Method for predicting marine facies shale chemical facies based on element logging curve

Abstract

The invention discloses a method for predicting marine facies shale chemical facies based on an element logging curve. The method comprises the following steps: firstly, carrying out mathematical statistical analysis by utilizing SPSS software, screening out an element logging curve with characteristics, and summarizing a plurality of element combinations with paleoenvironmental significance; normalization processing and principal component analysis are carried out on screened data, principal components are extracted to carry out dimension reduction processing on the data, shale chemical phase types are divided by utilizing system clustering analysis and k-means clustering analysis on the basis of principal component scores, and the contribution degree of each variable to clustering is determined through single-factor variance analysis. Finally, rock core samples are collected for petrology observation, and the accuracy of the shale chemical phase division result is verified. According to the method, the marine facies shale chemical facies can be quickly and accurately identified under the condition of lack of rock core samples.

Description

Technical Field

[0001] The present invention relates to the technical field of well logging data processing, and in particular to a method for predicting the chemical phase of marine shale based on element logging curves. Background Art

[0002] Mudstone is widely distributed across the globe, accounting for 70% or even 80% of all sedimentary rocks. It is not only the most important source rock for conventional oil and gas, but also a crucial reservoir rock for unconventional oil and gas (such as shale oil and shale gas), with enormous oil and gas resources. In recent years, my country has made a series of major breakthroughs in the exploration and development of shale oil and gas, and shale oil and gas will also be one of the key areas for increasing reserves and production in my country. Mudstone chemofacies refer to sedimentary units with specific geochemical characteristics, delineated by the chemical composition of mudstone (such as major elements, trace elements, isotopes, and organic matter). These geochemical characteristics are closely related to the sedimentary environment at the time of mudstone formation (such as redox conditions, paleoproductivity, and terrigenous input conditions). They can be used to reconstruct paleoenvironments, predict source rock quality, or assess the exploration and development potential of mudstone reservoirs. Therefore, the study and classification of shale chemical phase characteristics are of great significance in the evaluation of source rocks and comprehensive evaluation of shale oil and gas reservoirs, and can guide the next exploration and deployment of conventional and unconventional oil and gas.

[0003] Traditional methods for chemical phase division of mudstones are mainly based on the correlation between geochemical parameters and sedimentary environments, and are comprehensively judged through indicators such as element ratios, organic geochemical characteristics, and isotopic composition. They mainly include the following division methods: (1) Organic geochemical index method. The potential phases of mudstone source rocks are divided by indicators such as organic matter abundance, organic matter type, and organic matter maturity. For example, mudstones can be divided into organic-rich anoxic phases (high TOC and high HI values) and organic-poor oxidized phases (low TOC and low HI values) using total organic carbon (TOC) content and hydrogen index (HI); (2) Isotope geochemistry method. Carbon and sulfur isotopes can reflect the biogeochemical cycle process and divide mudstones into strong sulfate reduction phases (extremely low δ34S values) and high productivity phases (low δ13Corg values); (3) Major element combination analysis method. By using the content, ratio or triangular relationship (such as Si-Al-Ca) of major elements (such as Si, Al, Ca, Mg, K, etc.), the provenance, sedimentary environment and diagenesis of mud shale can be inferred, and the terrigenous clastic phase (high Al and K content), siliceous biofacies (high Si content) and carbonate mixed phase (high Ca content) can be divided. (4) Trace element combination analysis method. By using the ratio of some special trace elements (such as redox sensitive metal elements, nutritional trace elements, etc.), the paleoenvironment and paleoproductivity conditions during the formation of mud shale can be characterized, thereby dividing it into anoxic or sulfidic phase (high Mo concentration, high U / Th ratio, etc.), high productivity phase (high P, Ba content, etc.), high salinity marine phase (such as high Sr / Ba and B / Ga ratio), etc.; (5) Statistical analysis method. By using statistical methods to analyze major elements, trace elements and other organic geochemical parameters, key geochemical parameters can be extracted and different mud shale chemical phase combinations can be divided, such as high productivity anoxic phase, low productivity-oxidized phase, etc. However, these methods require drilling coring and laboratory testing of organic geochemical composition, isotopic composition, and elemental geochemical composition to perform chemical phase classification. These methods are limited not only by the core intervals of drilling and coring, but also by the long time and high cost of laboratory testing. This makes it impossible to quickly and comprehensively classify shale chemical phases and conduct comprehensive evaluation of source rocks (or unconventional reservoirs) across the entire wellbore, thus hindering further oil and gas exploration and deployment. Therefore, establishing a correlation between elemental logging curves and shale chemical phases can help quickly and accurately identify shale chemical phase types, thereby providing guidance for the next step of oil and gas exploration and deployment. Summary of the Invention

[0004] Based on this, the embodiment of the present application provides a method for predicting the chemical phase of marine shale based on element logging curves, which can overcome the problems existing in the existing technology and can be widely used in the intelligent identification of the chemical phase of marine shale and the comprehensive evaluation research of shale reservoirs.

[0005] In a first aspect, a method for predicting the chemical phase of marine shale based on element logging curves is provided, the method comprising:

[0006] S1 Data Collection: Collect elemental logging data of marine shale formations within the preset depth range of the drilling profile in the study area, and obtain multiple data sets consisting of depth values ​​and initial elemental logging data. The elemental logging data includes major elements, trace elements, and some rare earth elements.

[0007] S2 Data Screening and Characteristic Analysis: SPSS software was used to perform mathematical statistical analysis on the initial element logging curves to obtain the minimum, maximum, median, and standard deviation of each set of element logging curve data. Characteristic element logging curves were screened based on the intersection of the median and standard deviation values. Finally, characteristic analysis was performed on the screened element logging curves to summarize a variety of element combinations with paleoenvironmental significance.

