A quantitative evaluation method for laminar flow based on electro-imaging wellbore cross-sectional projection

Abstract

This invention discloses a method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection, belonging to the field of applied geophysical logging technology. The method includes the following steps: data preprocessing; wellbore interface projection; projection image generation and truncated mean calculation; and quantitative statistical analysis of relative laminae density. Based on the imaging principles of dynamic and static images in electro-imaging logging data, this method aims to fully leverage the high resolution advantage equivalent to the original microelectrode curves to quantitatively evaluate the development density of laminae in the target well and achieve rapid and efficient batch processing of data across the entire well section.

Description

Technical Field

[0001] This invention belongs to the field of applied geophysical logging technology, specifically relating to a method for quantitative evaluation of stratigraphy based on electrical imaging wellbore cross-sectional projection. Background Technology

[0002] Tight clastic reservoirs have complex lithological assemblages, with the reservoir layer sandwiched between upper and lower source rocks and exhibiting thin interbedded layers, resulting in extremely strong spatial heterogeneity and anisotropy in reservoir parameters. Bedding is a layered structure in which rock properties vary vertically. The most basic unit of bedding is the laminar bedding, within which there are no other layered structures visible to the naked eye.

[0003] With the deepening of unconventional oil and gas exploration and development both domestically and internationally, scholars have gradually paid attention to the influence of sedimentary structures such as laminae on various aspects of reservoirs or source rocks, including organic matter distribution, pore structure, and rock mechanical properties. Laminae typically refer to layered structures with a thickness of less than 1 cm. Due to limitations in conventional downhole measurement methods, research on laminae has largely focused on core experiments and oil and gas geology, with relatively few methods for evaluating laminae using well logging data, thus hindering exploration and development efficiency.

[0004] Microresistivity imaging logging, developed from formation dip logging, is currently the most widely used method in electrical imaging logging. It uses an array of button electrodes attached to the plates to measure hundreds of conductivity curves, which can reflect the small relative changes in formation around the wellbore. It has extremely high vertical resolution, making fine evaluation of the stratigraphy possible.

[0005] In electro-imaging logging, laminae exhibit alternating light and dark variations in both dynamic and static images. Laminae at different angles appear as sinusoidal curves of varying amplitudes. Lai Fuqiang, Wang Changjiang, and others used a watershed algorithm to identify sandy laminae in static electro-imaging images after filling blank strips. Kumar, an engineer at Schlumberger, proposed an electro-imaging vertical slicing technique that allows direct observation of sedimentary and geological structures, potentially replacing core samples to some extent. Li Zhenlin et al. applied Hough transform to extract laminae boundaries from vertical slice images after edge detection. Tian Han et al. statistically analyzed the extreme value changes of all microelectrode curves after filling blank strips to reflect the degree of laminae development and compared the number of visible laminae on vertical slice images, concluding that the original microelectrode curves have higher resolution than the slice images. Duan Chaowei et al. similarly believed that the original microelectrode curves inherit the resolution of electro-imaging logging and automatically extracted laminae based on the derivative characteristics of the microelectrode curves.

[0006] Therefore, current methods for evaluating laminae using electrical imaging logging data can be broadly categorized into two types: those based on dynamic and static images and those based on microelectrode curves. Images offer greater visual appeal, while raw microelectrode curves provide higher resolution. There is currently no laminae evaluation method that combines the advantages of both approaches. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention provides a quantitative evaluation method for laminae based on electro-imaging wellbore cross-sectional projection. Starting from the imaging principles of dynamic and static images in electro-imaging logging data, it aims to fully leverage the high-resolution advantage equivalent to the original microelectrode curves to quantitatively evaluate the development density of laminae in the target well and achieve rapid and efficient batch processing of data across the entire well section.

[0008] The objective of this invention is achieved through the following technical solution:

[0009] A method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection includes the following steps:

[0010] Step 1: Data preprocessing;

[0011] Step 2: Wellbore interface projection;

[0012] Step 3: Generate the projected image and calculate the truncated mean;

[0013] Step 4: Quantitative statistical analysis of relative laminar density.

[0014] Preferably, step one includes generating an electrical imaging dynamic map and picking up the formation dip angle.

[0015] Preferably, step two includes displacement depth calculation, pixel matrix reconstruction, and inverse distance weighted interpolation to fill in missing values.

