A method and apparatus for determining full borehole high resolution logging data

Abstract

The specification provides a method and device for determining full-bore high-resolution logging data. The method comprises: obtaining open hole logging data and high-resolution core logging data; inputting the open hole logging data and the high-resolution core logging data into a high-resolution logging data model to obtain full-bore high-resolution logging data; wherein the high-resolution logging data model is trained in the following manner: obtaining target open hole logging data samples and target high-resolution core logging data samples; decomposing the target high-resolution core logging data samples to obtain target smoothed data and residual data; removing deviated data from the target open hole logging data samples and the target smoothed data; processing a multi-level network model based on the target open hole logging data samples after removing the deviated data, the target smoothed data and the residual data to obtain the high-resolution logging data model. Based on the above method, full-bore high-resolution logging data can be accurately and quickly obtained.

Description

Technical Field

[0001] This specification relates to the field of petroleum exploration and development technology, and in particular to a method and apparatus for determining high-resolution logging data across the entire well section. Background Technology

[0002] Shale oil and gas reserves are abundant and have become an important alternative resource for oil. The "horizontal well + hydraulic fracturing" technology is a key technology for the development of unconventional shale oil and gas. The artificial fracture network morphology and other parameters formed by hydraulic fracturing simulation are the core of formulating a reasonable development plan for shale oil and gas.

[0003] The key technical aspect of hydraulic fracturing simulation is establishing a reservoir geomechanical model. Traditional methods rely on conventional open-hole logging data, such as dipole sonic logging, to build this model, then use fracturing simulation to understand fracture growth patterns and morphologies. However, this method has significant limitations. It cannot accurately identify and characterize highly developed mechanically weak surfaces and their properties in shale formations. The resulting reservoir geomechanical model cannot reflect the impact of these weak surfaces on the longitudinal propagation of artificial fractures. The artificial fracture morphology obtained through hydraulic fracturing simulation may differ significantly from the actual underground conditions, thus hindering the development of effective shale oil and gas development plans. With the development of core logging technology, high-resolution core logging instruments can obtain centimeter-level (approximately 1 cm) logging curve data. However, due to the high cost of core sampling and the difficulty of coring the entire well section or multiple wells, it is impossible to establish a high-precision reservoir geomechanical model, resulting in low exploration and development efficiency for shale oil and gas. Existing methods cannot determine high-resolution logging data for the entire well section quickly and easily at a low cost, thus hindering the development of reasonable shale oil and gas development plans and leading to low shale oil and gas development efficiency.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This specification provides a method and apparatus for determining high-resolution logging data for the entire well section, in order to solve the problem that existing technologies cannot accurately, quickly, and cost-effectively determine high-resolution logging data for the entire well section, resulting in low efficiency in shale oil and gas development.

[0006] On the one hand, the embodiments of this specification provide a method for determining high-resolution logging data for the entire well section, including:

[0007] Acquire open-hole logging data and high-resolution core logging data;

[0008] The open-hole logging data and high-resolution core logging data are input into the high-resolution logging data model to obtain high-resolution logging data for the entire well section;

[0009] The high-resolution logging data model is trained as follows: acquiring target open-hole logging data samples and target high-resolution core logging data samples; decomposing the target high-resolution core logging data samples to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; removing deviation data from the target open-hole logging data samples and target smoothed data; and processing a multi-level network model based on the target open-hole logging data samples, target smoothed data, and residual data after removing deviation data to obtain the high-resolution logging data model.

[0010] Furthermore, the method also includes:

[0011] Obtain initial open-hole logging data samples;

[0012] The initial open-hole logging data sample is preprocessed to determine whether the preprocessed initial open-hole logging data sample is complete.

[0013] If not, then obtain the logging data sample of the target adjacent well, and perform missing reconstruction on the preprocessed initial open-hole logging data sample based on the logging data sample of the target adjacent well to obtain the target open-hole logging data sample.

[0014] Furthermore, the target open-hole logging data sample includes at least one of the following: natural gamma, density, spontaneous potential, borehole diameter, resistivity, P-wave transit time, S-wave transit time, and compensated neutrons.

[0015] Further, the decomposition of the target high-resolution core logging data sample to obtain the target smoothed data and residual data corresponding to the target high-resolution core logging data sample includes:

[0016] Obtain an initial sliding window and continuously adjust the initial sliding window at preset intervals;

[0017] Based on the initial sliding window before adjustment and the initial sliding window after adjustment, multiple sets of high-resolution core logging data samples with different smoothness are determined in sequence.

[0018] The target smoothed data is determined by comparing the target high-resolution core logging data sample with the high-resolution core logging data sample with different smoothing degrees.

[0019] The residual data is determined based on the target high-resolution core logging data sample and the target smoothed data.

[0020] Further, the comparison of the target high-resolution core logging data sample with the high-resolution core logging data samples with different smoothing degrees to determine the target smoothed data includes:

[0021] The relative errors between the target high-resolution core logging data sample and the high-resolution core logging data samples with different smoothness levels are calculated sequentially.

[0022] By comparing the relative error, when the relative error is less than a preset error threshold, the target smoothed data is determined from high-resolution core logging data samples with different smoothness levels.

[0023] Furthermore, the removal of deviation data from the target open-hole logging data samples and the target smoothed data includes:

[0024] Calculate the difference ratio between the target open-hole logging data sample and the target smoothed data;

[0025] Based on the difference ratio data, determine the standard deviation corresponding to the difference ratio data;

[0026] Based on the standard deviation, remove the off-target data from the target open-hole logging data sample and the target smoothed data.

[0027] Furthermore, the multi-level network model includes a trend prediction module and a residual prediction module. Correspondingly, the multi-level network model, processed based on the target open-hole logging data sample after removing deviation data, the target smoothed data, and the residual data, yields a high-resolution logging data model, including:

[0028] The target open-hole logging data sample after removing the deviation data and the target smoothed data are input into the trend prediction module to obtain the trend prediction data sample for the entire well section;

[0029] The target open-hole logging data sample and residual data after removing the deviation data are input into the residual prediction module to obtain the residual prediction data sample for the entire well section.

[0030] Based on the trend prediction data sample and the residual prediction data sample of the entire well section, a high-resolution logging data sample of the entire well section is obtained.

[0031] The high-resolution logging data sample of the entire well section is compared and verified with the target high-resolution core logging data sample. When the comparison and verification results meet the preset accuracy requirements, a high-resolution logging data model is obtained.

[0032] On the other hand, embodiments of this specification also provide a device for determining high-resolution logging data for the entire well section, including:

[0033] The acquisition module is used to acquire open-hole logging data and high-resolution core logging data;

[0034] The prediction module is used to input the open-hole logging data and high-resolution core logging data into the high-resolution logging data model to obtain high-resolution logging data for the entire well section.