[0008] S3 data dimensionality reduction: The normalization method was used to process the selected element logging curve data to eliminate the influence of different dimensions. SPSS software was used to perform principal component analysis on the normalized element logging curve data to extract the first n principal components. Based on the extracted principal components and their factor loadings, the paleoenvironmental significance represented by different principal components was analyzed.

[0009] S4 Shale chemical facies classification: Based on the extracted first n principal component scores, a systematic cluster analysis was performed using SPSS software. The number of cluster categories k was determined using the elbow rule. Subsequently, a k-means cluster analysis was performed on the principal component scores to obtain the locations of k cluster centers, their distances from the cluster centers, and the classification results. The contribution of different variables to the clustering was determined using a one-way analysis of variance. Finally, k shale chemical facies types were defined based on the paleoenvironmental significance represented by the classification groups and principal components.

[0010] S5 accuracy verification: The core sample and boundary sample data of k clusters are identified based on the distance from the cluster center. The core samples with the closest depth to the core samples and boundary samples are collected. The samples are prepared into rock thin sections and scanning electron microscope optical slices. Petrological observations are carried out using a polarizing microscope and a field emission scanning electron microscope. The accuracy of the shale chemical phase division results is verified based on the mineral distribution and characteristic mineral assemblage characteristics.

[0011] Optionally, in S1, the collected element logging curve data includes at least Si, Al, K, Na, Ca, Mg, Fe, Ti, V, Cr, Mn, P, S, Ni, Cu, Zn, As, Se, Rb, Sr, Y, Zr, Nb, Mo, and Ag elements, where the unit of the main elements is %, and the unit of the remaining elements is ppm; the main elements are specifically Si, Al, K, Na, Ca, Mg, and Fe.

[0012] Optionally, in S2, when performing mathematical statistical analysis on the initial element logging curves using SPSS software, the minimum value, maximum value, median value and standard deviation of each set of element logging curve data are calculated, and characteristic element logging curves are screened out based on the threshold values ​​of median value >5ppm and SD>1.

[0013] Optionally, in S2, the element combinations summarized with paleoenvironmental significance include: terrigenous debris input element combinations, including Ti and Cr; mineral composition sensitive element combinations, including Al, Si, Ca and Mg; paleoproductivity element combinations, including P; paleo-sedimentary environment element combinations, including Fe, S, V and Mo; and special lithologic element combinations, including Mn.

[0014] Optionally, in S3, the specific calculation formula for normalization processing is:

[0015]

[0016] Among them, X i is the normalized data; x i is the data to be normalized; x min is the minimum value in the data set to be normalized; x max is the maximum value in the dataset to be normalized.

[0017] Optionally, in S3, during principal component analysis, the first n principal components are extracted according to the conditions that the eigenvalue is greater than 1 and the cumulative contribution rate is greater than 70%.

[0018] Optionally, in S4, a descending broken line graph of the combined clustering coefficient is drawn by the elbow rule in the system cluster analysis, and the number k of cluster categories is determined according to the degree of reduction of the coefficient.

[0019] Optionally, in S4, the specific steps of the k-means cluster analysis include:

[0020] Randomly initialize k cluster centers, calculate the Euclidean distance of each data point to the cluster center and assign the data point to the nearest cluster, recalculate the mean of all data points in the cluster as the new cluster center, and repeat the process until the change in the cluster center position is less than the set convergence threshold or the maximum number of iterations is reached.

[0021] Optionally, in S5, petrological observation of the core sample includes polarizing microscope observation of the rock thin section, and field emission scanning electron microscopy and X-ray energy spectrum analysis of the rock thin section to clarify the mineral distribution and characteristic mineral combination characteristics.

[0022] In a second aspect, a system for predicting the chemical phase of marine shale based on element logging curves is provided, the system comprising:

[0023] Data Collection Module: This module is used to collect elemental logging data of marine shale formations within a preset depth range of the drilling profile in the study area, obtaining multiple data sets consisting of depth values ​​and initial elemental logging data. The elemental logging data includes major elements, trace elements, and some rare earth elements.

[0024] Data screening and characteristic analysis module: used to perform mathematical statistical analysis on the initial element logging curves using SPSS software, obtain the minimum value, maximum value, median value and standard deviation of each set of element logging curve data, and screen out characteristic element logging curves based on the intersection diagram of the median value and the standard deviation value. Finally, characteristic analysis is performed on the screened element logging curves to summarize a variety of element combinations with paleoenvironmental significance;

[0025] Data dimensionality reduction processing module: used to process the selected element logging curve data using normalization methods to eliminate the influence of different dimensions, and use SPSS software to perform principal component analysis on the normalized element logging curve data to extract the first n principal components. Based on the extracted principal components and their factor loadings, the paleoenvironmental significance represented by different principal components is analyzed;

[0026] Shale chemical facies classification module: This module uses SPSS software to perform a systematic cluster analysis based on the extracted first n principal component scores. The number of cluster categories k is determined by the elbow rule. Subsequently, a k-means cluster analysis is performed on the principal component scores to obtain the locations of k cluster centers, their distances from the cluster centers, and the classification results. The contribution of different variables to the clustering is determined through one-way analysis of variance. Finally, k shale chemical facies types are defined based on the paleoenvironmental significance represented by the classification groups and principal components.

