Lithology identification system and identification method thereof

Abstract

The invention provides a lithology identification system and an identification method thereof. The system comprises a data preprocessing core module, a feature extraction core module and a lithology identification core module. In the data preprocessing stage, the system carries out standardization processing and core depth homing on logging data, and data quality and consistency are ensured. And the feature extraction module adopts a multi-parameter cross plot and three-dimensional cross plot technology and combines a factor analysis method to optimize logging parameter selection and improve the effectiveness of lithology identification features. And the lithology identification module comprehensively uses a density clustering neighbor method and a logging curve rapid identification and multi-parameter fine identification technology to realize accurate identification of complex lithology. According to the method, multiple technical means are creatively fused, the lithology recognition accuracy is improved to 90% or above, the method is particularly suitable for stratums with high heterogeneity, and powerful technical support is provided for exploration and development of unconventional oil and gas such as shale gas.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of address exploration, and in particular to a lithology identification system and an identification method thereof. BACKGROUND

[0002] In the field of oil and gas exploration and development, lithology identification is a core link for evaluating reservoir characteristics, predicting oil and gas distribution, and formulating development plans. However, existing lithology identification technologies face multiple technical bottlenecks, especially under complex geological conditions such as continental shale gas, the limitations of traditional methods become increasingly prominent.

[0003] Disadvantages of the prior art: Low identification accuracy of single parameter: Traditional methods mainly rely on a single logging curve (such as natural gamma, acoustic time difference) for lithology identification, but in continental shale gas layers with interbedded sand and mud and rapid lithology changes, a single parameter cannot distinguish complex lithology combinations such as fine sandstone, argillaceous sandstone, and sandy mudstone, and the identification coincidence rate is generally less than 65%.

[0004] Multiple solutions problem is prominent: Logging curves have overlapping responses to lithology, for example, high natural gamma values may correspond to mudstone or radioactive sandstone, leading to multiple solutions in traditional crossplot methods and increasing exploration risks.

[0005] Lack of three-dimensional spatial representation: Existing technologies can only provide one-dimensional wellbore lithology profiles and cannot achieve quantitative representation of lithology in three-dimensional space, making it difficult to meet modern development needs such as horizontal well trajectory optimization and fracturing segment selection.

[0006] Insufficient intelligence: Traditional lithology identification relies on manual crossplot plotting and experience-based identification, with low automation and processing efficiency that cannot meet the needs of massive logging data interpretation.

[0007] To solve the above problems, the applicant proposes a lithology identification system and an identification method thereof. SUMMARY

[0008] The purpose of the present application is to provide a lithology identification system and an identification method thereof to solve the problems in the prior art.

[0009] To achieve the above purpose, the present application provides the following technical solutions: a lithology identification system, comprising: A data acquisition module for acquiring logging data, including natural gamma (GR), acoustic time difference (AC), density (DEN), compensated neutron (CNL), and resistivity (RT) curves; A data preprocessing module including a standardization processing unit and a core homing unit for eliminating instrument errors and depth deviations; The lithology identification model library comprises a density clustering near neighbor method model, a multi-parameter crossplot model, a three-dimensional lithology identification template, a fractal dimension prediction model and a principal factor analysis model. The intelligent matching engine automatically selects or combines the lithology identification model according to the formation characteristics. The result output module generates a lithology profile, a lithology distribution three-dimensional model and a lithology parameter table.

[0010] Optionally, the standardization processing unit of the data preprocessing module adopts a dynamic frequency crossplot technique, and establishes a full-oilfield logging curve calibration standard through key well standard layer data, and the error correction precision reaches ± 2%; the core homing unit combines a gamma curve and an acoustic travel time curve, and realizes sub-meter matching of logging depth and core depth through a particle swarm optimization algorithm.

[0011] Optionally, the multi-parameter crossplot model in the lithology identification model library comprises six groups of crossplot charts including GR-DEN, AC-CNL and RT-TH, a dynamic threshold interval is set for each group of charts, and a support vector machine (SVM) algorithm is used to optimize the lithology classification boundary.