[0035] The model training module is used to acquire target open-hole logging data samples and target high-resolution core logging data samples; decompose the target high-resolution core logging data samples to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; remove deviation data from the target open-hole logging data samples and target smoothed data; and process a multi-level network model based on the target open-hole logging data samples, target smoothed data and residual data after removing deviation data to obtain a high-resolution logging data model.

[0036] In another aspect, this application also provides an apparatus including a processor and a memory for storing processor-executable instructions, wherein the processor executes the instructions to implement the method for determining high-resolution logging data for the entire well section in the above embodiments.

[0037] In another aspect, this application also provides a computer-readable storage medium storing computer instructions thereon, wherein the computer-readable storage medium executes the instructions to implement the method for determining high-resolution logging data for the entire well section in the above embodiments.

[0038] This specification provides a method and apparatus for determining high-resolution logging data for the entire well section. First, open-hole logging data and high-resolution core logging data are acquired. Second, the open-hole logging data and high-resolution core logging data are input into a high-resolution logging data model to obtain high-resolution logging data for the entire well section. The high-resolution logging data model is trained as follows: a target open-hole logging data sample and a target high-resolution core logging data sample are acquired; the target high-resolution core logging data sample is decomposed to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data sample; deviation data is removed from the target open-hole logging data sample and the target smoothed data; a multi-level network model is processed based on the target open-hole logging data sample, target smoothed data, and residual data after removing deviation data to obtain the high-resolution logging data model. In the embodiments of this specification, by acquiring the target open-hole logging data sample and the target high-resolution core logging data sample, the utilization rate and value of the core samples in the target open-hole logging data sample and the target high-resolution core logging data sample can be effectively improved. By decomposing the target high-resolution core logging data samples and removing deviation data from the target smoothed data and target open-hole logging data samples obtained from the decomposition, the training accuracy and training effect of the model can be improved. By processing a multi-level network model based on the target open-hole logging data samples after deviation data removal, the target smoothed data, and the residual data, a high-quality high-resolution logging data model can be trained. This high-resolution logging data model can then be used to accurately and quickly predict high-resolution logging data for the entire well section, compensating for the difficulty in obtaining high-resolution logging data from non-core sections or non-cored wells. Obtaining high-resolution logging data for the entire well section lays the foundation for subsequent high-precision reservoir mechanics models and the rational optimization of fracturing schemes. Attached Figure Description

[0039] To more clearly illustrate the embodiments of this specification, the accompanying drawings used in the embodiments will be briefly introduced below. The drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 These are mechanically weak surface diagrams of lithological interfaces and bedding in shale reservoirs provided in the embodiments of this specification. (a) and (b) are relatively homogeneous shale with millimeter-level bedding; (c) is a thin layer of fine calcareous silt interbedded with shale, with a thickness of 2-3 mm; (d) is homogeneous mudstone; (e) is liquefied siltstone, with silt veins sinking into the mudstone and shale, forming a complex sand-mud mixture; (f) is a thin layer of fine calcareous silt interbedded with shale, at the bottom of the first section; (g) is a thin layer of gray fine sandstone with calcareous mudstone in the middle.

[0041] Figure 2 This is a flowchart illustrating a method for determining high-resolution logging data across the entire well section, as provided in the embodiments of this specification.

[0042] Figure 3 This is a sliding window determination diagram provided in the embodiments of this specification;

[0043] Figure 4 This is a graph showing the change of the trend line under different windows provided in the embodiments of this specification;

[0044] Figure 5 This is a high-resolution logging curve decomposition diagram of the core section provided in the embodiments of this specification;

[0045] Figure 6 This is a schematic diagram of the removal deviation provided in the embodiments of this specification;

[0046] Figure 7 This is a schematic diagram of the deviation area between the conventional open-hole logging curve and the high-resolution logging curve of the core section provided in the embodiments of this specification;

[0047] Figure 8 This is a schematic diagram of the model training results provided in the embodiments of this specification;

[0048] Figure 9 This is a schematic diagram illustrating the verification of model results provided in the embodiments of this specification;

[0049] Figure 10 This is a schematic diagram of the structural composition of a device for determining high-resolution logging data across the entire well section, as provided in the embodiments of this specification.

[0050] Figure 11 This is a schematic diagram of the structural composition of an electronic device provided in one embodiment of this specification. Detailed Implementation

[0051] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0052] Shale oil and gas reserves are abundant and have become an important alternative resource for oil. "Horizontal wells + hydraulic fracturing" is a key technology for the development of unconventional shale oil and gas. Parameters such as the artificial fracture network morphology formed by hydraulic fracturing, including fracture height, fracture length, fracture complexity, proppant distribution, and inter-section and inter-well fracturing interference, are core to formulating a rational shale oil and gas development plan. The key technical point of hydraulic fracturing simulation is establishing a reservoir geomechanical model. Unconventional shale formations exhibit rapid vertical lithological changes, strong heterogeneity, and highly developed mechanically weak surfaces such as lithological interfaces, bedding, or foliation. Figure 1 As shown, artificial fractures, in addition to being limited by high-stress shielding layers during vertical propagation, are also prone to asymmetrical propagation and turning along horizontally weak mechanical surfaces, thus severely restricting their vertical extension. Current methods rely on conventional open-hole logging data such as dipole sonic logging to establish geomechanical models, and then use fracturing simulations to understand fracture growth patterns and morphologies. However, this method has significant limitations. The main reason is that the resolution of open-hole logging instruments is only tens of centimeters, making it impossible to precisely identify and characterize the highly developed weak mechanical surfaces and their properties in shale formations. Consequently, the reservoir geomechanical models established in this way cannot reflect the influence of these weak mechanical surfaces on the vertical propagation of artificial fractures. The artificial fracture morphology obtained through fracturing simulations may differ significantly from the actual underground conditions, making it difficult to accurately guide the optimization design of fracturing schemes.

[0053] With the development of core logging technology, high-resolution core logging instruments can now obtain logging curve data at the centimeter level (about 1cm). However, due to the high cost of core sampling, it is very difficult to achieve core sampling of the entire well section and multiple wells, which makes it impossible to establish a high-precision reservoir geomechanical model, resulting in low exploration and development efficiency of shale oil and gas.

[0054] The current methods are unable to obtain high-resolution logging data for the entire well section accurately and efficiently at a low cost, thus hindering the development of reasonable shale oil and gas development plans and resulting in low shale oil and gas development efficiency.

[0055] To address the aforementioned problems with existing methods, this specification introduces a method and apparatus for determining high-resolution logging data across the entire well section. This method can accurately and quickly acquire high-resolution logging data across the entire well section, thereby enabling the establishment of a high-precision reservoir mechanics model and laying the foundation for the rational optimization of fracturing schemes.