[0027] Accuracy verification module: used to identify the core sample and boundary sample data of k clusters based on the distance from the cluster center, collect core samples with the closest depth value to the core sample and boundary sample, prepare the samples into rock thin sections and scanning electron microscope optical slices, use polarizing microscope and field emission scanning electron microscope for petrological observation, and verify the accuracy of the shale chemical phase division results based on the mineral distribution and characteristic mineral combination characteristics.

[0028] The beneficial effects of the technical solutions provided in the embodiments of the present application include at least:

[0029] The present invention realizes a method for predicting the chemical phases of marine shale based on element logging curve data. Especially in the absence of core (rock cuttings) samples, it effectively divides the chemical phases of various types of shale, reduces the influence of human subjective factors, and provides important technical support for the next step of oil and gas exploration deployment. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other implementation drawings based on the provided drawings without inventive effort.

[0031] Figure 1 It is a workflow framework diagram of the present invention;

[0032] Figure 2 Schematic diagram of characteristic parameter screening based on the median and SD of element logging curve data in an embodiment of the present invention;

[0033] Figure 3 Schematic diagram of factor loadings of different element variables in the first five principal components in an embodiment of the present invention;

[0034] Figure 4 A descending line graph of clustering coefficients obtained by systematic clustering analysis in an embodiment of the present invention;

[0035] Figure 5 It is an overlay projection diagram of the principal component factor loading and the k-means cluster analysis results in an embodiment of the present invention;

[0036] Figure 6 Typical rock thin sections and SEM images of rock chip samples in the examples of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0038] In the description of the present invention, the terms "comprise", "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may also include other steps or units that are not explicitly listed but are inherent to these processes, methods, products or apparatuses, or steps or units that are added based on further optimization schemes conceived by the present invention.

[0039] Existing technologies include a method, system, equipment and storage medium for chemical phase division of lacustrine shale (CN118962078B). This invention patent uses a portable dispersive X-ray fluorescence spectrometer to test the element content of core samples, and adopts clustering algorithm and random forest algorithm to divide and predict the chemical phases of lacustrine shale.

[0040] However, this method requires core sampling and elemental testing, and cannot analyze the chemical phases of uncored shale sections. In addition, this method is mainly applicable to the chemical phase division of lacustrine shale and has not been applied to marine shale.

[0041] The second prior art is a scientific and efficient method for identifying the geochemical phases of shale geological sweet spots (CN118393105A). This invention patent determines the shale lithofacies type and the vertical distribution of the lithofacies by sampling and analyzing multiple wells, and finely characterizes the geochemical phases of the drilling profile through cluster analysis of geochemical elements, thereby dividing high-quality sweet spots.

[0042] This method requires a large number of core samples as well as mineralogical and elemental geochemical tests, and cannot analyze the chemical phases of the uncored shale. In addition, the shale geochemical phases proposed in this patent are still divided in the same way as the traditional shale lithofacies division scheme. On the basis of shale lithofacies, the sedimentary environment of different lithofacies is analyzed to select the shale lithofacies type that is most favorable for shale gas exploration. Therefore, the shale geochemical phase does not belong to the category of mud shale chemical phases.

[0043] The present invention relates to a method for predicting the chemical phases of marine shale based on element logging curves, which can overcome the above-mentioned problems or at least partially solve the above-mentioned problems. In particular, it provides an intelligent identification method for the chemical phases of marine shale that does not require coring and indoor analysis and testing. The method can be widely used in the intelligent identification of the chemical phases of marine shale and the comprehensive evaluation of shale reservoirs.

[0044] The present invention provides a method for predicting the chemical phase of marine shale based on element logging curves, the method comprising:

[0045] S1 Data Collection: Collect element logging curve data of marine shale formations within the preset depth range of the drilling profile in the study area to obtain multiple data sets consisting of depth values ​​and initial element logging curves.

[0046] Specifically, the collection of elemental well logging data sets involved collecting elemental well logging data for marine shale formations within a certain depth range of the drilling profile in the study area, including major elements, trace elements, and some rare earth elements, to obtain multiple data sets consisting of depth values ​​and initial elemental well logging curves.

[0047] S2 Data screening and characteristic analysis: SPSS software was used to perform mathematical statistical analysis on the initial element logging curves to obtain the minimum, maximum, median and standard deviation of each set of element logging curve data. Characteristic element logging curves were screened out based on the intersection diagram of the median and standard deviation values. Finally, the screened element logging curves were subjected to characteristic analysis to summarize a variety of element combinations with paleoenvironmental significance.

[0048] Among them, the screening and characteristic analysis of element logging curves. First, the initial element logging curves described in step S1 are subjected to mathematical and statistical analysis using SPSS software to obtain the minimum, maximum, median and standard deviation (SD) of each set of element logging curve data; secondly, due to the limited accuracy of element logging curve testing, the lower the element test value, the lower the test accuracy, and the less suitable it is to be adopted. However, the standard deviation of the logging curve data can reflect the degree of dispersion of the data, which helps to understand the fluctuation of the data. Therefore, the smaller the standard deviation value of a certain element logging curve, the less fluctuation the element logging curve data is, which is not conducive to the division of shale chemical phases. On the contrary, the larger the standard deviation value of a certain element logging curve, the more effective the element logging curve can be in characterizing different chemical phases. Therefore, a two-dimensional scatter analysis was performed on the median and SD values ​​of the above-mentioned element logging curve group, and a certain threshold was selected to screen out characteristic element logging curves. Finally, the screened element logging curves were subjected to characteristic analysis, and a variety of element combinations with paleoenvironmental significance were summarized, such as the terrigenous debris input element combination (including Ti and Cr, etc.), the mineral composition sensitive element combination (Al, Si, Ca and Mg), the paleoproductive element combination (including P, etc.), the paleosedimentary environment element combination (including Fe, S, V and Mo, etc.), and the special lithologic element combination (such as Mn, etc.).