[0012] Optionally, the three-dimensional lithology identification template is constructed based on well-seismic joint data, a sequential Gaussian simulation algorithm is used to realize lithology body spatial distribution prediction, the grid resolution reaches 0.5 m x 0.5 m x 0.1 m, and the lithology prediction coincidence rate is greater than or equal to 88%.

[0013] Optionally, the fractal dimension prediction model establishes a quantitative relationship between fractal dimension D and lamella cumulative thickness H through analysis of the frequency distribution of silty lamella thickness: H = 15.2 x e^(0.8D), and the correlation coefficient R² is 0.91.

[0014] Optionally, the intelligent matching engine automatically selects the identification model according to a formation heterogeneity index I: When I < 0.3, the density clustering near neighbor method model is enabled; When 0.3 ≤ I < 0.6, the multi-parameter crossplot model is enabled; When I ≥ 0.6, the three-dimensional lithology identification template and the fractal dimension prediction model are used for joint identification.

[0015] A lithology identification method comprises the following steps: Step one: logging curves are acquired through a data acquisition module, and instrument errors and depth deviations are eliminated through a data preprocessing module; Step two: an intelligent matching engine selects a lithology identification model according to a formation heterogeneity index; Step three: a selected model is applied to perform lithology identification, and a lithology profile and a three-dimensional distribution model are generated. Step four: the result output module generates a lithology parameter table containing porosity, permeability and TOC content correlation data.

[0016] Beneficial effects: multi-parameter collaborative identification mechanism: fusion of 12 types of logging parameters such as natural gamma, acoustic time difference, density, neutron porosity, construction of lithology sensitive parameter set, optimization of classification boundary through support vector machine algorithm, breakthrough of single parameter identification limitation.

[0017] Three-dimensional lithology identification template: based on well-seismic joint data, a three-dimensional lithology template is constructed, and a sequential Gaussian simulation algorithm is used to realize the spatial distribution prediction of lithology body, and the grid resolution is 0.5m*0.5m*0.1m, and the lithology prediction coincidence rate is greater than or equal to 88%.

[0018] Fractal dimension prediction model: for the micro lithology body such as silty lamina, the quantitative relationship model (H=15.2* e^(0.8D)) between fractal dimension (D) and lamina thickness (H) is innovated, and the sub-centimeter level lithology boundary is accurately identified.

[0019] Intelligent matching engine: develop the formation heterogeneity index (I) evaluation algorithm, automatically select the density clustering, cross plot method or three-dimensional template identification mode according to the I value, and realize the adaptive lithology interpretation under complex geological conditions. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 It is the data acquisition module schematic diagram of the embodiment of the present application.

[0021] Figure 2 It is the data preprocessing module schematic diagram of the embodiment of the present application.

[0022] Figure 3 It is the lithology identification module schematic diagram of the embodiment of the present application.

[0023] Figure 4 It is the result output module schematic diagram of the embodiment of the present application. DETAILED DESCRIPTION

[0024] The preferred embodiments of the present application are described below with reference to the accompanying drawings, so that the technical content of the present application is more clear and convenient to understand. The present application can be embodied in many different forms, and the protection scope of the present application is not limited to the embodiments mentioned in the text.

[0025] In the field of oil and gas exploration and development, lithology identification is the core link of evaluating reservoir characteristics and predicting oil and gas distribution. The traditional method is limited by the problems of insufficient single parameter analysis accuracy, poor adaptability to complex geological conditions, etc., and it is difficult to meet the exploration needs of unconventional oil and gas reservoirs such as continental shale gas. The present application proposes a lithology identification system and its identification method, which realizes the accurate identification and quantitative evaluation of complex lithology through multi-source data fusion, intelligent algorithm optimization and three-dimensional visualization characterization.