[0056] Based on the above approach, this specification proposes a method and apparatus for determining high-resolution logging data across the entire well section. First, open-hole logging data and high-resolution core logging data are acquired. Then, the open-hole logging data and high-resolution core logging data are input into a high-resolution logging data model to obtain high-resolution logging data for the entire well section. The high-resolution logging data model is trained as follows: target open-hole logging data samples and target high-resolution core logging data samples are acquired; the target high-resolution core logging data samples are decomposed to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; deviation data is removed from the target open-hole logging data samples and target smoothed data; a multi-level network model is processed based on the target open-hole logging data samples, target smoothed data, and residual data after removing deviation data to obtain the high-resolution logging data model. (See reference...) Figure 2 As shown in the embodiments of this specification, a method for determining high-resolution logging data for the entire well section is provided. In specific implementation, this method may include the following:

[0057] S201: Acquire open-hole logging data and high-resolution core logging data.

[0058] In some embodiments, the aforementioned open-hole logging data can be low-resolution conventional open-hole logging data for the entire well section (including core and non-core sections) obtained through open-hole logging, while the aforementioned high-resolution core logging data can be high-resolution logging data for the core section obtained using a high-resolution core logging instrument. The well measured by the open-hole logging and the high-resolution core logging instrument can be the same well, which can be designated as the target well. By acquiring open-hole logging data and high-resolution core logging data from the same well, the model can learn the correlation between the open-hole logging data and the high-resolution core logging data within the core section of the same well, laying the foundation for subsequently obtaining high-resolution logging data for the entire well section. The aforementioned open-hole logging data and high-resolution core logging data can include conventional open-hole logging curves (which can be represented by data points in 0.125m increments) and high-resolution core logging curve data, or high-resolution logging curves for the core section (which can be represented by data points in 0.01m increments).

[0059] In some embodiments, the conventional open-hole logging data described above may include at least one of the following: natural gamma ray (GR), density (RHOB), spontaneous potential (SP), diameter (CALI), resistivity (RT), P-wave transit time (DTCO), S-wave transit time (DTSM), and compensated neutron (CNL). The high-resolution logging curves for the core section described above may include at least one of the following: P-wave transit time (DTCO), natural gamma ray (GR), and density (RHOB). Since the open-hole logging data includes the open-hole logging data for the core section, it correlates with the high-resolution core logging data (or high-resolution logging data for the core section) within the core section. Modeling can be based on this correlation to ultimately predict the high-resolution logging data for the entire well section (prediction of high-resolution logging data for the core section plus prediction of high-resolution logging data for non-core sections). The prediction of high-resolution logging data for the entire well section will be explained separately later and will not be repeated here.

[0060] In some embodiments, the aforementioned full well section can be the depth segment from the surface to the bottom of the well, and the core segment, or coring segment, can be the depth segment from which the core is brought to the surface for coring. The aforementioned high-resolution core logging instrument obtains high-resolution core logging data by measuring the core in a laboratory, while open-hole logging involves running logging tools after drilling to measure and obtain conventional open-hole logging data. Therefore, the logging data obtained by the high-resolution core logging instrument and open-hole logging in the core segment may have discrepancies. Further processing can be performed on the acquired data to reduce these discrepancies. For example, preprocessing and missing data reconstruction can be performed on the acquired open-hole logging data to obtain complete open-hole logging data. Depth matching can be performed on the high-resolution core logging data based on the open-hole logging data. By preprocessing and missing data reconstruction on the open-hole logging data, and performing depth matching on the high-resolution core logging data, a data foundation can be laid for subsequently outputting more accurate high-resolution core logging data for the entire well section.

[0061] S202: Input the open-hole logging data and high-resolution core logging data into the high-resolution logging data model to obtain high-resolution logging data for the entire well section.

[0062] In some embodiments, during practical applications, the acquired open-hole logging data and high-resolution core logging data can be input into a pre-trained high-resolution logging data model to obtain high-resolution logging data for the entire well section. Predicting high-resolution logging data for the entire well section based on conventional open-hole logging data and high-resolution core logging data can effectively improve the utilization rate and value of core samples, compensating for the difficulty in obtaining high-resolution logging data in non-core sections or non-core wells, and providing a data foundation for establishing a more reliable reservoir geomechanical model. Obtaining high-resolution logging data for the entire well section provides a data foundation for establishing a more accurate and reliable reservoir geomechanical model, thereby enabling more rational simulation analysis of hydraulic fracture propagation, and thus enabling the rational and effective formulation of shale oil and gas development plans, effectively guiding the exploration and development of shale oil and gas.

[0063] In some embodiments, after obtaining high-resolution logging data for the entire well section, the method further includes:

[0064] Based on high-resolution logging data from the entire well section, a reservoir geomechanical model was determined.

[0065] The morphology of artificial fractures was obtained based on the reservoir geomechanical model;

[0066] Based on the morphology of artificial seams, a shale oil and gas development plan is determined, which then guides shale oil and gas development.

[0067] In some embodiments, high-resolution logging data from the entire well section can be used as training data to train a high-precision reservoir geomechanical model. Then, based on this high-precision model, hydraulic fracture propagation simulation analysis can be conducted to obtain parameters such as the artificial fracture network morphology formed by the hydraulic fracturing simulation. Finally, based on these parameters, a reasonable shale oil and gas development plan can be formulated, guiding shale oil and gas development.

[0068] S203: The high-resolution logging data model is trained and obtained in the following manner: acquiring target open-hole logging data samples and target high-resolution core logging data samples; decomposing the target high-resolution core logging data samples to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; removing deviation data from the target open-hole logging data samples and target smoothed data; and processing a multi-level network model based on the target open-hole logging data samples, target smoothed data, and residual data after removing deviation data to obtain the high-resolution logging data model.

[0069] In some embodiments, target open-hole logging data samples and target high-resolution core logging data samples can be acquired. Then, based on machine learning (e.g., multi-level network models), the interaction relationship between the target open-hole logging data samples and the target high-resolution core logging data samples is continuously learned, ultimately training a high-resolution logging data model. This high-resolution logging data model can accurately, quickly, and cost-effectively predict high-resolution logging data for the entire well section (including high-resolution logging data for the core section and high-resolution logging data for non-core sections), solving the problem of high costs and difficulty in accurately and quickly acquiring high-resolution logging data for non-core sections in existing core sampling methods.

[0070] In some embodiments, the target open-hole logging data sample can be preprocessed and then reconstructed from missing data (open-hole logging curve). In specific implementation, the process of acquiring the target open-hole logging data sample may include:

[0071] Obtain initial open-hole logging data samples;

[0072] The initial open-hole logging data sample is preprocessed, and it is determined whether the preprocessed initial open-hole logging data sample is complete.

[0073] If not, then obtain the logging data sample of the target adjacent well, and perform missing reconstruction on the preprocessed initial open-hole logging data sample based on the logging data sample of the target adjacent well to obtain the target open-hole logging data sample.