[0049] S3 data dimensionality reduction processing: The screened element logging curve data were processed using the normalization method to eliminate the influence of different dimensions, and principal component analysis was performed on the normalized element logging curve data using SPSS software to extract the first n principal components. Based on the extracted principal components and their factor loadings, the paleoenvironmental significance represented by different principal components was analyzed.

[0050] Among them, first, based on the problem of inconsistent dimensions among the element logging curve data screened out in step S2, the above-mentioned element logging curve data are processed by normalization method to eliminate the influence of different dimensions; secondly, SPSS software is used to perform principal component analysis on the above-mentioned normalized element logging curve data, and the first n principal components (PC1, PC2, ..., PCn) are extracted through the cumulative contribution rate and eigenvalue of the principal components, where n is a preset value and is less than the total number of the above-mentioned normalized logging curves; finally, based on the above-mentioned n principal components extracted and their factor loadings, the paleoenvironmental significance represented by different principal components is analyzed.

[0051] In the above step S3, the data normalization calculation method is as follows:

[0052]

[0053] Among them, X i is the normalized data; x i is the data to be normalized; x min is the minimum value in the data set to be normalized; x max is the maximum value in the dataset to be normalized.

[0054] In the above step S3, the method of selecting the principal components includes: (1) selecting the number of principal components based on the cumulative contribution rate being greater than 70%; (2) selecting the number of principal components based on the eigenvalue being greater than 1.

[0055] In the above step S3, the paleoenvironmental significance represented by the principal component is analyzed based on the factor loading of the principal component. The factor loading value of a certain element variable in the i-th principal component (PCi) is greater than 0 and the larger the value is, the stronger the positive influence of the element variable is in PCi. If the factor loading value is less than 0 and the smaller the value is, the stronger the negative influence of the element variable is in PCi.

[0056] S4 Mudstone chemical facies classification: Based on the extracted first n principal component scores, SPSS software was used for systematic cluster analysis. The number of cluster categories k was determined by the elbow rule. Subsequently, k-means cluster analysis was performed on the principal component scores to obtain the positions of k cluster centers, the distances from the cluster centers, and the classification results. The contribution of different variables to the clustering was clarified through one-way analysis of variance. Finally, based on the paleoenvironmental significance represented by the classification groups and principal components, k mudstone chemical facies types were defined.

[0057] Among them, k-means cluster analysis is used to classify mud shale chemical phases. First, based on the scores of the first n principal components extracted in step S3, a systematic cluster analysis is performed using SPSS software. The number of cluster categories, k, is determined using the "elbow rule." Second, k-means cluster analysis is performed on the scores of the first n principal components extracted in step S3 using SPSS software. The number of clusters is set to the k value determined above, thereby obtaining the locations of k cluster centers, the distances from the cluster centers, and the classification results. A one-way analysis of variance (ANOVA) is then performed to determine the contribution of different variables (i.e., the scores of the first n principal components) to the clustering. Finally, based on the k classification groups and the paleoenvironmental significance represented by the different principal components in step S3, k mud shale chemical phase types are defined.

[0058] In the above step S4, the system cluster analysis method is selected as "inter-group linkage" and the measurement interval is selected as "squared Euclidean distance".

[0059] In the above step S4, the number of cluster categories k is determined by the "elbow rule", that is, the "coefficient" in the combined clustering is used to make a descending line graph, and the k value is determined according to the degree of reduction of the coefficient.

[0060] In the above step S4, the steps of the k-means clustering algorithm are as follows: (1) randomly initialize k points, i.e., cluster centers; (2) calculate the Euclidean distance of each data point to all cluster centers, and assign the data point to the cluster with the closest distance; (3) recalculate the mean of all data points in the current cluster, and use it as the cluster center for reallocation; (4) repeat the above cluster center calculation and allocation process until the position change of all cluster centers is less than the set convergence threshold, or the maximum number of iterations is reached, and finally k non-overlapping classification groups are obtained.

[0061] In the above step S4, the contribution of different variables to the clustering is analyzed by ANOVA, mainly based on the F value and significance (p), wherein the larger the F value of variable i is, i is a preset value and is less than the above n variables, representing that the variable i contributes more to the clustering, and the p value of variable i <0.05 represents that variable i has a significant contribution to the clustering.

[0062] S5 accuracy verification: The core sample and boundary sample data of k clusters are identified based on the distance from the cluster center. The core samples with the closest depth to the core samples and boundary samples are collected. The samples are prepared into rock thin sections and scanning electron microscope optical slices. Polarizing microscope and field emission scanning electron microscope are used for petrological observation. The accuracy of the shale chemical phase division results is verified based on the characteristics such as mineral distribution and characteristic mineral association.

[0063] The accuracy of the shale chemical phase division is verified. First, based on the "distance from the cluster center" in step S4 above, the core sample and boundary sample data of k clusters are identified based on the minimum and maximum distances, thereby representing the range covered by each sample cluster. Second, core (rock chip) samples with the closest depth values ​​to the core samples and boundary samples are collected, and the core (rock chip) samples are prepared into rock thin sections and scanning electron microscope (SEM) light sheets. Petrological observations are performed using a polarizing microscope and a field emission scanning electron microscope. Based on the characteristics of the mineral distribution and characteristic mineral assemblages in the core (rock chip) samples, the accuracy of the shale chemical phase division results is verified.