[0026] The system adopts a modular architecture design, including four core modules of data acquisition, preprocessing, lithology identification, and result output. The data acquisition module is compatible with 12 types of logging curves such as natural gamma, acoustic time difference, density, neutron porosity, and resistivity, supports parsing of multiple data formats such as LIS, LAS, and DLIS, and has a sampling interval of up to 0.125 m, ensuring the integrity and accuracy of the original data. The preprocessing module integrates three functions of data cleaning, standardization processing, and core orientation: the data cleaning uses wavelet transform denoising algorithm to effectively filter out high-frequency noise; the standardization processing introduces dynamic frequency crossplot technology, takes the standard layer of the key well as the reference, realizes the scaling unification of logging data in the whole oilfield through nonlinear mapping, and the error correction accuracy reaches ±2%; the core orientation innovates particle swarm optimization algorithm, combines the characteristics of gamma curve and acoustic time difference curve, realizes the sub-meter matching of logging depth and core depth, and the correlation coefficient of core orientation is improved from 0.72 to 0.91.

[0027] A lithology identification method, comprising the following steps: Step one: obtain the logging curve through the data acquisition module, and eliminate the instrument error and depth deviation through the data preprocessing module; Step two: the intelligent matching engine selects the lithology identification model according to the formation heterogeneity index; Step three: apply the selected model to identify the lithology, and generate the lithology profile and three-dimensional distribution model; Step four: the result output module generates the lithology parameter table, including the porosity, permeability, and TOC content correlation data.

[0028] In step two, the formation heterogeneity index (I) is calculated by the following formula: I = (σ_GR² + σ_AC² + σ_DEN²) / (μ_GR² + μ_AC² + μ_DEN²) Wherein σ is the standard deviation of logging parameters, and μ is the mean value of logging parameters.

[0029] In step three, the three-dimensional lithology identification template adopts phase control modeling technology, combines with seismic attribute to constrain the lithology boundary, and optimizes the geometric shape of the lithology body through Markov Chain Monte Carlo (MCMC) algorithm.

[0030] In step four, the lithology parameter table correlates the lithology and physical parameters through Bayesian network, establishes the porosity (φ)-lithology index (LI) relationship model: φ = 0.12 × LI^1.05 - 0.8, and the determination coefficient R² = 0.89.

[0031] The lithology identification module builds a four-level progressive identification system. The first-level identification uses the density clustering nearest neighbor method to automatically divide lithology clusters by calculating the accessibility of the density of logging data points, which is suitable for stable sedimentary strata with interbedded sand and mud. For complex lithology combinations, the multi-parameter crossplot technology is developed, integrating six sets of crossplot charts such as natural gamma-ray-density, acoustic travel time-neutron porosity, and combining support vector machine algorithm to optimize the classification boundary, breaking through the limitations of single parameter identification. At the three-dimensional lithology identification level, the sequential Gaussian simulation algorithm is introduced to fuse well and seismic data to build a three-dimensional lithology template with a grid resolution of 0.5m x 0.5m x 0.1m, realizing the quantitative prediction of lithology spatial distribution. For micro-lithology such as silt laminae, the fractal dimension prediction model is innovated, which establishes the quantitative relationship between fractal dimension and cumulative thickness by analyzing the frequency distribution of lamina thickness, with a prediction error of less than 15%.

[0032] The system is equipped with an intelligent matching engine that automatically selects the optimal identification strategy based on the formation heterogeneity index. This index quantifies the complexity of the formation by calculating the ratio of the standard deviation to the mean of the logging parameters: when the index is less than 0.3, the density clustering method is used; when it is between 0.3 and 0.6, the multi-parameter crossplot is enabled; and when it is greater than 0.6, the three-dimensional lithology template and fractal dimension prediction model are activated for joint identification. This mechanism enables the system to automatically adapt the identification method under different geological conditions, with a lithology identification compliance rate improved from 65% to 92%.

[0033] In the actual application in the Yan'an area of the Ordos Basin, the system carries out fine processing for the shale gas layer of the Chang 7 member. In the stable sedimentary area of the Chang 71 member, the density clustering method accurately identifies thick sandstone and mudstone, with a lithology profile and core observation coincidence degree of 90%. In the transitional facies belt of the Chang 72 member, the multi-parameter crossplot technology effectively distinguishes between argillaceous sandstone and sandy mudstone, with a porosity prediction error controlled within ±1.5%. In the complex facies transition area of the Chang 73 member, the three-dimensional lithology template combined with fractal dimension prediction successfully identifies 2mm-level silt laminae, with a cumulative thickness prediction error of less than 10%. The three-dimensional lithology distribution model output by the system directly guides the optimization of horizontal well trajectory, increasing the drilling rate from 75% to 92% and improving the accuracy of fracture segment selection by 30%.