[0074] In some embodiments, the initial open-hole logging data sample can be the raw open-hole logging data without any processing. Due to the influence of the skill level of the data acquisition personnel and the large time span of the raw open-hole logging data, the raw open-hole logging data has certain quality problems. Therefore, before establishing a high-resolution logging data model, it is necessary to preprocess the initial open-hole logging data sample (i.e., perform professional data analysis and data cleaning). The preprocessing can include numerical error analysis, outlier analysis, data missing, depth mismatch analysis, identification, and supplementation. For example, a well is usually logged several times, and the open-hole logging data (open-hole logging curves) obtained from each measurement are different. Between these open-hole logging data (open-hole logging curve) sequences, depth misalignment can occur due to issues such as the length of the measuring instrument, debugging, or operation. It is necessary to correct the depth of the open-hole logging data (open-hole logging curves). Some open-hole logging data (open-hole logging curves) show anomalies at certain depths, which need to be corrected through appropriate professional technical means.

[0075] In some embodiments, after the initial open-hole logging data sample has undergone the preprocessing described above, it can be determined whether the preprocessed initial open-hole logging data sample is complete, i.e., whether it is necessary to perform missing reconstruction on the initial open-hole logging data sample. If it is determined that the initial open-hole logging data sample is incomplete, it is necessary to further obtain logging data samples from the target adjacent well (which can be the adjacent well with complete logging data next to the target well as the target adjacent well). Then, based on the logging data samples from the target adjacent well, missing reconstruction is performed on the preprocessed initial open-hole logging data sample to obtain a complete target open-hole logging data sample. Machine learning techniques such as cross-plotting, multiple linear regression, and neural networks can be used to perform missing reconstruction on the initial open-hole logging data to ultimately obtain a complete target open-hole logging data sample. It should be noted that the above-described missing reconstruction methods are not limited to the examples described above. Those skilled in the art may make other changes based on the technical essence of the embodiments in this specification, but as long as the functions and effects achieved are the same as or similar to those in the embodiments of this specification, they should all be covered within the protection scope of the embodiments of this specification.

[0076] In some embodiments, the target open-hole logging data samples may include at least one of the following: natural gamma (GR), density (RHOB), spontaneous potential (SP), diameter (CALI), resistivity (RT), P-wave transit time (DTCO), S-wave transit time (DTSM), and compensated neutron data (CNL). Specifically, the more important data can be selected from the target open-hole logging data to train the model; for example, natural gamma (GR), density (RHOB), P-wave transit time (DTCO), and S-wave transit time (DTSM) can be selected for model training.

[0077] In some embodiments, high-resolution core logging curves are obtained on the surface using appropriate logging instruments. Specifically, the core is first drilled out using specialized tools, and then the high-resolution core logging instrument is used to measure the drilled core to obtain the high-resolution logging curve (or high-resolution core logging curve) of the core segment. During core drilling, rock fracturing or changes in the temperature and pressure system can cause the dissolution of specific components (such as salt or gypsum), resulting in errors in the core depth. Conventional open-hole logging curves, on the other hand, are obtained after drilling by running logging tools, and the corresponding depth is the actual depth. Therefore, the depth of the high-resolution core segment obtained after high-resolution logging of the drilled core in the laboratory will have deviations. It is necessary to compare the curve morphology of the high-resolution core segment with that of the conventional open-hole logging curve to correct the depth of the high-resolution core segment.

[0078] In some embodiments, the aforementioned target high-resolution core logging data sample can be obtained by: first acquiring a high-resolution core logging data sample, then performing depth matching between the high-resolution core logging data sample and the target open-hole logging data sample, ultimately obtaining the depth-matched target high-resolution core logging data sample. Obtaining the depth-matched target high-resolution core logging data sample can lay the foundation for subsequently establishing a high-precision and reliable high-resolution logging data model.

[0079] In some embodiments, after determining the target high-resolution core logging data sample, the target high-resolution core logging data sample can be decomposed to obtain target smoothing data and residual data corresponding to the target high-resolution core logging data sample. The target smoothing data may include the trend line of the target high-resolution core logging data sample and the residual line of the target high-resolution core logging data sample.

[0080] In some embodiments, the above-mentioned target high-resolution core logging data sample is decomposed to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data sample, which may include:

[0081] Obtain an initial sliding window and continuously adjust the initial sliding window at preset intervals;

[0082] Based on the initial sliding window before adjustment and the initial sliding window after adjustment, multiple sets of high-resolution core logging data samples with different smoothness are determined in sequence.

[0083] The target smoothed data is determined by comparing the target high-resolution core logging data sample with the high-resolution core logging data sample with different smoothing degrees.

[0084] The residual data is determined based on the target high-resolution core logging data sample and the target smoothed data.

[0085] In some embodiments, the target smoothed data (i.e., decomposing a trend line from the target high-resolution core logging data sample) can be determined using moving average techniques. Specifically, an initial sliding window (e.g., set to 0.02m) can be obtained first. The initial sliding window can be understood as the data point interval in the target high-resolution core logging data sample. Then, the initial sliding window can be continuously adjusted at preset intervals (e.g., 0.01m). Each adjustment of the initial sliding window yields the corresponding high-resolution core logging data sample. After adjusting the initial sliding window, multiple sets of high-resolution core logging data samples with different smoothing degrees under different sliding windows can be obtained. Finally, the target high-resolution core logging data sample can be compared sequentially with the multiple sets of high-resolution core logging data samples with different smoothing degrees to determine the target smoothed data.

[0086] In some embodiments, the above comparison of the target high-resolution core logging data sample and the high-resolution core logging data samples with different smoothing degrees to determine the target smoothed data may, in specific implementations, include:

[0087] The relative errors between the target high-resolution core logging data sample and the high-resolution core logging data samples with different smoothness levels are calculated sequentially.

[0088] By comparing the relative error, when the relative error is less than a preset error threshold, the target smoothed data is determined from high-resolution core logging data samples with different smoothness levels.

[0089] In some embodiments, the relative error described above can be calculated using the following formula:

[0090]

[0091] in, The relative error is given by n, where n is the number of data points. For high-resolution core logging data samples, These are high-resolution core logging data samples with varying degrees of smoothness.

[0092] In some embodiments, after calculating the relative error between the target high-resolution core logging data sample and high-resolution core logging data samples with different smoothness levels, the calculated relative error can be compared with a preset error threshold. When the relative error is less than the preset error threshold, the target smooth data can be determined from the high-resolution core logging data samples with different smoothness levels. This ensures that the obtained target smooth data is both closest to the target high-resolution core logging data sample and relatively smooth. At this time, the target smooth data corresponds to a most suitable target sliding window. That is, when the relative error is less than the preset error, the target sliding window can be determined (this target sliding window is obtained by adjusting the initial sliding window and then filtering from the adjusted initial sliding window). This target sliding window can find a balance between smoothness and information availability.

[0093] In some embodiments, after the target smoothed data is determined, the target high-resolution core logging data sample and the target smoothed data can be subtracted, and the result of the subtraction can be used as the residual data.

