Logging data missing filling method and device

By performing wavenumber-domain low-pass filtering and model training on the log data, the problem of low filling accuracy of log data is solved, and the reliability of log data and the accuracy of reservoir prediction is improved.

CN119066411APending Publication Date: 2024-12-03CHINA NAT PETROLEUM CORP +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310630714.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-30
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing logging data filling method has low accuracy and poor reliability, making it difficult to accurately predict the reservoir location, affecting the reservoir prediction results.

Method used

By performing wavenumber-domain low-pass filtering on the original well logging curve, low-frequency well logging curves are generated, missing data are predicted using model training, and filled with existing data, and finally a complete well logging curve is generated.

Benefits of technology

Improve the reliability and accuracy of logging data loss filling, ensuring the data integrity and accuracy of reservoir prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119066411A_ABST
    Figure CN119066411A_ABST
Patent Text Reader

Abstract

The invention discloses a method and a device for filling missing logging data. The method comprises the steps that wavenumber domain low-pass filtering is carried out on a plurality of original well logging curves in a to-be-researched area, a plurality of low-frequency well logging curves are obtained, and missing data exist in each original well logging curve and each low-frequency well logging curve; performing model training on the plurality of low-frequency logging curves, and predicting to obtain a plurality of low-frequency logging curves without missing data; the part, with missing data, of each original logging curve is filled with data, located at the same depth position, of the corresponding low-frequency logging curve without missing data, and a plurality of transition logging curves are obtained; and performing model training on the plurality of transition logging curves, and predicting to obtain a plurality of complete logging curves without missing data. According to the method, the reliability and precision of filling the missing logging data can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of petroleum exploration technology, and in particular to a method and device for filling missing well logging data. Background Art

[0002] In oil and gas exploration technology, the logging curve constructed from logging data is used to predict the location of the reservoir, and its quality will directly affect the prediction results of the reservoir. In actual production, due to factors such as expansion, instrument failure, and improper operation, it is easy to cause missing logging data and incomplete logging curves, which brings huge challenges to subsequent reservoir prediction work such as modeling and inversion.

[0003] In the related art, the method for filling missing well logging data is mainly a single filling method, that is, filling the missing part by default value, mean value or mode. Although this method has a fast processing speed, it has low filling accuracy, poor reliability, and is prone to serious data deviation. Summary of the invention

[0004] In view of this, the present application provides a method and device for filling missing well logging data, which can improve the reliability and accuracy of filling missing well logging data.

[0005] Specifically, the following technical solutions are included:

[0006] In a first aspect, an embodiment of the present application provides a method for filling missing well logging data, the method comprising:

[0007] The multiple original well logging curves in the study area are subjected to wave number domain low-pass filtering to obtain multiple low-frequency well logging curves, in which each original well logging curve and each low-frequency well logging curve has missing data;

[0008] Performing model training on the plurality of low-frequency logging curves to predict and obtain a plurality of low-frequency logging curves without missing data;

[0009] For each part of the original well logging curve with missing data, the part is filled with data at the same depth position on the corresponding low-frequency well logging curve without missing data, so as to obtain multiple transition well logging curves;

[0010] Model training is performed on the multiple transition logging curves to predict and obtain multiple complete logging curves without missing data.

[0011] In some embodiments, before performing wavenumber domain low-pass filtering on the multiple original well logging curves in the study area to obtain multiple low-frequency well logging curves, the method further includes:

[0012] Acquire well logging data in the area to be studied, wherein the well logging data includes a plurality of original well logging curve data;

[0013] The multiple original logging curve data are preprocessed to obtain multiple original logging curves.

[0014] In some embodiments, the preprocessing of the plurality of original well logging curve data to obtain the plurality of original well logging curves comprises:

[0015] Detecting each original logging curve data, marking invalid values ​​in each original logging curve, wherein the data corresponding to the invalid values ​​represent missing data;

[0016] For each original logging curve data, the data point at the current depth position is marked as an invalid value, and the data point at the current depth position of each original logging curve is deleted.