[0016] Preferably, the displacement depth calculation includes the following steps: The open-hole wellbore is approximately a cylinder in the inclination direction. The sedimentary structures in the formation are represented by cross-sections with different dip angles on the cylinder. Unfolding the outer surface of the cylinder along a certain vertical tangent yields an imaging logging image. The original layered structure is thus stretched into the shape of a sine function curve. The original dynamic and static images are projected onto the vertical cross-section along a fixed tangent. Thus, the layers are displayed as straight lines with the same slope and dip angle on the projected image, while maintaining the corresponding thickness. Two approximate semi-circular projection images are generated on both sides of the cross-section. The projection image can be obtained by calculating the displacement of the data points of the imaging matrix in depth and reconstructing the matrix array according to the depth. The total number of columns in the imaging logging image is 360, corresponding to 0 to 360° of the top and bottom surfaces of the cylinder. Therefore, the column number of the matrix data can be equivalent to the angle relative to true north. The displacement depth calculation formula for the k-th imaging matrix data point is as follows:

[0017] ΔDepth=r·tan(Dip).cos(k-Azimuth) (1);

[0018]

[0019] Where ΔDepth is the displacement depth of the k-th imaging matrix data point;

[0020] r is the well radius at the current depth;

[0021] Dip is the construction dip angle at the current depth;

[0022] k is the column number to which the current data point belongs;

[0023] Azimuth is the azimuth angle at the current depth;

[0024] C1 and C2 are wellbore curves.

[0025] Preferably, in step two, the original depth of the data point is added to the displacement depth to obtain the new depth of the data point. Then, based on the new depth of each data point, the data is aligned to the nearest original depth column to achieve the reconstruction of the matrix array indexed by depth.

[0026] Preferably, during the array reconstruction process, a small number of data points may be missing in certain positions of the array due to errors. Therefore, inverse distance weighted interpolation is used to fill in the missing positions of the matrix.

[0027] Preferably, the interpolation includes the following steps:

[0028] (1) Determine the location to be interpolated and the location of known sample points;

[0029] (2) Calculate the weights of the known sample points;

[0030] (3) Solve for the weighted average.

[0031] Preferably, the coordinates and values ​​of data points within a certain range of the matrix array are first searched for.

[0032] For each missing position, calculate its distance to the nearby sample points found in the search, and convert the distance into a weight using Equation 3. The smaller the distance between the sample point and the missing position, the greater the impact on the interpolation. That is, the weight is inversely proportional to the distance.

[0033]

[0034] Where: w i The weight of the i-th sample point;

[0035] d i This represents the distance between the sample point and the missing location.

[0036] p can be 2 or 3, representing the Euclidean distance or the Manhattan distance, respectively.

[0037] After calculating the weight of each sample point, the function values ​​of each sample point are weighted and averaged to obtain the function value at the missing position, as shown in the following formula:

[0038]

[0039] Where z(x, y) is the interpolation result for the missing position;

[0040] The coordinates of the i-th sample point are (x... i y i );

[0041] N is the number of sample points;

[0042] w i The weight of the i-th sample point;

[0043] z i Let be the imaging value of the i-th sample point.

[0044] Preferably, in step three, after the projected image is generated, the array obtained by reading data along a fixed tilt angle is the array vector in the same direction as the texture.

[0045] Preferably, it includes the following steps:

[0046] (1) Cutting, translation and flipping of semi-circular projection images;

[0047] (2) Extraction of tilt angle guidance array;

[0048] (3) Calculation of truncated mean.

[0049] Preferably, the original image, after being projected onto the cross section, will produce two semi-circular projection images. The azimuth angle is the angle between the No. 1 electrode and the due north direction. The column number of each data column can be regarded as the angle with due north. When the interpolation between the column number and the azimuth angle is ±90°, the value of cos(k-Azimuth) is 0. Cutting along this value line and translating the image portion exceeding 90° to the left side will generate a pair of mirror-symmetric images. Cutting the image along the central axis and flipping the right image will yield two similar projection images. The texture on the two projection images retains the original electro-imaging values, only stretched in morphology into oblique lines with slopes consistent with the orientation.

[0050] The electrodes of the electrical imaging logging instrument cannot cover all points around the well, and there are gaps between the electrodes, which are displayed as blank stripes on dynamic and static maps. By taking the value of each row vector along the dip direction across the blank strip, all information of the covering texture can be obtained.

[0051] After removing the abnormally high conductivity and abnormally high resistivity portions of the interfering stratum information, the arithmetic mean of the valid data portion is calculated using the following formula:

[0052]

[0053] Where σ is the conductivity value or the pixel value of the dynamic or static image;

[0054] α is the truncation coefficient;

[0055] n is the number of data items;

[0056] m is the number of data points removed.