[0094] In some embodiments, the removal of deviation data from the target open-hole logging data sample and the target smoothed data described above may, in specific implementation, include:

[0095] Calculate the difference ratio between the target open-hole logging data sample and the target smoothed data;

[0096] Based on the difference ratio data, determine the standard deviation corresponding to the difference ratio data;

[0097] Based on the standard deviation, remove the off-target data from the target open-hole logging data sample and the target smoothed data.

[0098] In some embodiments, after determining the target smoothing data, the difference ratio between the target open-hole logging data sample and the target smoothing data can be calculated according to the following formula:

[0099]

[0100] Where θ is the difference ratio between the target open-hole logging data sample and the target smoothed data. The target smoothed data is obtained by decomposing the target high-resolution core logging data sample. This is a sample of naked-eye logging data for the target.

[0101] After identifying the discrepancies, the standard deviation can be calculated based on these discrepancies. Then, based on the calculated standard deviation, the deviation data in the target open-hole logging data sample and the target smoothed data are removed. For example, the standard deviation can be... Data outside the target open-hole logging data sample is considered as deviation data, and then the deviation data is removed. By removing the deviation data from the target open-hole logging data sample and the target smoothed data, and then using the deviation-removed target open-hole logging data sample and the target smoothed data for model training, the accuracy and effectiveness of model training can be improved.

[0102] In some embodiments, the multi-level network model described above may include a trend prediction module and a residual prediction module. Accordingly, the multi-level network model, which processes the target open-hole logging data sample, target smoothed data, and residual data after removing deviation data to obtain a high-resolution logging data model, may, in specific implementations, include:

[0103] The target open-hole logging data sample after removing the deviation data and the target smoothed data are input into the trend prediction module to obtain the trend prediction data sample for the entire well section;

[0104] The target open-hole logging data sample and residual data after removing the deviation data are input into the residual prediction module to obtain the residual prediction data sample for the entire well section.

[0105] Based on the trend prediction data sample and the residual prediction data sample of the entire well section, a high-resolution logging data sample of the entire well section is obtained.

[0106] The high-resolution logging data sample of the entire well section is compared and verified with the target high-resolution core logging data sample. When the comparison and verification results meet the preset accuracy requirements, a high-resolution logging data model is obtained.

[0107] In some embodiments, since high-resolution core logging datasets are generally relatively small, most machine learning models, such as neural network models, require a large dataset to train a high-quality model. In this embodiment, the prediction module can employ the Kernel Ridge Regression (KRR) algorithm, and the residual prediction module can employ the Out-of-Tree (ET) algorithm, thus achieving the goal of training a high-quality model with a smaller dataset. Specifically, the KRR algorithm can learn and predict trend prediction data samples for the entire well section, and the ET algorithm can learn and predict residual prediction data samples for the entire well section. Finally, the predicted trend prediction data samples and the predicted residual prediction data samples for the entire well section can be recombined to obtain the high-resolution logging data samples for the entire well section.

[0108] In some embodiments, after obtaining high-resolution logging data samples for the entire well section, the high-resolution logging data samples of the core section within the high-resolution logging data samples for the entire well section can be compared and verified with the target high-resolution core logging data samples. When the comparison and verification results meet the preset accuracy requirements, a high-resolution logging data model can be obtained. Otherwise, it indicates that the current model training error is too large, and it is necessary to continue training the multi-level network model based on the target open-hole logging data samples after removing the deviation data, the target smoothed data, and the residual data. That is, the trend prediction data samples for the entire well section are re-predicted based on the kernel ridge regression (KRR) algorithm, and the trend prediction data samples for the entire well section are re-predicted based on the out-of-tree (ET) algorithm using the residual prediction module, and the high-resolution logging data samples for the entire well section are obtained again. Training stops when the verification results meet the preset accuracy requirements or the number of model iterations reaches the preset number of iterations.

[0109] By predicting high-resolution logging data for the entire well section based on open-hole logging data and high-resolution core logging data, the utilization rate and value of the core can be maximized, making up for the difficulty in obtaining high-resolution logging data in non-core sections or non-core wells, and providing a data foundation for establishing a more reliable reservoir geomechanical model.

[0110] The above method will be described below with reference to a specific embodiment. However, it is worth noting that this specific embodiment is only for better illustration of this application and does not constitute an improper limitation of this application.

[0111] Before implementation, conventional open-hole logging data and high-resolution core logging data are first acquired. Then, the conventional open-hole logging data undergoes data analysis and cleaning, including missing curve reconstruction. This data is then combined with the high-resolution core logging data for depth matching, ultimately yielding complete target open-hole logging data and target high-resolution core logging data. After obtaining the target high-resolution core logging data, it can be decomposed to determine the corresponding trend line and residual line. Finally, deviations in the target open-hole logging data and trend line are removed. The target open-hole logging data after deviation removal, the trend line, and the residual line are then used to train a multi-level network model, resulting in the final high-resolution logging data model.

[0112] In practice, open-hole logging data and high-resolution core logging data can be acquired. These data are then input into a trained high-resolution logging data model, outputting high-resolution logging data for the entire well section. This method allows for accurate, efficient, and low-cost determination of high-resolution logging data for the entire well section, thus better characterizing the bedding properties of unconventional shale reservoirs. This lays the foundation for establishing a high-precision reservoir mechanics model and optimizing fracturing strategies.

[0113] In a specific scenario example, the method for determining high-resolution logging data for the entire well section provided in the embodiments of this specification can be applied to predict high-resolution logging data for the entire well section, maximizing the utilization rate and value of the core, compensating for the difficulty in obtaining high-resolution logging data in non-core sections or non-core wells, and providing a data foundation for establishing a more reliable reservoir geomechanical model. In specific implementation, the following steps may be included.

[0114] (1) Data analysis and cleaning

[0115] Data analysis and cleaning can be understood as: analyzing, identifying, and supplementing conventional open-hole logging data obtained from open-hole logging from a professional perspective, addressing numerical errors, outliers, missing data, depth mismatches, etc.

[0116] (2) Reconstruction of missing curves in conventional open-hole logging data

[0117] The reconstruction of missing curves in conventional open-hole logging data can be understood as: determining whether there are missing curves in the conventional open-hole logging data after data analysis and cleaning, such as missing curves for neutrons and density. In this case, neighboring wells with complete data can be selected, and machine learning techniques such as cross plots, multiple linear regression, and neural networks can be used to establish a missing curve reconstruction model to obtain the missing curves.

[0118] (3) Analysis of high-resolution core logging data

[0119] High-resolution core logging data analysis can be understood as: performing depth matching between high-resolution core logging data and conventional open-hole logging data.

[0120] (4) Decomposition of high-resolution core logging data

[0121] High-resolution core logging data decomposition can be understood as follows: using moving average technology, the high-resolution core logging data (high-resolution logging data of the core segment, such as the high-resolution logging curve of the core segment) is decomposed into two profiles, corresponding to two roughness levels, namely trend and residual.