[0034] The technical solution innovation points are embodied in four aspects: the four-level progressive lithology identification system is created for the first time, realizing quantitative characterization of lithology from single well profile to three-dimensional space; the intelligent matching engine is developed to establish the quantitative correspondence between formation complexity and identification method; the fractal dimension prediction model is proposed to break through the bottleneck of micro-lithology body identification accuracy; and the lithology-property-gas-bearing property comprehensive evaluation platform is constructed to correlate the lithology identification results with parameters such as porosity, permeability, TOC, etc., forming the shale gas reservoir comprehensive evaluation capability. In the field test in Ordos Basin, the single well logging interpretation time is shortened by 60%, the lithology identification coincidence rate reaches 92%, effectively supporting the efficient use of shale gas reserves, and having significant economic value and social significance.

[0035] Embodiment one In the shale gas exploration of Chang 7 member of Yanchang Formation in Yan'an area of Ordos Basin, aiming at the characteristics of interbedded sand and mud development and fast vertical and horizontal changes of lithology, the lithology identification system constructed by the present application is used for fine evaluation. Eight types of lithology such as fine sandstone, argillaceous sandstone, sandy mudstone and carbonaceous shale are developed in this member, and the identification coincidence rate of the traditional single parameter method is less than 65%. The system first deploys a high-precision logging array to obtain 12 logging curves such as natural gamma, acoustic time difference, density, neutron porosity and resistivity, with a sampling interval of 0.125 meters. Through dynamic frequency crossplot method, data standardization processing is carried out, the 15-meter-thick mudstone section of Yan 10 well area is selected as the standard layer, the natural gamma and density, the acoustic time difference and the neutron porosity crossplot are constructed, and the parameter distribution range of the standard layer is determined. The least square method is applied to calibrate 38 wells in the whole area to eliminate instrument errors. Then the particle swarm optimization algorithm is developed to realize core depth homing, 32-meter full-diameter core data and corresponding logging curves of Yan 8 well are input, matching parameter range is set, and after 15 iterations, the optimal matching is realized, so that the correlation coefficient of core and logging curve is improved from 0.72 to 0.91. In the lithology identification link, a three-dimensional lithology identification template is constructed, density clustering, decision tree and neural network algorithms are integrated, logging parameters and derived parameters are input, and model parameters are optimized through cross-validation. The hierarchical identification strategy is implemented, first to distinguish three major categories of sandstone, mudstone and siltstone, and then to subdivide six subcategories such as siliceous sandstone and siliceous mudstone. Finally, three-dimensional geological model is constructed by integrating well and seismic data, sequential Gaussian simulation algorithm is used to realize lithology spatial distribution prediction, lithology probability body data is generated, and interactive visualization platform is used to support dynamic browsing and attribute query of lithology profile. This embodiment improves the lithology identification coincidence rate from 65% to 91%, shortens the single well processing time by 70%, discovers thick sandstone reservoir in Yan 6 well area, obtains high-yield gas flow in gas testing, and provides accurate geosteering for horizontal well trajectory optimization.