[0064] In the above step S5, the petrological observation of the core (chip) sample can first be carried out by observing the rock thin section, and then the rock thin section is carbon-plated (or platinum-plated) and then subjected to scanning electron microscopy and X-ray energy spectrum (EDS) analysis, so as to qualitatively observe and quantitatively characterize specific minerals and assemblages.

[0065] In the above step S5, the characteristics of the mineral distribution and characteristic mineral combination in the core (cutting) sample include: (1) the distribution of rock-forming minerals (such as quartz, feldspar, calcite, dolomite), which can characterize the main mineral composition and terrigenous debris input conditions in the mud shale; (2) the distribution of plankton (such as siliceous radiolarians) and organic matter, which can characterize the paleoproductivity conditions in the mud shale; (3) the distribution of sulfides (such as raspberry pyrite and diagenetic pyrite), which can characterize the redox conditions of bottom water or pore water; (4) the distribution of special minerals (such as apatite and manganese-containing minerals), which can characterize the presence of metal ore layers (such as phosphorite layers or manganese ore layers).

[0066] In an optional embodiment of the present application, this embodiment provides a method for predicting the chemical phase of marine shale based on element logging curves. The specific implementation process is as follows: Figure 1 This example uses the Lower Cambrian marine shale interval in Well X in the Sichuan Basin as the research object. The data set is composed of element logging curves. Principal component analysis and k-means cluster analysis are used to intelligently identify the chemical phases of the shale. The following steps are included:

[0067] Step S1: Collection of elemental log data sets. Elemental log data, including major elements, trace elements, and some rare earth elements, are collected for the marine shale formation within a certain depth range of the well profile in the study area. Multiple data sets consisting of depth values ​​and initial elemental log curves are obtained.

[0068] Step S2: Screening and Characteristic Analysis of Element Log Curves. Using SPSS software, perform mathematical and statistical analysis on the initial element log curves described in Step S1. The minimum, maximum, median, and standard deviation (SD) of each element log curve data set are obtained. Characteristic element log curves are screened based on the intersection of the median and SD values. Finally, characteristic analysis is performed on the selected element log curves, and various element combinations with paleoenvironmental significance are summarized.

[0069] Step S3: Dimensionality reduction of the elemental log data set. The elemental log data selected in step S2 are processed using normalization methods to eliminate the influence of different dimensions. SPSS software is then used to perform principal component analysis on the normalized elemental log data, extracting the first n principal components (PC1, PC2, ..., PCn). Finally, based on the extracted n principal components and their factor loadings, the paleoenvironmental significance of each principal component is analyzed.

[0070] Step S4: Use k-means cluster analysis to classify the chemical phases of the mudstone. Based on the scores of the first n principal components extracted in step S3 above, a systematic cluster analysis is performed using SPSS software. The number of cluster categories k is determined using the "elbow rule". Subsequently, a k-means cluster analysis is performed on the scores of the first n principal components to obtain the locations of the k cluster centers, the distances from the cluster centers, and the classification results. A one-way analysis of variance (ANOVA) is then used to clarify the contribution of different variables to the clustering. Finally, based on the above k classification groups and the paleoenvironmental significance represented by the different principal components in step S3 above, k mudstone chemical phase types are defined.

[0071] Step S5: Verification of the accuracy of shale chemical phase division. According to the "distance from the cluster center" in the above step S4, the core sample and boundary sample data of k clusters are identified according to the minimum distance and the maximum distance, thereby representing the range covered by each sample cluster. Subsequently, by collecting core (rock chip) samples with the closest depth value to the core sample and the boundary sample, the core (rock chip) samples are prepared into rock thin sections and scanning electron microscope (SEM) light sheets, and petrological observations are performed using a polarizing microscope and a field emission scanning electron microscope. According to the characteristics of the mineral distribution and characteristic mineral combination in the core (rock chip) samples, the accuracy of the shale chemical phase division results is verified.

[0072] In step S1, element logging data for the Lower Cambrian shale section of Well X was collected at a depth range of 4500 to 5240 meters. The Lower Cambrian includes the Maidiping Formation, the Qiongzhusi Formation, and the bottom of the Canglangpu Formation, and also includes dolomite at the top of the Upper Sinian Dengying Formation. These data were used for chemical comparison of the shale. The element logging test interval was 1 meter, and a total of 745 sets of data were accumulated.

[0073] In step S1, the element logging curve data collected from the Lower Cambrian shale section of Well X include Na, Mg, Al, Si, P, S, K, Ca, Cd, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Rb, Sr, Y, Zr, Nb, Mo, and Ag, where the unit of the main elements (including Si, Al, K, Na, Ca, Mg, and Fe) is %, and the unit of the remaining elements is ppm.

[0074] In step S2, the initial element logging curves described in step S1 are subjected to mathematical statistical analysis using SPSS software to obtain the minimum, maximum, median, and SD values ​​of each set of element logging curve data as shown in Table 1. An intersection diagram (e.g., Figure 2 ), the median value > 5ppm and SD > 1 were selected as the thresholds to screen out characteristic element logging curves, among which the major elements mainly include Si, Al, Ca, Mg, Ti, P, Mn, Fe and S ( Figure 2 A), trace elements mainly include V, Cr and Mo ( Figure 2 B). Among these 12 elements, Ti and Cr can indicate detrital input; higher Ti and Cr levels indicate stronger detrital input. Elements that can indicate mineral composition include Al, Si, Ca, and Mg. A higher Al content indicates a higher clay mineral content, a higher Si content indicates a higher quartz mineral content, and higher Ca and Mg contents indicate a higher carbonate content. A high Ca content, while a low Mg content, indicates a calcite enrichment, while high Ca and Mg contents simultaneously indicate a dolomite enrichment. Elements that can indicate paleoproductivity include P; higher P levels indicate higher paleoproductivity. Abnormally high P levels can also indicate the presence of phosphate deposits. Elements that can indicate paleo-sedimentary environments include Fe, S, V, and Mo; higher levels of these elements indicate a more reducing sedimentary environment. Other elements (such as Mn) can indicate the presence of special mineral deposits (such as manganese-bearing deposits).