[0122] This method allows for the summation and averaging of data points within a specific window, with continuous calculations along the entire profile to obtain a "moving" average result, such as... Figure 3 As shown. Before performing a "moving" average, the size of the sliding window needs to be determined to capture sufficient information. If the window size is too large, the results will include more irrelevant information (loss of resolution). If the window size is too small, the results will fail to reflect the main trend. The appropriate window size can be determined by gradually increasing the window size and comparing the relative error between the smoothed curve and the original curve. When the error curve flattens, it indicates that increasing the window size no longer provides any benefit and results in the loss of more information. See also Figure 4 As shown, Figure 4 The horizontal axis represents depth, and the vertical axis represents P-wave transit time. The high-resolution logging curves for the core section are constructed using data points at intervals of 0.01m. Figure 4 The system sets window sizes of 0.02m, 0.03m, and 0.04m, which can be used to extract trend lines at different window sizes from high-resolution logging curves of core sections using moving average techniques. Figure 4 As can be seen, when the window size is set too large, the trend line becomes smoother, and the deviation from the high-resolution logging curve of the core segment becomes greater. If the window size is set too small, the trend line will be closer to the high-resolution logging curve of the core segment, but it will not be smooth enough. Therefore, by gradually increasing the window size and comparing the relative error between the smoothed trend line and the high-resolution logging curve of the core segment under different windows, a suitable window size and a suitable smoothed trend line can be determined when the relative error is less than a preset error threshold. That is, the window size of the moving average technique can be determined at the turning point where the relative error between the smoothed trend line and the high-resolution logging curve of the core segment under different windows begins to stabilize. After determining the trend line using the moving average technique, the difference between the high-resolution core logging curve of the core segment and the trend line can be calculated as the residual. (See also...) Figure 5 As shown, from Figure 5 The trend lines and residual lines can be seen from the high-resolution logging curves of the core section.

[0123] (5) Remove deviations

[0124] Conventional open-hole logging data is a smoothed estimate of high-resolution core logging data. Therefore, ideally, the trend of high-resolution core logging data should be very close to that of conventional open-hole logging data. However, in real-world data, these two types of data deviate locally. When the training dataset is small, this deviation affects the accuracy of predictions. Therefore, a term can be used... Describe the similarity between the following two sets of data:

[0125]

[0126] in, It can be the difference ratio between the trends of conventional open-hole logging data (conventional open-hole logging curves) and high-resolution core logging data (or high-resolution logging curves of core segments). The trend is obtained by decomposing high-resolution core logging data (or high-resolution logging curves of core segments). It is conventional open-hole logging data (conventional open-hole logging curve). The distribution can be determined based on the standard deviation. Calculations can be performed to obtain a standard deviation. Data points outside this range are considered deviations. See also Figure 6 As shown, Figure 6 The left side shows that the trend lines derived from the high-resolution logging curves of the core section and the conventional open-hole logging curves may have significant local deviations. Figure 6 The right side shows the trend line derived from the high-resolution logging curves after removing the core section, and the curve after significant deviation from the conventional open-hole logging curve. (See also...) Figure 7 As shown, Figure 7 Zone A, Zone B, and Zone C are regions where the trend lines decomposed from the high-resolution logging of the core section deviate from the conventional open-hole logging curve by more than one standard deviation. These are called deviation areas or deviation data. Deviation areas or deviation data can include opposite trends (e.g., Zone A and Zone B) and large differences (e.g., Zone C).

[0127] (6) Model training

[0128] A training dataset is constructed using the removed trend line, conventional open-hole logging curves, and residual lines. The model is then trained based on this dataset. The model uses the Kernel Ridge Regression (KRR) algorithm to learn and predict the "trend" component. The Out-of-Tree (ET) algorithm is used to learn and predict the "residual" component. The prediction of high-resolution logging data across the entire well section is based on the recombination of the global segment's trend prediction and the overall residual prediction; that is, high-resolution data prediction for the entire well section = trend prediction for the entire well section + residual prediction for the entire well section. See also... Figure 8 As shown, Figure 8 The training results of the model are shown, that is, the final output of the model is the sum of the predicted trend line and the residual line of the whole well section, which yields the high-resolution logging curve or high-resolution logging data of the whole well section.

[0129] (7) Model validation

[0130] The results can be compared and verified using measured high-resolution core logging data and high-resolution core logging data predicted by the model for the entire well section. (See reference...) Figure 9 As shown, if the predicted high-resolution logging curve meets the preset accuracy requirements after comparison and verification with the measured high-resolution logging curve of the core section, the model construction ends and a high-resolution logging data model can be obtained (this high-resolution logging data model is used to output high-resolution logging data or high-resolution logging curve of the entire well section). If the error is too large, return to (6) to continue training the model until the error is less than the preset error threshold or the accuracy is higher than the preset accuracy threshold, then stop training the model.

[0131] (8) Output of results

[0132] The results of the predicted high-resolution logging curves can be output numerically.

[0133] The above solution can achieve the following technical effects:

[0134] (1) Based on conventional open-hole logging data and high-resolution core logging data, high-resolution logging data for the entire well section can be predicted, which can maximize the utilization rate and value of the core, make up for the deficiency that it is difficult to obtain high-resolution logging data in non-core sections or non-core wells, and provide a data basis for establishing a more reliable reservoir geomechanical model.

[0135] (2) By using machine learning methods to repeatedly calculate a large amount of historical data, that is, to calculate the model relationship between conventional open-hole logging data and high-resolution core logging data, the amount of calculation can be reduced and the calculation accuracy can be improved.

[0136] (3) In unconventional shale reservoirs, lithological interfaces, bedding, and other mechanically weak surfaces are extremely well developed. These beddings have a significant impact on fracturing operations. However, these beddings are very thin (centimeter level), and conventional open-hole logging data has a resolution of only tens of centimeters, making it impossible to identify these mechanically weak surfaces. Only high-resolution logging data (such as P-wave and S-wave transit time and volume density curves) can better characterize the properties of these beddings. By acquiring high-resolution logging data for the entire well section, we can lay the foundation for establishing a high-precision reservoir mechanical model and more rationally carrying out fracturing scheme optimization.

[0137] Although this specification provides the following examples or appendices Figure 10The methods, steps, or apparatus structures shown may include more or fewer combined operational steps or module units based on conventional or non-inventive methods. In steps or structures where there is no logically necessary causal relationship, the execution order of these steps or the module structure of the apparatus is not limited to the execution order or module structure shown in the embodiments or drawings of this specification. When the methods or module structures described are applied in actual devices, servers, or terminal products, they can be executed sequentially or in parallel according to the methods or module structures shown in the embodiments or drawings (e.g., in parallel processor or multi-threaded processing environments, or even distributed processing or server cluster implementation environments).

[0138] Based on the above-described method for determining high-resolution logging data across the entire well section, this specification also provides an embodiment of a device for determining high-resolution logging data across the entire well section. For example... Figure 10 As shown, the device may specifically include the following modules:

[0139] The acquisition module 1001 can be used to acquire open-hole logging data and high-resolution core logging data;

[0140] The prediction module 1002 can be used to input the open-hole logging data and high-resolution core logging data into the high-resolution logging data model to obtain high-resolution logging data for the entire well section.