[0057] Preferably, the cutoff values ​​between the main peak and the abnormally high conductance and abnormally high resistance portions are found using a threshold segmentation method. The array is truncated at these cutoff points, and the effective data in the middle is extracted and the average value is calculated. First, a frequency distribution histogram is plotted on the extracted vector array. Then, the inter-class variance at all data points is calculated using OTSU. The point corresponding to the maximum value of the inter-class variance is the first cutoff point, i.e., the cutoff point that maximizes the separation of the original array. The first cutoff point is either the cutoff point with the abnormally high conductance portion or the cutoff point with the abnormally high resistance portion. Abnormal data outside the cutoff point is truncated, and the remaining array is retained. Then, the second cutoff point is found using OTSU calculation, and the array is truncated again. The final remaining array is the effective array after removing abnormally high and low conductance data. The formula for calculating the maximum inter-class variance is as follows:

[0058]

[0059] Where, ρ 2 The inter-class variance of the data on both sides of this point;

[0060] MG is the global mean;

[0061] P1 represents the data frequency to the left of the current threshold;

[0062] m is the cumulative mean of all thresholds;

[0063] By replacing the original cross-sectional projection image with the truncated mean value and filling it along the dip direction, a smooth cross-sectional projection image is generated, which can intuitively show the development of thicker bedding.

[0064] Preferably, the interface locations of alternating lamellar layers correspond to abrupt changes in electrical properties, and correspondingly, in the truncated mean curves reflecting typical characteristics... The above is displayed as a "jumping peak," where the laminae with relatively high conductivity in the surrounding rock bulge to the right on the curve, i.e., a truncated mean curve. The "peak" of the mean curve; the laminae with relatively high resistivity to the surrounding rock show a leftward dip on the curve, i.e., a truncated mean curve. The "trough"; the more developed the laminae, the more pronounced the truncated mean curve. The more frequent the fluctuations, the faster the peaks and valleys alternate. Therefore, the density of laminar development can be reflected by quantifying the peak and valley frequencies of the curve.

[0065] Preferably, the peaks and troughs of the curve correspond to the extreme points of the curve, and the three sufficient conditions for determining extreme values ​​are:

[0066] (1) The function is differentiable at this point, and the signs are opposite on both sides;

[0067] (2) The first derivative is zero, and the second derivative is not zero;

[0068] (3) If the lowest non-zero derivative of the function at a point is of an even order, then it is an extreme point.

[0069] Preferably, the actual logging curve is a scatter line with equal sampling intervals. When magnified to a certain scale, it usually appears as a linearly interpolated broken line. Even if the truncated mean curve has the same sampling interval as electrical imaging logging, three sampling points are taken as a unit, and one sampling point is taken as the number of advance steps. The logging values ​​of each sampling point are compared with the two adjacent sampling points in a loop to find the maximum or minimum point, which is the extreme point of the curve.

[0070] After finding the extreme point using the above method, it is necessary to quantify the peak and valley frequencies within the range. Using the logging depth as an index, a sliding window is set from top to bottom. The window size can be set, and the sliding step size is one sampling interval. The number of peaks and valleys appearing in the window after each sliding is counted, corresponding to the center depth of the window. The obtained value is the change in the laminar frequency within the range, reflecting the stable development of the formation around the wellbore.

[0071] The beneficial effects of this technical solution are as follows:

[0072] I. This invention provides a quantitative evaluation method for striations based on electro-imaging wellbore cross-section projection. Based on the principle of electro-imaging logging, the wellbore image is projected onto a vertical cross-section along the north-south axis of the wellbore. This causes the striations, which originally appear as sinusoidal functions in the dynamic and static electro-imaging logging images, to appear as straight lines with different slopes on the cross-section. By taking values ​​along the straight lines and calculating the truncated mean, the average conductivity curve can be obtained. Furthermore, the peak and valley frequencies of the average conductivity curve within a fixed window are statistically analyzed to achieve quantitative statistics on striation development density.

[0073] II. The present invention provides a quantitative evaluation method for stratigraphy based on electro-imaging wellbore cross-section projection. For the first time, a curve reflecting the electrical changes of stratigraphy is extracted from the electro-imaging logging image of the wellbore cross-section projection. This curve is subjected to secondary threshold segmentation to remove the influence of abnormal high and low values ​​other than stratigraphy. The truncated mean is used as the image base value, and the projection image is filled along the original dip direction to obtain a high-precision cross-sectional projection image, which can intuitively display the distribution of stratigraphy after removing interference information, and serve as a reference for core depth positioning.