[0017] In some embodiments, the detecting each original well logging curve data and marking the invalid value in each original well logging curve includes:

[0018] For each piece of original logging curve data, in response to the data value being equal to a preset value, marking the data value as an invalid value; or,

[0019] For each piece of original logging curve data, in response to a data value being greater than a first threshold or less than a second threshold, the data value is marked as an invalid value.

[0020] In some embodiments, the preprocessing of the plurality of original well logging curve data to obtain the plurality of original well logging curves further comprises:

[0021] Based on the invalid values ​​in each original logging curve, the missing ratio of each original logging curve is calculated;

[0022] Based on the missing ratio of each original well logging curve, a plurality of original well logging curves after missing ratio sorting are obtained, wherein the missing ratio sorting includes sorting from high to low missing ratio and sorting from low to high missing ratio.

[0023] In some embodiments, the plurality of low-frequency logging curves include a target low-frequency logging curve, the target low-frequency logging curve includes a first layer segment and a second layer segment, wherein the first layer segment is a layer segment where a valid value is located, and the second layer segment is a layer segment where an invalid value is located, and the model training is performed on the plurality of low-frequency logging curves to predict a plurality of low-frequency logging curves without missing data, comprising:

[0024] For the target low-frequency logging curve, the following first processing process is performed:

[0025] The data of the first layer segment of the target low-frequency logging curve and the data of the corresponding layer segments of other low-frequency logging curves other than the target low-frequency logging curve at the same depth as the first layer segment are respectively used as the output and input of the model training to obtain a first training model, wherein the missing data of the corresponding layer segments of other low-frequency logging curves at the same depth as the first layer segment are obtained by using polynomial interpolation;

[0026] Substituting data of corresponding layers of other low-frequency logging curves, except the target low-frequency logging curve, at the same depth as the second layer segment into the first training model to obtain predicted data of the second layer segment of the target low-frequency logging curve;

[0027] Filling the predicted data of the second layer segment of the target low-frequency logging curve into the second layer segment of the target low-frequency logging curve to obtain a low-frequency logging curve without missing data;

[0028] Each low-frequency logging curve is taken as a target low-frequency logging curve one by one, and the first processing process as described above is performed to obtain the plurality of low-frequency logging curves without missing data.

[0029] In some embodiments, the step of taking each low-frequency logging curve as a target low-frequency logging curve and executing the first processing as described above comprises:

[0030] According to the missing order, each low-frequency logging curve is used as a target low-frequency logging curve, and the first processing process as described above is performed, wherein the missing order is obtained according to the missing ratio of the original logging curve, and the missing ratio of the original logging curve is obtained according to the invalid value in the original logging curve;

[0031] The missingness ranking is a ranking from low to high missingness ratio.

[0032] In some embodiments, the plurality of transition logging curves include a target transition logging curve, the target transition logging curve includes a third layer segment and a fourth layer segment, wherein the third layer segment is a layer segment where a valid value is located, and the fourth layer segment is a layer segment where an invalid value is located, and the model training is performed on the plurality of transition logging curves to predict a plurality of complete logging curves without missing data, comprising:

[0033] For the target transition logging curve, the following second processing process is performed:

[0034] The model training is performed using the data of the third layer segment of the target transition logging curve and the data of the corresponding layer segments of other transition logging curves other than the target transition logging curve at the same depth as the third layer segment as the output and input of the model training, respectively, to obtain a second training model;

[0035] Substituting data of corresponding layers of other transition logging curves except the target transition logging curve at the same depth as the fourth layer segment into the second training model to obtain predicted data of the fourth layer segment of the target transition logging curve;

[0036] Filling the predicted data of the fourth layer segment of the target transition logging curve into the fourth layer segment of the target transition logging curve to obtain a complete logging curve without missing data;

[0037] Each transition logging curve is taken as a target transition logging curve one by one, and the second processing process as described above is performed to obtain the plurality of complete logging curves without missing data.

