (3), which includes an authigenic mineral phase identification unit (31), a diagenetic intensity quantification unit (32), and a diagenetic co-occurrence sequence drawing unit (33). The authigenic mineral phase identification unit (31) identifies diagenetic event sequences based on microscopic analysis and age data. The diagenetic intensity quantification unit (32) and the diagenetic co-occurrence sequence drawing unit (33) work together to trace the diagenetic process and quantify its impact.

5. The method for constructing a differential evolution model of heterogeneous tight reservoirs according to claim 1, characterized in that, The differential evolution law coupling process is executed by the differential evolution law coupling module (4), which includes an influencing factor screening unit (41), a factor correlation analysis unit (42), and an evolution difference quantification unit (43). The influencing factor screening unit (41) is used to screen sedimentary microfacies, diagenetic alteration intensity, tectonic activity frequency, and burial depth from multi-source data as core influencing factors. The factor correlation analysis unit (42) establishes the correlation function between factors and reservoir properties through grey relational analysis or algorithm modeling. The evolution difference quantification unit (43) is used to clarify the contribution of each factor.

6. The method for constructing a differential evolution model of heterogeneous tight reservoirs according to claim 5, characterized in that, The algorithm is modeled using multiple regression analysis or random forest machine learning algorithm; the correlation function is a function used to dynamically characterize the mathematical relationship between the key influencing factors and reservoir porosity and permeability parameters.

7. The method for constructing a differential evolution model of heterogeneous tight reservoirs according to claim 1, characterized in that, The model verification and optimization process is executed by the model construction and optimization module (5), which consists of a model framework building unit (51), a parameter assignment unit (52), a model verification unit (53), and a model optimization unit (54). The model framework building unit (51) uses three-dimensional visualization technology to build a multi-dimensional model framework. The parameter assignment unit (52) substitutes the correlation function and quantitative parameters into the multi-dimensional model framework. The model verification unit (53) uses cross-validation algorithm to calculate the model error using verification well data that did not participate in the modeling. The model optimization unit (54) iteratively optimizes the model based on the model error and the regional geological background correction factor.

8. The method for constructing a differential evolution model of heterogeneous tight reservoirs according to claim 1, characterized in that, The method also includes a result output process, which is executed by the result output module (6). The result output module (6) includes a data report generation unit (61), a three-dimensional model display unit (62), and a result export unit (63), which are used to generate data reports, display three-dimensional models, and export application results to meet the needs of different scenarios of scientific research analysis and exploration reporting.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for constructing a differential evolution model of heterogeneous tight reservoirs as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for constructing a differential evolution model of heterogeneous tight reservoirs as described in any one of claims 1 to 8.

Images