[0074] Third, the present invention provides a quantitative evaluation method for laminar density based on electro-imaging wellbore cross-section projection. Through a statistical method using a sliding window, it quantifies the relative change in laminar density within a range, reflecting the changing trend of the number of laminar layers in the vertical direction of the formation corresponding to the wellbore depth.

[0075] IV. The present invention provides a quantitative evaluation method for bedding based on the projection of the wellbore section by electrical imaging, which fully leverages the high resolution advantage of electrical imaging logging, improves the accuracy of bedding or bedding logging evaluation, and provides a scientific basis for evaluating source rocks and reservoir parameters affected by bedding development. Attached Figure Description

[0076] Figure 1 This is a flowchart of the present invention;

[0077] Figure 2 This is a schematic diagram illustrating the principles of electro-imaging dynamic and static images and wellbore cross-section projection images in this invention;

[0078] Figure 3 This is a schematic diagram illustrating the principle of the half-circumference projection image of the cross section after cutting, translation, and reconstruction in this invention.

[0079] Figure 4 This is a data cleaning diagram illustrating the process of removing outlier information when calculating the truncated mean in this invention.

[0080] Figure 5 This is a schematic diagram of the cutoff point for secondary threshold segmentation calculation using OTSU in this invention;

[0081] Figure 6 This is a schematic diagram illustrating the peak and trough frequencies of the truncated mean curve obtained by using a sliding window in this invention.

[0082] Figure 7 This is a schematic diagram of the projection image generation process of a well at a depth of 4232-4238m in Embodiment 3 of the present invention;

[0083] Figure 8 This is a graph showing the results of quantitative evaluation of laminar flow in a well at a depth of 4232-4238m in Example 3 of this invention. Detailed Implementation

[0084] The present invention will be further described in detail below with reference to embodiments, but the implementation of the present invention is not limited thereto.

[0085] Example 1

[0086] like Figure 1 As shown, this invention provides a method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection, comprising the following steps:

[0087] Step 1: Data preprocessing;

[0088] Step 2: Wellbore interface projection;

[0089] Step 3: Generate the projected image and calculate the truncated mean;

[0090] Step 4: Quantitative statistical analysis of relative laminar density.

[0091] Example 2

[0092] This invention provides a method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection, comprising the following steps:

[0093] 1. Wellbore cross-sectional projection

[0094] An open-hole wellbore can be approximated as a cylinder along the wellbore inclination direction. Sedimentary structures in the formation appear as cross-sections with different dip angles on this cylinder. Unfolding the outer surface of the cylinder along a vertical tangent yields an imaging logging image. The original layered structure is thus stretched into a sinusoidal curve. Due to the complexity of the sinusoidal curve, it is not conducive to the acquisition and evaluation of minute structures such as laminae and fractures. Therefore, the original dynamic and static images are projected onto a vertical cross-section along a fixed tangent. Thus, the laminae appear on the projected image as straight lines with a slope consistent with the dip angle. Figure 2 The dashed lines in the diagram represent angled textures (1 for angled textures and 2 for horizontal textures), maintaining the corresponding thickness. Two approximate semi-circular projection images are generated on both sides of the cross-section (i.e., N~S, S~N). Figure 2 ).

[0095] The projected image can be obtained by calculating the displacement of the data points in the imaging matrix at depth and reconstructing the matrix array according to depth. The total number of columns in the imaging logging image is 360, corresponding to 0 to 360° from the top and bottom surfaces of the cylinder. Therefore, the column number of the matrix data can be equivalent to the angle relative to true north. The formula for calculating the displacement depth of the k-th imaging matrix data point is as follows:

[0096] ΔDepth=r·tan(Dip).cos(k-Azimuth) (1);

[0097]

[0098] Where ΔDepth is the displacement depth of the k-th imaging matrix data point;

[0099] r is the well radius at the current depth;

[0100] Dip is the construction dip angle at the current depth;

[0101] k is the column number to which the current data point belongs;

[0102] Azimuth is the azimuth angle at the current depth;

[0103] C1 and C2 are wellbore curves.

[0104] The new depth of a data point is obtained by adding its original depth to its displacement depth. Then, the data is aligned to the nearest column of the original depth based on the new depth of each data point, thus reconstructing the matrix array indexed by depth.