[0038] In some embodiments, the step of taking each transition logging curve as a target transition logging curve one by one and performing the second processing as described above comprises:

[0039] According to the missing order, each transition logging curve is used as a target transition logging curve, and the second processing process as described above is performed, wherein the missing order is obtained according to the missing ratio of the original logging curve, and the missing ratio of the original logging curve is obtained according to the invalid value in the original logging curve;

[0040] The missingness ranking is a ranking from low to high missingness ratio.

[0041] In a second aspect, an embodiment of the present application provides a device for filling missing well logging data, the device comprising:

[0042] A low-pass filtering module is used to perform low-pass filtering in the wave number domain on multiple original logging curves in the study area to obtain multiple low-frequency logging curves, wherein each original logging curve and each low-frequency logging curve has missing data;

[0043] A first prediction module is used to perform model training on the plurality of low-frequency logging curves to predict a plurality of low-frequency logging curves without missing data;

[0044] A filling module is used to fill in the missing data of each original logging curve with the data at the same depth position on the corresponding low-frequency logging curve without missing data, so as to obtain multiple transition logging curves;

[0045] The second prediction module is used to perform model training on the plurality of transition logging curves to predict a plurality of complete logging curves without missing data.

[0046] The method for filling missing well logging data provided in the embodiment of the present application performs low-pass filtering in the wave number domain on multiple original well logging curves in the study area to obtain multiple low-frequency well logging curves that can display the low-frequency trend form, and then performs model training on multiple low-frequency well logging curves to predict multiple complete low-frequency well logging curves without missing data, and fill the part of the original well logging curve with missing data using the data on the corresponding low-frequency well logging curve without missing data to obtain multiple transition well logging curves. Based on the model training of multiple transition well logging curves, multiple well logging curves without missing data can be predicted, and these multiple well logging curves without missing data are well logging curves that have completed data missing filling. This method makes full use of the existing effective well logging data, and combines with model training on the basis of considering the overall trend of missing data to predict well logging curves with complete data, so that the reliability and accuracy of well logging data missing filling are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 A flowchart of a method for filling missing well logging data provided in an embodiment of the present application;

[0049] Figure 2 A real well logging curve diagram without missing well logging data provided in an embodiment of the present application;

[0050] Figure 3 A well logging curve diagram with missing well logging data provided in an embodiment of the present application;

[0051] Figure 4 A logging curve diagram after filling a missing logging data filling method provided in an embodiment of the present application;

[0052] Figure 5 A structural block diagram of a device for filling missing well logging data provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0054] Well logging measures geophysical parameters such as conductivity, acoustic properties, radioactivity, and electrochemical properties of rock formations through geophysical methods. By interpreting and applying logging curves that represent these parameters, it is possible to identify underground reservoirs, which is of great significance for oil and gas exploration and development. The quality of logging curves directly affects the results of reservoir prediction.

[0055] However, in actual production, due to factors such as expansion, instrument failure, improper operation, and cost control, some layers and some wells of logging curves are often missing, resulting in reduced reliability and fewer types of logging curves, which poses a huge challenge to subsequent reservoir prediction work such as modeling and inversion. Therefore, it is necessary to fill in the missing parts of logging data to ensure the integrity and richness of the data.

[0056] At present, the following two problems are often encountered in the process of filling missing well logging data: (1) Different well logging curves are essentially the embodiment of different physical characteristics of the same reservoir. There is a certain mapping relationship between the well logging data. However, due to the complex formation conditions and anisotropy, traditional physical models or empirical models are generally linear or simple regression, which are difficult to accurately describe the relationship between well logging curves, and the filling accuracy of missing values ​​is not high; (2) Due to the different missing layers of various curves in the same well, or the different types of missing curves in different wells, it is difficult to use the same model / method to fill missing data for multiple wells in the work area. The variability of missing data types and missing areas increases the difficulty of applying traditional models. The commonly used single filling method has a fast processing speed, but has low filling accuracy, poor reliability, and is prone to serious data deviation.