[0141] The model training module 1003 can be used to acquire target open-hole logging data samples and target high-resolution core logging data samples; decompose the target high-resolution core logging data samples to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; remove deviation data from the target open-hole logging data samples and target smoothed data; and process a multi-level network model based on the target open-hole logging data samples, target smoothed data and residual data after removing deviation data to obtain a high-resolution logging data model.

[0142] In some embodiments, the model training module 1003 may be further used to obtain initial open-hole logging data samples; preprocess the initial open-hole logging data samples, and determine whether the preprocessed initial open-hole logging data samples are complete; if not, obtain logging data samples of the target adjacent well, and perform missing reconstruction on the preprocessed initial open-hole logging data samples based on the logging data samples of the target adjacent well to obtain the target open-hole logging data samples.

[0143] In some embodiments, the target open-hole logging data sample in the model training module 1003 may include at least one of the following: natural gamma, density, spontaneous potential, well diameter, resistivity, P-wave transit time, S-wave transit time, and compensated neutrons.

[0144] In some embodiments, the model training module 1003 described above can be specifically used to obtain an initial sliding window and continuously adjust the initial sliding window at preset intervals; based on the initial sliding window before adjustment and the initial sliding window after adjustment, sequentially determine multiple sets of high-resolution core logging data samples with different smoothness levels; compare the target high-resolution core logging data sample with the high-resolution core logging data samples with different smoothness levels to determine the target smoothed data; and determine the residual data based on the target high-resolution core logging data sample and the target smoothed data.

[0145] In some embodiments, the model training module 1003 may also be used to sequentially calculate the relative error between the target high-resolution core logging data sample and the high-resolution core logging data samples with different smoothness; compare the relative error, and when the relative error is less than a preset error threshold, determine the target smooth data from the high-resolution core logging data samples with different smoothness.

[0146] In some embodiments, the model training module 1003 may further be used to calculate the difference ratio data between the target open-hole logging data sample and the target smoothed data; determine the standard deviation corresponding to the difference ratio data based on the difference ratio data; and remove deviation data from the target open-hole logging data sample and the target smoothed data based on the standard deviation.

[0147] In some embodiments, the multi-level network model may include a trend prediction module and a residual prediction module model. Specifically, the training module 1003 may also be used to input the target open-hole logging data sample after removing deviation data and the target smoothed data into the trend prediction module to obtain a trend prediction data sample for the entire well section; input the target open-hole logging data sample after removing deviation data and the residual data into the residual prediction module to obtain a residual prediction data sample for the entire well section; obtain a high-resolution logging data sample for the entire well section based on the trend prediction data sample and the residual prediction data sample for the entire well section; compare and verify the high-resolution logging data sample for the entire well section with the target high-resolution core logging data sample, and obtain a high-resolution logging data model when the comparison and verification result meets the preset accuracy requirements.

[0148] It should be noted that the units, devices, or modules described in the above embodiments can be implemented by computer chips or physical entities, or by products with certain functions. For ease of description, the above devices are described by dividing them into various modules according to their functions. Of course, in implementing this specification, the functions of each module can be implemented in one or more software and / or hardware, or the module that implements the same function can be implemented by a combination of multiple sub-modules or sub-units, etc. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection between the devices or units shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0149] As can be seen from the above, the device for determining high-resolution logging data across the entire well section, based on the embodiments provided in this specification, can acquire high-resolution logging data across the entire well section. This solves the problems of low resolution in conventional open-hole logging data, which prevents reservoir geomechanical models constructed using such data from effectively reflecting the mechanically weak features of lithological interfaces, bedding, and foliation in shale reservoirs, as well as the strong vertical heterogeneity of shale reservoirs, and further fails to accurately reflect the impact of hydraulic fracture propagation. Acquiring high-resolution logging data across the entire well section provides a data foundation for establishing more accurate reservoir geomechanical models and lays the foundation for developing reasonable shale oil and gas development plans and conducting more reasonable and reliable hydraulic fracture propagation simulation analyses.

[0150] This specification also provides an electronic device for determining a method based on high-resolution logging data across the entire well section, including a processor and a memory for storing processor-executable instructions. Specifically, the processor can execute the following steps according to the instructions: acquiring open-hole logging data and high-resolution core logging data; inputting the open-hole logging data and high-resolution core logging data into a high-resolution logging data model to obtain high-resolution logging data across the entire well section; wherein the high-resolution logging data model is trained as follows: acquiring target open-hole logging data samples and target high-resolution core logging data samples; decomposing the target high-resolution core logging data samples to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; removing deviation data from the target open-hole logging data samples and target smoothed data; and processing a multi-level network model based on the target open-hole logging data samples, target smoothed data, and residual data after removing deviation data to obtain the high-resolution logging data model.

[0151] To execute the above instructions more accurately, please refer to... Figure 11As shown in the embodiments of this specification, another specific electronic device is also provided, wherein the electronic device includes a network communication port 1101, a processor 1102 and a memory 1103, and the above structures are connected by internal cables so that the various structures can perform specific data interaction.

[0152] Specifically, the network communication port 1101 can be used to acquire open-hole logging data and high-resolution core logging data.

[0153] The processor 1102 is specifically used to input the open-hole logging data and high-resolution core logging data into a high-resolution logging data model to obtain high-resolution logging data for the entire well section. The high-resolution logging data model is trained as follows: acquiring target open-hole logging data samples and target high-resolution core logging data samples; decomposing the target high-resolution core logging data samples to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; removing deviation data from the target open-hole logging data samples and target smoothed data; and processing a multi-level network model based on the target open-hole logging data samples, target smoothed data, and residual data after removing deviation data to obtain the high-resolution logging data model.

[0154] The memory 1103 can be used to store the corresponding instruction program.

[0155] In this embodiment, the network communication port 1101 can be a virtual port bound to different communication protocols, thereby enabling the sending or receiving of different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0156] In this embodiment, the processor 1102 can be implemented in any suitable manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc. This specification is not limiting.

[0157] In this embodiment, the memory 1103 may include multiple layers. In a digital system, anything that can store binary data can be a memory. In an integrated circuit, a circuit with storage function but no physical form is also called a memory, such as RAM, FIFO, etc. In a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0158] This specification also provides a computer storage medium for a method of determining high-resolution logging data across the entire well section. The computer storage medium stores computer program instructions that, when executed, perform the following: acquiring open-hole logging data and high-resolution core logging data; inputting the open-hole logging data and high-resolution core logging data into a high-resolution logging data model to obtain high-resolution logging data across the entire well section; wherein the high-resolution logging data model is trained as follows: acquiring target open-hole logging data samples and target high-resolution core logging data samples; decomposing the target high-resolution core logging data samples to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; removing deviation data from the target open-hole logging data samples and target smoothed data; and processing a multi-level network model based on the target open-hole logging data samples, target smoothed data, and residual data after removing deviation data to obtain the high-resolution logging data model.