[0105] 2. Filling in missing values

[0106] During the array reconstruction process, a small number of data points may be missing in certain positions due to errors. Therefore, inverse distance weighted interpolation is used to fill these missing positions. The interpolation steps are as follows:

[0107] (1) Determine the location to be interpolated and the location of known sample points.

[0108] First, search for the coordinates and values ​​of data points within a certain range of the matrix array that are far from the empty space.

[0109] (2) Calculate the weights of the known sample points

[0110] For each missing position, calculate its distance to the nearby sample points found in the search, and convert the distance into a weight using Equation 3. The smaller the distance between the sample point and the missing position, the greater its influence on the interpolation, that is, the weight is inversely proportional to the distance.

[0111]

[0112] Where: w i The weight of the i-th sample point;

[0113] d i This represents the distance between the sample point and the missing location.

[0114] p can be 2 or 3, representing the Euclidean distance or the Manhattan distance, respectively.

[0115] (3) Solving for the weighted average

[0116] After calculating the weight of each sample point, the function values ​​of each sample point are weighted and averaged to obtain the function value at the missing position, as shown in the following formula:

[0117]

[0118] Where z(x, y) is the interpolation result for the missing position;

[0119] The coordinates of the i-th sample point are (x... i y i );

[0120] N is the number of sample points;

[0121] w i The weight of the i-th sample point;

[0122] z i Let be the imaging value of the i-th sample point.

[0123] 3. Calculation of the truncated mean curve

[0124] After the projected image is generated, the array obtained by reading data along a fixed tilt angle is the array vector in the same direction as the striations. To obtain more accurate information, the following processing is also required:

[0125] (1) Cutting, translation and flipping of semi-circular projection image

[0126] After the original image is projected onto the cross section, two semi-circular projection images will be generated. When the instrument is measuring in the well, rotation may occur. The azimuth angle is usually defined as the angle between the No. 1 electrode and true north. The column number of each data column can be considered as the angle with true north. When the interpolation between the column number and the azimuth angle is ±90°, the value of cos(k-Azimuth) is 0. Cutting along this line and shifting the image portion exceeding 90° to the left will generate the desired image. Figure 3 The pair of mirror-symmetric images shown can be cut along the central axis and the right image flipped to obtain two similar projected images. At this time, the textures on the two projected images (dashed lines in 3 in the figure) retain the original electro-imaging values, only stretched in morphology into oblique lines with slopes consistent with the orientation.

[0127] (2) Extraction of tilt angle guidance array

[0128] The electrodes of electro-optical imaging logging instruments like FMI cannot cover all points around the wellbore, leaving gaps that appear as blank bands on dynamic and static images. Previous studies have typically prioritized filling these blank bands using various interpolation methods or deep learning algorithms. However, actual electro-optical imaging dynamic and static images often contain significant noise, and instrument contamination and image distortion caused by changes in the wellbore environment can all interfere with the filling process. Since striations are minute structures, filling errors can significantly interfere with striation identification. Therefore, instead of filling the blank bands, we take values ​​for each row vector along the dip direction, extending beyond the blank bands, to obtain complete striation information.

[0129] (3) Calculation of truncated mean

[0130] In electro-optical imaging, high-conductivity components such as wellbore collapse, conductive minerals, and dissolution cavities have low brightness, while high-resistivity components such as lenses and gravel have high brightness, both of which interfere with the evaluation of laminar flow. Therefore, the concept of truncated mean from statistics is introduced to remove abnormally high-conductivity and abnormally high-resistivity data that interfere with laminar flow information, and then the arithmetic mean of the effective data is calculated, as shown in the following formula:

[0131]

[0132] Where σ is the conductivity value or the pixel value of the dynamic or static image;

[0133] α is the truncation coefficient;

[0134] n is the number of data items;

[0135] m is the number of data points removed.

[0136] The abnormal high resistivity and high conductivity affecting the laminar surface evaluation vary at different depths in electrophysiological imaging images, therefore, data cannot be simply truncated at a fixed percentage. The anomalous portion usually has a clear dividing line with the data portion of the effective reflective laminar surface, and the overall data does not follow a "single main peak" normal distribution, but rather a "multi-peak distribution." Therefore, a threshold segmentation method can be used to find the cutoff values ​​on both sides of the main peak and the abnormal high conductivity and high resistivity portions. The array is then truncated at these cutoff points, and the effective data in the middle is extracted to calculate the average value. Figure 4 (As shown).