[0057] In order to solve the problems existing in the prior art, an embodiment of the present application provides a method for filling missing well logging data. Figure 1 A flowchart of a method for filling missing well logging data provided in an embodiment of the present application, see Figure 1 , the method comprises the following steps:

[0058] Step 101 , low-pass filtering is performed in the wavenumber domain on multiple original logging curves in the study area to obtain multiple low-frequency logging curves, wherein each original logging curve and each low-frequency logging curve has missing data.

[0059] By performing low-pass filtering in the wavenumber domain on multiple original logging curves in the study area, multiple low-frequency logging curves can be obtained, and the low-frequency logging curves can reflect the overall trend.

[0060] Among them, the original logging curve corresponds to the low-frequency logging curve one by one, and the trends and change trends of the two are consistent. Since each original logging curve has missing data, each low-frequency logging curve also has missing data. The fact that each original logging curve has missing data means that the logging data of multiple original logging curves at the same depth are not complete, that is, at the same depth, at least one original logging curve has missing data.

[0061] For example, the number of original well logging curves may be seven.

[0062] Prior to this step, the method for filling missing well logging data provided in the embodiment of the present application also includes: obtaining well logging data in the area to be studied, wherein the well logging data includes multiple original well logging curve data; preprocessing the multiple original well logging curve data to obtain multiple original well logging curves.

[0063] It can be understood that each data point in the logging curve data corresponds to a depth value and values ​​of multiple physical properties.

[0064] Furthermore, the process of performing low-pass filtering in the wavenumber domain on multiple original well logging curves in the study area can be as follows:

[0065] (1) Perform wavenumber domain transformation on the logging data in the depth domain in the area to be studied, so that the depth domain logging data is transformed into the wavenumber domain:

[0066] U(k)=∫U(z)e ikz dz ①

[0067]

[0068] In the formula: U(k) represents the wave number domain logging data, z represents the depth of the well, i marks the complex domain, k represents the wave number, v represents the acoustic wave velocity, and w represents the main frequency of the logging data. It can be understood that the logging data obtained in the area to be studied are all depth domain data.

[0069] (2) Determine the main frequency w2 of the seismic data at the location of the well to be studied in the study area, and use the above formula ② to calculate the wave number k2 corresponding to the main frequency of the seismic data.

[0070] (3) The wave number domain data is low-pass filtered using a Chebyshev low-pass filter, with the cutoff frequency set to the wave number k2 corresponding to the main frequency of the earthquake, to obtain the filtered low-frequency wave number domain logging data U(k).

[0071] (4) The filtered low-frequency wavenumber domain logging data is transformed back to the depth domain to obtain the depth domain low-frequency logging data:

[0072] U(z)=∫U(k)e -ikz dk

[0073] Where: U(z) represents the depth domain logging data.

[0074] In some embodiments, multiple original logging curve data are preprocessed to obtain multiple original logging curves, including: detecting each original logging curve data, marking invalid values ​​in each original logging curve, wherein the data corresponding to the invalid values ​​represent missing data; for each original logging curve data, corresponding data points in response to the current depth position are marked as invalid values, and the data points at the current depth position of each original logging curve are deleted.

[0075] Here, invalid values ​​can also be called null values, indicating that the data here is missing.

[0076] When the data points of each original logging curve at the current depth position are marked as invalid values, it means that there is no valid value at the current depth position. Therefore, the data points of each original logging curve at the current depth position can be deleted to facilitate subsequent model training.