[0075] Table 1 Statistics of element logging curves for the Cambrian shale layer in Well X

[0076]

[0077] The unit of main elements (including Si, Al, K, Na, Ca, Mg, and Fe) is %, and the unit of other elements is ppm.

[0078] In step S3, the 12 element logging curves selected in step S2 are normalized using SPSS software to eliminate the influence of different dimensions. The data normalization calculation method is as follows:

[0079]

[0080] Among them, X i is the normalized data; x i is the data to be normalized; x min is the minimum value in the data set to be normalized; x max is the maximum value in the dataset to be normalized.

[0081] In step S3, principal component analysis was performed on the above 12 normalized element logging curve data using SPSS software. The eigenvalues ​​and cumulative contribution rates corresponding to the first 10 principal components are shown in Table 2. Based on the eigenvalue > 1 and the cumulative contribution rate > 70%, the first 5 principal components with a cumulative contribution rate of 73.29% were selected, indicating that these 5 principal components contain 73.29% of the information of the total original variables and can basically represent the entire data set.

[0082] Table 2. Principal component contribution analysis of element logging curves of the Cambrian shale interval in Well X

[0083] Principal component number Eigenvalue Cumulative contribution rate (%) PC1 3.76 3136 PC2 1.60 44.72 PC3 1.27 55.27 PC4 1.16 64.96 PC5 1.00 73.29 PC6 0.86 80.43 PC7 0.69 86.15 PC8 0.56 90.84 PC9 0.47 94.78 PC10 0.40 98.12

[0084] In step S3, based on the first five principal components and the component matrix extracted by the principal component analysis, the factor loading of different element variables in each principal component is plotted as follows: Figure 3As shown in the figure, PC1 is mainly loaded with Si, Al, and Fe, indicating that the mineral composition is dominated by clay minerals and is relatively enriched in pyrite, representing a clayey clastic rock and a relatively reducing sedimentary environment. PC1 is mainly loaded with Ca and Mg, indicating that the mineral composition is dominated by carbonate minerals. PC2 is mainly loaded with Ti and Cr, indicating a relatively high terrigenous clastic input. PC2 is mainly loaded with Fe, reflecting a relatively reducing sedimentary environment. PC3 is mainly loaded with Mn, which may reflect the presence of manganese-bearing ore layers. PC3 is mainly loaded with P, which may reflect a relatively high paleoproductivity or the presence of phosphate layers. PC4 is mainly loaded with S, which may reflect a relatively enriched sulfide. PC4 is mainly loaded with Mo, which may reflect a strongly reducing (or sulfiding) sedimentary environment. PC5 is mainly loaded with P, which may reflect a relatively high paleoproductivity or the presence of phosphate layers. PC5 is mainly loaded with Mg, reflecting a relatively high dolomite content.

[0085] In step S4, SPSS software was used to perform a systematic cluster analysis on the first five principal component scores extracted in step S3. The analysis method was "Inter-group linkage" and the measurement interval was "Square Euclidean distance". A descending line graph was drawn by combining the "coefficients" in the clustering. Figure 4 ), according to the “elbow rule”, it is found that the turning point where the coefficient decreases significantly to slowly decreases is at k=8, thus determining the number of cluster categories to be 8.

[0086] In step S4, SPSS software is used to perform k-means cluster analysis on the first five principal component scores extracted in step S3 above, with the number of clusters set to 8 as determined above. First, eight points, i.e., cluster centers, are randomly initialized. Second, the Euclidean distance from each data point to all cluster centers is calculated, and the data point is assigned to the cluster with the closest distance. Third, the mean of all data points in the current cluster is recalculated and used as the cluster center for reallocation. Finally, the above cluster center calculation and assignment process is repeated until the change in the position of all cluster centers is less than the set convergence threshold or the maximum number of iterations is reached, ultimately resulting in eight non-overlapping classification groups, and the eight cluster center positions, distances to the cluster centers, and classification results are obtained.

[0087] In step S4, the one-way analysis of variance (ANOVA) data of the k-means cluster analysis are shown in Table 3. According to the distribution of F values ​​and significance (p), it can be found that the p values ​​of all variables are less than 0.05, indicating that all variables have significant contributions to the clustering. However, the F values ​​of all variables are ranked from large to small as follows: PC1>PC5>PC3>PC4>PC2, indicating that PC1 contributes the most to the clustering and PC2 contributes the least to the clustering.

[0088] Table 3 ANOVA data table of k-means cluster analysis

[0089]

[0090] In step S4, based on the results of the principal component analysis in step S3 and the k-means cluster analysis, the principal component scores and principal component factor loadings are superimposed and projected as follows: Figure 5 As shown in the figure, the first group of samples has an abnormally low PC2 score, reflecting an oxidized mud shale facies; the second group of samples has an abnormally low PC1 and PC5 scores, reflecting an oxidized dolomite facies; the third group of samples has an abnormally low PC3 and PC5 scores, reflecting a phosphorus-rich calcareous mud shale facies (possibly a phosphorite layer); the fifth group of samples has an abnormally high PC5 score and relatively low PC3 and PC4 scores, and only contains two samples, showing characteristics of rich P and rich Cr, reflecting a phosphorus-bearing mud shale facies; the sixth group of samples has an abnormally low PC3 score and relatively high PC5 score, reflecting a manganese-rich calcareous mud shale facies; the seventh group of samples has an abnormally low PC4 score and relatively low PC1 score, reflecting a strongly reduced (or sulfided) calcareous mud shale facies; the eighth group of samples has a relatively high PC1 score, covering most of the samples, and is an anoxic clayey mud shale facies; the fourth group of samples is between the second and eighth groups and is an anoxic calcareous mud shale facies.