[0137] The specific steps are as follows ( Figure 5 As shown in the diagram, firstly, a frequency distribution histogram is plotted on the extracted vector array. Then, the inter-class variance (OTSU) is calculated for all data points. Comparing the calculated OTSU for all data points, the point corresponding to the maximum value is the first cutoff point, which is the cutoff point that maximizes the separation of the original array. The first cutoff point may be the cutoff point with the abnormally high conductance or the abnormally high resistance. Abnormal data outside the cutoff point are truncated, and the remaining array is retained. Then, the second cutoff point is found using OTSU calculation, and the array is truncated again. The final remaining array is the effective array after removing abnormally high and low conductance data. The formula for calculating the maximum inter-class variance is as follows:

[0138]

[0139] Where, ρ 2 The inter-class variance of the data on both sides of this point;

[0140] MG is the global mean;

[0141] P1 represents the data frequency to the left of the current threshold;

[0142] m is the cumulative mean of all thresholds;

[0143] By replacing the original cross-sectional projection image with the truncated mean value and filling it along the dip direction, a smooth cross-sectional projection image is generated, which can intuitively show the development of thicker bedding.

[0144] 4. Calculation of relative laminar density

[0145] The locations of alternating lamellar interfaces typically correspond to points of abrupt changes in electrical properties; correspondingly, these points are reflected in the truncated mean curves of typical characteristic reactions. Displayed as a "jumping peak," the laminae with relatively high conductivity relative to the surrounding rock show a rightward bulge on the curve, i.e., a truncated mean curve. The "peak" of the mean curve; the laminae with relatively high resistivity to the surrounding rock show a leftward dip on the curve, i.e., a truncated mean curve. The "trough" of the curve. The more developed the laminae, the more pronounced the truncated mean curve. The more frequent the fluctuations, the faster the peaks and valleys alternate. Therefore, the density of laminar development can be reflected by quantifying the peak and valley frequencies of the curve.

[0146] like Figure 6 The peaks and troughs of a curve correspond to its extreme points. The three sufficient conditions for determining extreme values ​​are:

[0147] (1) The function is differentiable at this point, and the signs are opposite on both sides;

[0148] (2) The first derivative is zero, and the second derivative is not zero;

[0149] (3) If the lowest non-zero derivative of the function at a point is of an even order, then it is an extreme point.

[0150] Actual well logging curves are scattered points connected at equal sampling intervals. When magnified to a certain scale, they usually appear as a linearly interpolated broken line. Even if the truncated mean curve has the same sampling interval as electrical imaging logging, it is still impossible to accurately calculate the function expression of the curve at certain small fluctuations. Therefore, by taking three sampling points as a unit and one sampling point as the advance step, the logging values ​​of each sampling point are cyclically compared with the logging values ​​of the two adjacent sampling points to find the maximum or minimum point, which is the extreme point of the curve.

[0151] After finding the extreme points using the above method, it is necessary to quantify the peak and valley frequencies within the range. Using the logging depth as an index, a sliding window is set from top to bottom. The window size can be set (here, 1m), and the sliding step size is one sampling interval. The number of peak and valley frequencies appearing in the window after each sliding is counted, corresponding to the center depth of the window. The obtained value is the change in the laminar frequency within the range, reflecting the stability and development of the formation around the wellbore.

[0152] Example 3

[0153] This embodiment uses the quantitative evaluation method of laminae based on the cross-sectional projection of the wellbore by electro-imaging as described in Embodiment 2. The data source is a natural gas well with a well depth of 4232-4238m. The core data shows that the laminae in this section are relatively well-developed.

[0154] 1. Use logging software to generate dynamic electrical imaging maps, automatically pick up formation dip angles, and use linear interpolation to ensure that the depth of each sampling point corresponding to the button electrode has formation dip angle data.

[0155] 2. First, calculate the displacement depth of each button electrode and reconstruct the original image. Figure 7 The pixel matrix of a) is used to fill in missing values ​​to obtain the primary projection image (a). Figure 7 b); then the primary projection image is cut, translated, and reconstructed into a semi-circular projection image that is mirror-symmetrical to the central axis of the wellbore. Figure 7 c); Finally, extract the array of semi-circular projection images along the tilt angle direction, use OTSU to search for the cutoff point, calculate the truncated mean curve, fill the array along the original tilt angle direction, and generate the final cross-sectional projection image. Figure 7 d) Compared to the previous versions of the image, factors that affect the laminar flow indication, such as drill string induced fractures, have been filtered out. Only strip-shaped laminar flows and their original structural dip angles are retained in the image.