[0077] In some embodiments, invalid values ​​include system abnormal values ​​and response range invalid values. Therefore, each original well logging curve data is detected and the invalid values ​​in each original well logging curve are marked, including:

[0078] In the case where the invalid value is a system abnormal value, for each original logging curve data, in response to the data value being equal to a preset value, the data value is marked as an invalid value; or, in the case where the invalid value is an invalid value in the response range, for each original logging curve data, in response to the data value being greater than a first threshold or less than a second threshold, the data value is marked as an invalid value.

[0079] That is, when the data value is a preset value representing a system abnormal value, the data value is marked as an invalid value; when the data value does not fall within the threshold interval formed by the first threshold and the second threshold, the data value is marked as an invalid value.

[0080] In some embodiments, preprocessing the multiple original logging curve data to obtain the multiple original logging curves also includes: calculating the missing ratio of each original logging curve based on the invalid values ​​in each original logging curve; and obtaining the multiple original logging curves after missing ratio sorting based on the missing ratio of each original logging curve, wherein the missing ratio sorting includes sorting the missing ratio from high to low and sorting the missing ratio from low to high.

[0081] By calculating the missing ratio of the original logging curve, data preparation is done for subsequent model training.

[0082] It can be understood that, for an original well logging curve, its missing ratio is the ratio of the number of invalid values ​​in the original well logging curve to the number of all values ​​(invalid values ​​and valid values) in the original well logging curve.

[0083] Step 102: Perform model training on multiple low-frequency logging curves to predict and obtain multiple low-frequency logging curves without missing data.

[0084] The multiple low-frequency logging curves include a target low-frequency logging curve, and the target low-frequency logging curve includes a first layer segment and a second layer segment, wherein the first layer segment is a layer segment where a valid value is located, and the second layer segment is a layer segment where an invalid value is located.

[0085] It can be understood that the target low-frequency logging curve can be any one of multiple low-frequency logging curves.

[0086] This step may include the following sub-steps:

[0087] Step 1: For the target low-frequency logging curve, perform the following first processing:

[0088] (1) The data of the first layer segment of the target low-frequency logging curve and the data of the corresponding layer segments of other low-frequency logging curves other than the target low-frequency logging curve at the same depth as the first layer segment are used as the output and input of model training respectively to obtain a first training model, wherein the missing data of the corresponding layer segments of other low-frequency logging curves at the same depth as the first layer segment are obtained by using polynomial interpolation.

[0089] The first training model may be a linear regression model.

[0090] For example, when there are seven low-frequency logging curves and the well depth is 1000 meters, when the first segment of the target low-frequency logging curve is the segment of 0-800 meters and the second segment is the segment of 800-1000 meters, the input of the model training can be the data segment corresponding to 0-800 meters of the other six low-frequency logging curves except the target low-frequency logging curve, and the output can be the data segment corresponding to 0-800 of the target low-frequency logging curve.

[0091] (2) Substituting data of corresponding layers of other low-frequency logging curves at the same depth as the second layer segment except the target low-frequency logging curve into the first training model to obtain predicted data of the second layer segment of the target low-frequency logging curve.

[0092] For example, when the first segment of the target low-frequency logging curve is the segment of 0-800 meters and the second segment is the segment of 800-1000 meters, the data segments corresponding to 800-1000 meters of the other six low-frequency logging curves except the target low-frequency logging curve can be used as input, and the output is the data segment corresponding to 800-1000 meters of the target low-frequency logging curve.

[0093] (3) Filling the predicted data of the second layer segment of the target low-frequency logging curve into the second layer segment of the target low-frequency logging curve to obtain a low-frequency logging curve without missing data.

[0094] The predicted data of the second layer segment of the target low-frequency logging curve obtained through model training and learning can reflect the trend of the target low-frequency logging curve. Filling it into the second layer segment of the target low-frequency logging curve can make the target low-frequency logging curve a logging curve with complete data, that is, a low-frequency logging curve without missing data.