[0159] In this embodiment, the storage medium includes, but is not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), Cache, Hard Disk Drive (HDD), or Memory Card. The memory can be used to store computer program instructions. The network communication unit can be an interface configured according to standards specified in the communication protocol for network connection communication.

[0160] While this specification provides the steps of operation for the methods described in the embodiments or flowcharts, more or fewer steps may be included based on conventional or non-inventive means. The order of steps listed in the embodiments is merely one possible order of execution among many steps and does not represent the only possible order. In actual device or client product execution, the methods shown in the embodiments or drawings may be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed data processing environment). The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or apparatus. Without further limitations, the presence of other identical or equivalent elements in a process, method, product, or apparatus that includes said elements is not excluded. The terms "first," "second," etc., are used to denote names and do not indicate any particular order.

[0161] Those skilled in the art will also know that, besides implementing the controller using purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller function as logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices within it used to implement various functions can also be considered structures within that hardware component. Alternatively, the devices used to implement various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0162] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0163] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this specification can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of this specification can essentially be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments of this specification.

[0164] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. This specification can be used in numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable electronic devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0165] Although this specification has been described by way of examples, those skilled in the art will recognize that many variations of this specification are possible without departing from its spirit, and it is intended that the appended claims cover such variations without departing from the spirit of this specification.

Claims

1. A method for determining high-resolution logging data across the entire well section, characterized in that, include: Acquire open-hole logging data and high-resolution core logging data; The open-hole logging data and high-resolution core logging data are input into the high-resolution logging data model to obtain high-resolution logging data for the entire well section; The high-resolution logging data model is trained as follows: target open-hole logging data samples and target high-resolution core logging data samples are acquired; the target high-resolution core logging data samples are decomposed to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; deviation data is removed from the target open-hole logging data samples and target smoothed data; a multi-level network model is processed based on the target open-hole logging data samples, target smoothed data and residual data after removing deviation data to obtain the high-resolution logging data model. The removal of deviation data from the target open-hole logging data samples and the target smoothed data includes: Calculate the difference ratio between the target open-hole logging data sample and the target smoothed data; Based on the difference ratio data, determine the standard deviation corresponding to the difference ratio data; Based on the standard deviation, remove the off-target data from the target open-hole logging data sample and the target smoothed data; The multi-level network model includes a trend prediction module and a residual prediction module. Correspondingly, the multi-level network model, processed based on the target open-hole logging data sample after removing deviation data, target smoothed data, and residual data, yields a high-resolution logging data model, including: The target open-hole logging data sample after removing the deviation data and the target smoothed data are input into the trend prediction module to obtain the trend prediction data sample for the entire well section; The target open-hole logging data sample and residual data after removing the deviation data are input into the residual prediction module to obtain the residual prediction data sample for the entire well section. Based on the trend prediction data sample and the residual prediction data sample of the entire well section, a high-resolution logging data sample of the entire well section is obtained. The high-resolution logging data sample of the entire well section is compared and verified with the target high-resolution core logging data sample. When the comparison and verification results meet the preset accuracy requirements, a high-resolution logging data model is obtained.

2. The method according to claim 1, characterized in that, The method further includes: Obtain initial open-hole logging data samples; The initial open-hole logging data sample is preprocessed to determine whether the preprocessed initial open-hole logging data sample is complete. If not, then obtain the logging data sample of the target adjacent well, and perform missing reconstruction on the preprocessed initial open-hole logging data sample based on the logging data sample of the target adjacent well to obtain the target open-hole logging data sample.

3. The method according to claim 2, characterized in that, The target open-hole logging data sample includes at least one of the following: natural gamma, density, spontaneous potential, borehole diameter, resistivity, P-wave transit time, S-wave transit time, and compensated neutrons.

4. The method according to claim 1, characterized in that, The process of decomposing the target high-resolution core logging data sample to obtain the target smoothed data and residual data corresponding to the target high-resolution core logging data sample includes: Obtain an initial sliding window and continuously adjust the initial sliding window at preset intervals; Based on the initial sliding window before adjustment and the initial sliding window after adjustment, multiple sets of high-resolution core logging data samples with different smoothness are determined in sequence. The target smoothed data is determined by comparing the target high-resolution core logging data sample with the high-resolution core logging data sample with different smoothing degrees. The residual data is determined based on the target high-resolution core logging data sample and the target smoothed data.

5. The method according to claim 4, characterized in that, The comparison of the target high-resolution core logging data sample with high-resolution core logging data samples with different smoothing degrees to determine the target smoothed data includes: The relative errors between the target high-resolution core logging data sample and the high-resolution core logging data samples with different smoothness levels are calculated sequentially. By comparing the relative error, when the relative error is less than a preset error threshold, the target smoothed data is determined from high-resolution core logging data samples with different smoothness levels.

6. A device for determining high-resolution logging data across the entire well section, characterized in that, include: The acquisition module is used to acquire open-hole logging data and high-resolution core logging data; The prediction module is used to input the open-hole logging data and high-resolution core logging data into the high-resolution logging data model to obtain high-resolution logging data for the entire well section. The model training module is used to acquire target open-hole logging data samples and target high-resolution core logging data samples; the target high-resolution core logging data samples are decomposed to obtain target smoothed data and residual data corresponding to the target high-resolution core logging data samples; Remove off-target data from the target open-hole logging data samples and the target smoothed data; A high-resolution logging data model is obtained by processing a multi-level network model based on the target open-hole logging data sample after removing deviation data, target smoothed data, and residual data. The removal of deviation data from the target open-hole logging data samples and the target smoothed data includes: Calculate the difference ratio between the target open-hole logging data sample and the target smoothed data; Based on the difference ratio data, determine the standard deviation corresponding to the difference ratio data; Based on the standard deviation, remove the off-target data from the target open-hole logging data sample and the target smoothed data; The multi-level network model includes a trend prediction module and a residual prediction module. Correspondingly, the multi-level network model, processed based on the target open-hole logging data sample after removing deviation data, target smoothed data, and residual data, yields a high-resolution logging data model, including: The target open-hole logging data sample after removing the deviation data and the target smoothed data are input into the trend prediction module to obtain the trend prediction data sample for the entire well section; The target open-hole logging data sample and residual data after removing the deviation data are input into the residual prediction module to obtain the residual prediction data sample for the entire well section. Based on the trend prediction data sample and the residual prediction data sample of the entire well section, a high-resolution logging data sample of the entire well section is obtained. The high-resolution logging data sample of the entire well section is compared and verified with the target high-resolution core logging data sample. When the comparison and verification results meet the preset accuracy requirements, a high-resolution logging data model is obtained.

7. An electronic device, characterized in that, The method includes a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to implement the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, It stores computer instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 5.

Images