[0156] 3. Evaluation results of the laminar flow pattern are as follows: Figure 8 As shown, the first track represents depth; the second and third tracks represent static and dynamic images, respectively; the fourth track is the final cross-sectional projection image; the fifth track is the truncated mean curve; and the last track represents the peak-valley frequency count from the dynamic sliding window (Window = 1m), i.e., the relative densification density of the window, showing relatively high values ​​in the depth ranges of 4233–4234m and 4237–4238m, with the densification distribution consistent with core observations. Simultaneously, the truncated mean image can serve as a reference for core depth correction, improving the accuracy of core positioning in logging depth.

[0157] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications or equivalent changes made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for quantitative evaluation of laminae based on electrical imaging wellbore cross-sectional projection, characterized in that, Includes the following steps: Step 1: Data preprocessing; Step 2: Wellbore interface projection; Step 3: Generate the projected image and calculate the truncated mean; Step 4: Quantitative statistical analysis of relative laminar density.

2. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 1, characterized in that: Step one includes generating an electrical imaging dynamic map and picking up the formation dip angle.

3. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 2, characterized in that: Step two includes displacement depth calculation, pixel matrix reconstruction, and inverse distance weighted interpolation to fill in missing values.

4. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 3, characterized in that: The displacement depth calculation includes the following steps: The wellbore of an open-hole well is approximately a cylinder along the wellbore inclination direction. The sedimentary structures in the formation appear as cross-sections with different dip angles on this cylinder. Unfolding the outer surface of the cylinder along a vertical tangent yields an imaging logging image. The original layered structure is thus stretched into a sinusoidal curve. Projecting the original dynamic and static images along a fixed tangent onto the vertical cross-section, the layers appear as straight lines with the same slope and dip angle, maintaining their corresponding thickness. Two approximate semi-circular projection images are generated on both sides of the cross-section. The projection image is obtained by calculating the displacement of the data points in the imaging matrix at depth and reconstructing the matrix array according to depth. The total number of columns in the imaging logging image is 360, corresponding to 0–360° on the top and bottom surfaces of the cylinder. Therefore, the column number of the matrix data can be equivalent to the angle relative to true north. The formula for calculating the displacement depth of the k-th imaging matrix data point is as follows: ΔDepth=r·tan(Dip)·cos(k-Azimuth) (1); Where ΔDepth is the displacement depth of the k-th imaging matrix data point; r is the well radius at the current depth; Dip is the construction dip angle at the current depth; k is the column number to which the current data point belongs; Azimuth is the azimuth angle at the current depth; C1 and C2 are wellbore curves.

5. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 4, characterized in that: In step two, the original depth of the data point is added to the displacement depth to obtain the new depth of the data point. Then, based on the new depth of each data point, the data is aligned to the nearest original depth column to realize the reconstruction of the matrix array indexed by depth.

6. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 5, characterized in that: During the array reconstruction process, a small number of data points may be missing in certain positions due to errors. Therefore, inverse distance weighted interpolation is used to fill in the missing positions in the matrix.

7. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 6, characterized in that: The interpolation includes the following steps: (1) Determine the location to be interpolated and the location of known sample points; (2) Calculate the weights of the known sample points; (3) Solve for the weighted average.

8. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 7, characterized in that: First, search for the coordinates and values ​​of data points within a certain range of the matrix array that are far from the empty space; For each missing position, calculate its distance to the nearby sample points found in the search, and convert the distance into a weight using Equation 3. The smaller the distance between the sample point and the missing position, the greater the impact on the interpolation. That is, the weight is inversely proportional to the distance. Where: w i The weight of the i-th sample point; d i This represents the distance between the sample point and the missing location. p can be 2 or 3, representing the Euclidean distance or the Manhattan distance, respectively. After calculating the weight of each sample point, the function values ​​of each sample point are weighted and averaged to obtain the function value at the missing position, as shown in the following formula: Where z(x, y) is the interpolation result for the missing position; The coordinates of the i-th sample point are (x... i y i ); N is the number of sample points; w i The weight of the i-th sample point; z i Let be the imaging value of the i-th sample point.

9. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 8, characterized in that: In step three, after the projected image is generated, the array obtained by reading data along a fixed tilt angle is the array vector in the same direction as the texture.

10. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 9, characterized in that: Specifically, the following steps are included: (1) Cutting, translation and flipping of semi-circular projection images; (2) Extraction of tilt angle guidance array; (3) Calculation of truncated mean.

11. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 10, characterized in that: After the original image is projected onto the cross section, two semi-circular projection images are generated. The azimuth angle is the angle between the No. 1 electrode and the due north direction. The column number of each data column can be regarded as the angle with due north. When the interpolation of the column number and the azimuth angle is ±90°, the value of cos(k-Azimuth) is 0. Cutting along this value line and translating the image part exceeding 90° to the left side can generate a pair of mirror-symmetric images. Cutting the image along the central axis and flipping the right image can obtain two similar projection images. The texture on the two projection images retains the original electro-imaging values, only stretched in morphology into oblique lines with the same slope and attitude. The electrodes of the electrical imaging logging instrument cannot cover all points around the well, and there are gaps between the electrodes, which are displayed as blank stripes on dynamic and static maps. By taking the value of each row vector along the dip direction across the blank strip, all information of the covering texture can be obtained. After removing the abnormally high conductivity and abnormally high resistivity portions of the interfering stratum information, the arithmetic mean of the valid data portion is calculated using the following formula: Where σ is the conductivity value or the pixel value of the dynamic or static image; α is the truncation coefficient; n is the number of data items; m is the number of data points removed.

12. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 11, characterized in that: The cutoff values ​​of the abnormally high conductance and abnormally high resistance parts on both sides of the main peak are found by threshold segmentation. The array is truncated at the cutoff point, and the effective data in the middle is extracted to calculate the average value. First, the frequency distribution histogram of the extracted vector array is statistically plotted. Then, calculate the inter-class variance at all data points using OTSU. Compare the inter-class variance calculated at all data points. The point corresponding to the maximum value is the first cutoff point, which is the cutoff point that maximizes the separation of the original array. The first cutoff point is either the cutoff point with the abnormally high-conductivity part or the cutoff point with the abnormally high-resistivity part. Outlier data outside the cutoff point is truncated, and the remaining array is retained. Then, the second cutoff point is found through OTSU calculation, and the array is truncated again. The final remaining array is the effective array after removing outlier high and low conductance. The formula for calculating the maximum inter-class variance is as follows: Where, ρ 2 The inter-class variance of the data on both sides of this point; MG is the global mean; P1 represents the data frequency to the left of the current threshold; m is the cumulative average of all threshold values; By replacing the values ​​of the original cross-sectional projection image with the truncated mean and filling along the dip direction, a smooth cross-sectional projection image is generated, which can intuitively show the development of thicker bedding.

13. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 12, characterized in that: The locations of the alternating lamellar interfaces correspond to abrupt changes in electrical properties, and correspondingly, these changes are reflected in the truncated mean curves of typical characteristic reactions. The curve above shows a "jumping peak," where the laminae with relatively high conductivity relative to the surrounding rock bulge to the right, i.e., a truncated mean curve. The "peak" of the mean curve; the laminae with relatively high resistivity to the surrounding rock show a leftward dip on the curve, i.e., a truncated mean curve. The "trough"; the more developed the laminae, the more pronounced the truncated mean curve. The more frequent the fluctuations, the faster the peaks and valleys alternate. Therefore, the density of laminar development can be reflected by quantifying the peak and valley frequencies of the curve.

14. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 13, characterized in that: The peaks and troughs of a curve correspond to its extreme points. The three sufficient conditions for determining extreme points are: (1) The function is differentiable at this point, and the signs are opposite on both sides; (2) The first derivative is zero, and the second derivative is not zero; (3) If the lowest non-zero derivative of the function at a point is of an even order, then it is an extreme point.

15. The method for quantitative evaluation of laminae based on electro-imaging wellbore cross-sectional projection according to claim 14, characterized in that: The actual logging curve is a scatter line with equal sampling intervals. When magnified to a certain scale, it usually appears as a linearly interpolated broken line. Even if the truncated mean curve has the same sampling interval as electrical imaging logging, three sampling points are taken as a unit, and one sampling point is taken as the advance step. The logging values ​​of each sampling point are compared with the two adjacent sampling points in a loop to find the maximum or minimum point, which is the extreme point of the curve. After finding the extreme point using the above method, it is necessary to quantify the peak and valley frequencies within the range. Using the logging depth as an index, a sliding window is set from top to bottom. The window size can be set, and the sliding step size is one sampling interval. The number of peaks and valleys appearing in the window after each sliding is counted, corresponding to the center depth of the window. The obtained value is the change in the laminar frequency within the range, reflecting the stable development of the formation around the wellbore.

Images