[0036] Embodiment two In the exploration of marine shale in Wufeng-Longmaxi Formation in Jiaoshiba area of Sichuan Basin, silt laminae developed in organic-rich shale are identified. The thickness of the laminae is 0.2-2 mm, which controls the permeability of reservoir and the effect of fracturing reconstruction. The error of traditional imaging logging interpretation is more than 30%. Firstly, the micro-resistivity scanning imaging logging is deployed to obtain 360-degree borehole wall images with a resolution of 0.2 inch. Image enhancement algorithm is developed to remove noise spots by median filtering, to enhance contrast by histogram equalization, and to realize lamina boundary sharpening by wavelet transform. Then, the quantitative relationship between lamina thickness and fractal dimension is established. The frequency distribution of lamina thickness is obtained by statistical observation of 200 meters of core in 5 wells, and the fractal dimension is calculated to establish the prediction model and determine the model parameters by cross-validation. In the recognition of multi-source data fusion, the joint recognition workflow is constructed to extract lamina dip angle and strike from imaging logging, to calculate organic carbon content from conventional logging, and to constrain lamina spatial distribution combined with seismic attributes. Random forest classifier is developed to input 8 characteristic variables such as lamina density, fractal dimension and organic carbon content, to determine key parameters by characteristic importance sorting, and to realize automatic classification of lamina type. This example improves the identification accuracy of silt laminae from 62% to 88%, reduces the thickness prediction error from ±0.5 mm to ±0.2 mm, reveals the directional arrangement characteristics of lamina, provides a basis for the optimization of horizontal well section, and improves the activation rate of natural fractures by 40% in the volume fracturing reconstruction, and improves the ultimate recovery of single well by 25%.

[0037] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be realized in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present application is defined by the appended claims rather than the above description, and all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any reference signs in the claims should not be regarded as limiting the claims involved.

[0038] In addition, it should be understood that although the present specification is described in terms of embodiments, not every embodiment contains only one independent technical solution, and the description of the specification is only for the sake of clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that those skilled in the art can understand.

Claims

1. A lithology identification system characterized by, It comprises: a data acquisition module for obtaining logging data, including natural gamma, acoustic time difference, density, compensated neutron and resistivity curves; a data preprocessing module including a standardization processing unit and a core homing unit for eliminating instrument errors and depth deviations; a lithology identification model library integrating density clustering nearest neighbor method model, multi-parameter crossplot model, three-dimensional lithology identification template, fractal dimension prediction model and principal factor analysis model; an intelligent matching engine for automatically selecting or combining lithology identification models according to the characteristics of the formation; a result output module for generating lithology profile, lithology distribution three-dimensional model and lithology parameter table.

2. The lithology identification system according to claim 1, wherein: the standardization processing unit of the data preprocessing module adopts dynamic frequency crossplot technology to establish a full-oilfield logging curve calibration standard through key well standard layer data, and the error correction accuracy reaches ±2%; the core homing unit combines gamma curve and acoustic time difference curve to realize sub-meter matching of logging depth and core depth through particle swarm optimization algorithm.

3. The lithology identification system according to claim 1, wherein: the multi-parameter crossplot model in the lithology identification model library includes six groups of crossplot charts such as GR-DEN, AC-CNL and RT-TH, and each group of chart sets dynamic threshold interval to optimize the lithology classification boundary through support vector machine algorithm.

4. The lithology identification system according to claim 1, wherein: the three-dimensional lithology identification template is constructed based on well-seismic joint data, and adopts sequential Gaussian simulation algorithm to realize lithology body spatial distribution prediction, with grid resolution reaching 0.5m×0.5m×0.1m, and lithology prediction coincidence rate ≥88%.

5. The lithology identification system according to claim 1, wherein: the fractal dimension prediction model establishes the quantitative relationship between fractal dimension D and lamina cumulative thickness H through analyzing the frequency distribution of silt lamina thickness: H=15.2×e^(0.8D), and the correlation coefficient R²=0.

91.

6. The lithology identification system according to claim 1, wherein: the intelligent matching engine automatically selects the identification model according to the formation heterogeneity index I: when I<0.3, the density clustering nearest neighbor method model is enabled; when 0.3≤I<0.6, the multi-parameter crossplot model is enabled; when I≥0.6, the three-dimensional lithology identification template and fractal dimension prediction model are used for joint identification.

7. A lithology identification method characterized by, It comprises the following steps: Step one: obtain logging curves through the data acquisition module, and eliminate instrument errors and depth deviations through the data preprocessing module; Step two: select the lithology identification model according to the formation heterogeneity index through the intelligent matching engine; Step three: apply the selected model to identify lithology, and generate lithology profile and three-dimensional distribution model; Step four: the result output module generates lithology parameter table, including porosity, permeability and TOC content correlation data.

Images