[0091] In step S5, according to the "distance from the cluster center" in step S4 above, the core sample and boundary sample data of 8 clusters are found according to the minimum distance and the maximum distance, totaling 16, thereby representing the range covered by each sample cluster.

[0092] In step S5, 16 rock chip samples with the closest depth to the core sample and the boundary sample are collected, and the rock chip samples are prepared into rock thin sections, polished, and petrologically observed using a polarizing microscope. The rock chip samples are named according to their mineral composition and their relative content. Subsequently, the rock thin sections are observed in more detail using field emission scanning electron microscopy and X-ray energy spectrum (EDS) analysis to clarify the mineral distribution and characteristic mineral combination in the rock chip samples, thereby determining the paleoenvironmental background of the rock chip sample deposition and verifying the accuracy of the shale chemical phase division results. For example, Figure 6Typical rock thin sections and SEM images of rock chip samples are given (A is a rock thin section, showing a large amount of dolomite, lacking organic matter and pyrite; B is an SEM image, showing a large amount of apatite minerals, associated with carbonate minerals; C is an SEM image, showing a large amount of manganese-containing minerals, including rhodochrosite and pyroxenite; D is an SEM image, showing a large amount of clay minerals and berry-shaped pyrite. The white cross represents the EDS measurement point). Specifically, the second group of oxidized dolomite phases mainly develops dolomite, lacking organic matter and other reducing minerals (such as pyrite) ( Figure 6 A); Group 3 phosphorus-rich calcareous shale facies and Group 6 manganese-rich calcareous shale facies contain large amounts of apatite minerals and manganese-containing minerals ( Figure 6 B, C); The anoxic clayey shale phase of Group 7 contains a large amount of clay minerals and spherical pyrite ( Figure 6 D), indicating a relatively low sedimentation rate and a relatively reducing depositional environment.

[0093] The present application also provides a system for predicting the chemical phase of marine shale based on element logging curves, which may include:

[0094] Data Collection Module: This module is used to collect elemental logging data of marine shale formations within a preset depth range of the drilling profile in the study area, obtaining multiple data sets consisting of depth values ​​and initial elemental logging data. The elemental logging data includes major elements, trace elements, and some rare earth elements.

[0095] Data screening and characteristic analysis module: used to perform mathematical statistical analysis on the initial element logging curves using SPSS software, obtain the minimum value, maximum value, median value and standard deviation of each set of element logging curve data, and screen out characteristic element logging curves based on the intersection diagram of the median value and the standard deviation value. Finally, characteristic analysis is performed on the screened element logging curves to summarize a variety of element combinations with paleoenvironmental significance;

[0096] Data dimensionality reduction processing module: used to process the selected element logging curve data using normalization methods to eliminate the influence of different dimensions, and use SPSS software to perform principal component analysis on the normalized element logging curve data to extract the first n principal components. Based on the extracted principal components and their factor loadings, the paleoenvironmental significance represented by different principal components is analyzed;

[0097] Shale chemical facies classification module: This module uses SPSS software to perform a systematic cluster analysis based on the extracted first n principal component scores. The number of cluster categories k is determined by the elbow rule. Subsequently, a k-means cluster analysis is performed on the principal component scores to obtain the locations of k cluster centers, their distances from the cluster centers, and the classification results. The contribution of different variables to the clustering is determined through one-way analysis of variance. Finally, k shale chemical facies types are defined based on the paleoenvironmental significance represented by the classification groups and principal components.

[0098] Accuracy verification module: used to identify the core sample and boundary sample data of k clusters based on the distance from the cluster center, collect core samples with the closest depth value to the core sample and boundary sample, prepare the samples into rock thin sections and scanning electron microscope optical slices, use polarizing microscope and field emission scanning electron microscope for petrological observation, and verify the accuracy of the shale chemical phase division results based on the characteristics such as mineral distribution and characteristic mineral combination.

[0099] The specific limitations of the system for predicting the chemical phases of marine shale based on elemental well logging curves can be found in the limitations of the method for predicting the chemical phases of marine shale based on elemental well logging curves described above and will not be further elaborated here. Each module in the aforementioned system for predicting the chemical phases of marine shale based on elemental well logging curves can be implemented in whole or in part through software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0100] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0101] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for predicting the chemical phase of marine shale based on element logging curves, characterized in that: The method comprises: S1 Data Collection: Collect elemental logging data of marine shale formations within the preset depth range of the drilling profile in the study area, and obtain multiple data sets consisting of depth values ​​and initial elemental logging data. The elemental logging data includes major elements, trace elements, and some rare earth elements. S2 Data Screening and Characteristic Analysis: SPSS software was used to perform mathematical statistical analysis on the initial element logging curves to obtain the minimum, maximum, median, and standard deviation of each set of element logging curve data. Characteristic element logging curves were screened based on the intersection of the median and standard deviation values. Finally, characteristic analysis was performed on the screened element logging curves to summarize a variety of element combinations with paleoenvironmental significance. S3 data dimensionality reduction: The normalization method was used to process the selected element logging curve data to eliminate the influence of different dimensions. SPSS software was used to perform principal component analysis on the normalized element logging curve data to extract the first n principal components. Based on the extracted principal components and their factor loadings, the paleoenvironmental significance represented by different principal components was analyzed. S4 Shale chemical facies classification: Based on the extracted first n principal component scores, a systematic cluster analysis was performed using SPSS software. The number of cluster categories k was determined using the elbow rule. Subsequently, a k-means cluster analysis was performed on the principal component scores to obtain the locations of k cluster centers, their distances from the cluster centers, and the classification results. The contribution of different variables to the clustering was determined using a one-way analysis of variance. Finally, k shale chemical facies types were defined based on the paleoenvironmental significance represented by the classification groups and principal components. S5 accuracy verification: The core sample and boundary sample data of k clusters are identified based on the distance from the cluster center. The core samples with the closest depth to the core samples and boundary samples are collected. The samples are prepared into rock thin sections and scanning electron microscope optical slices. Petrological observations are carried out using a polarizing microscope and a field emission scanning electron microscope. The accuracy of the shale chemical phase division results is verified based on the mineral distribution and characteristic mineral assemblage characteristics.

2. The method according to claim 1, characterized in that In S1, the collected element logging curve data include at least Si, Al, K, Na, Ca, Mg, Fe, Ti, V, Cr, Mn, P, S, Ni, Cu, Zn, As, Se, Rb, Sr, Y, Zr, Nb, Mo, and Ag elements, where the unit of the main elements is %, and the unit of the remaining elements is ppm; the main elements are specifically Si, Al, K, Na, Ca, Mg, and Fe.

3. The method according to claim 1, characterized in that In S2, when SPSS software was used to perform mathematical statistical analysis on the initial element logging curves, the minimum, maximum, median, and standard deviation of each set of element logging curve data were calculated, and characteristic element logging curves were screened out based on the thresholds of median value >5ppm and SD>1.

4. The method according to claim 1, wherein In S2, the element combinations with paleoenvironmental significance summarized include: terrigenous debris input element combination, including Ti and Cr; mineral composition sensitive element combination, including Al, Si, Ca and Mg; paleoproductivity element combination, including P; paleosedimentary environment element combination, including Fe, S, V and Mo; and special lithologic element combination, including Mn.

5. The method according to claim 1, wherein In S3, the specific calculation formula for normalization is: Among them, X i is the normalized data; x i is the data to be normalized; x min is the minimum value in the data set to be normalized; x max is the maximum value in the dataset to be normalized.

6. The method according to claim 1, characterized in that In S3, during principal component analysis, the first n principal components are extracted based on the conditions that the eigenvalue is greater than 1 and the cumulative contribution rate is greater than 70%.

7. The method according to claim 1, characterized in that In S4, the elbow rule in the system cluster analysis is used to draw a descending line graph of the combined clustering coefficient, and the number of cluster categories k is determined according to the degree of reduction of the coefficient.

8. The method according to claim 1, characterized in that In S4, the specific steps of k-means cluster analysis include: Randomly initialize k cluster centers, calculate the Euclidean distance of each data point to the cluster center and assign the data point to the nearest cluster, recalculate the mean of all data points in the cluster as the new cluster center, and repeat the process until the change in the cluster center position is less than the set convergence threshold or the maximum number of iterations is reached.

9. The method according to claim 1, characterized in that In S5, petrological observations of core samples included polarizing microscope observations of rock thin sections, as well as field emission scanning electron microscopy and X-ray energy spectrum analysis of rock thin sections to clarify the mineral distribution and characteristic mineral assemblage features.

10. A system for predicting the chemical phase of marine shale based on element logging curves, characterized in that: The system comprises: Data Collection Module: This module is used to collect elemental logging data of marine shale formations within a preset depth range of the drilling profile in the study area, obtaining multiple data sets consisting of depth values ​​and initial elemental logging data. The elemental logging data includes major elements, trace elements, and some rare earth elements. Data screening and characteristic analysis module: used to perform mathematical statistical analysis on the initial element logging curves using SPSS software, obtain the minimum value, maximum value, median value and standard deviation of each set of element logging curve data, and screen out characteristic element logging curves based on the intersection diagram of the median value and the standard deviation value. Finally, characteristic analysis is performed on the screened element logging curves to summarize a variety of element combinations with paleoenvironmental significance; Data dimensionality reduction processing module: used to process the selected element logging curve data using normalization methods to eliminate the influence of different dimensions, and use SPSS software to perform principal component analysis on the normalized element logging curve data to extract the first n principal components. Based on the extracted principal components and their factor loadings, the paleoenvironmental significance represented by different principal components is analyzed; Shale chemical facies classification module: This module uses SPSS software to perform a systematic cluster analysis based on the extracted first n principal component scores. The number of cluster categories k is determined by the elbow rule. Subsequently, a k-means cluster analysis is performed on the principal component scores to obtain the locations of k cluster centers, their distances from the cluster centers, and the classification results. The contribution of different variables to the clustering is determined through one-way analysis of variance. Finally, k shale chemical facies types are defined based on the paleoenvironmental significance represented by the classification groups and principal components. Accuracy verification module: used to identify the core sample and boundary sample data of k clusters based on the distance from the cluster center, collect core samples with the closest depth value to the core sample and boundary sample, prepare the samples into rock thin sections and scanning electron microscope optical slices, use polarizing microscope and field emission scanning electron microscope for petrological observation, and verify the accuracy of the shale chemical phase division results based on the mineral distribution and characteristic mineral combination characteristics.

Images