[0095] Step 2: Take each low-frequency logging curve as the target low-frequency logging curve one by one, and execute the first processing process in step 1 above to obtain multiple low-frequency logging curves without missing data.

[0096] It can be understood that after the first processing process of step one, a target low-frequency logging curve without missing data can be obtained, and by performing the first processing process in step one on each low-frequency logging curve one by one, multiple low-frequency logging curves without missing data can be obtained.

[0097] In some embodiments, this step includes: sorting by missing values, taking each low-frequency logging curve as a target low-frequency logging curve, and executing the first processing process as above, wherein the missing value sorting is obtained according to the missing value ratio of the original logging curve, and the missing value ratio of the original logging curve is obtained according to the invalid value in the original logging curve; wherein the missing value sorting is sorting by missing value ratio from low to high.

[0098] By performing the above first processing process on multiple low-frequency logging curves according to the missing sorting from low to high, the predicted low-frequency logging curve can be made more consistent with the actual low-frequency trend.

[0099] After this step is completed, the low-frequency regression of the invalid value part of all low-frequency logging curves is completed.

[0100] Step 103, for each original logging curve with missing data, fill in the missing data at the same depth position on the corresponding low-frequency logging curve without missing data to obtain multiple transition logging curves.

[0101] Based on step 102, multiple complete low-frequency logging curves without missing data can be obtained. This step is equivalent to filling the missing data part of the original logging curve with the corresponding part of the obtained low-frequency logging curve without missing data, so as to obtain multiple transition logging curves with complete data.

[0102] It can be understood that the transition logging curve is equivalent to a logging curve obtained by performing preliminary patching on the original logging curve, and is still not the final logging curve.

[0103] Step 104 , model training is performed on the multiple transition logging curves to predict and obtain multiple complete logging curves without missing data.

[0104] The multiple transition logging curves include a target transition logging curve, and the target transition logging curve includes a third layer segment and a fourth layer segment, wherein the third layer segment is a layer segment where a valid value is located, and the fourth layer segment is a layer segment where an invalid value is located.

[0105] It can be understood that the target transition logging curve can be any one of a plurality of past logging curves.

[0106] This step may include the following sub-steps:

[0107] Step 1: For the target transition logging curve, perform the following second processing:

[0108] (1) Model training is performed using the data of the third layer segment of the target transition logging curve and the data of the corresponding layer segments of other transition logging curves other than the target transition logging curve at the same depth as the third layer segment as the output and input of model training, respectively, to obtain a second training model.

[0109] The second training model may be an extreme tree model.

[0110] This training method is the same as the training method in step 1 (1) of step 102, and will not be described in detail here.

[0111] (2) Substituting data of corresponding layers of other transition logging curves other than the target transition logging curve at the same depth as the fourth layer into the second training model, the predicted data of the fourth layer of the target transition logging curve is obtained.

[0112] This substitution method is the same as the substitution method in step 1 (2) of step 102, and will not be repeated here.

[0113] (3) Fill the predicted data of the fourth layer segment of the target transition logging curve into the fourth layer segment of the target transition logging curve to obtain a complete logging curve without missing data.

[0114] This filling method is the same as the filling method in step 1 (3) of step 102, and will not be repeated here.

[0115] Step 2: Take each transition logging curve as the target transition logging curve one by one, and execute the second processing as above to obtain multiple complete logging curves without missing data.

[0116] It can be understood that after the second processing process of step one, a complete logging curve without missing data can be obtained, and by performing the second processing process in step one on each transition logging curve one by one, multiple complete logging curves without missing data can be obtained.

[0117] In some embodiments, this step includes: sorting by missing values, taking each transition logging curve as a target transition logging curve, and executing the second processing process as above, wherein the missing value sorting is obtained according to the missing value ratio of the original logging curve, and the missing value ratio of the original logging curve is obtained according to the invalid value in the original logging curve; wherein the missing value sorting is sorting by missing value ratio from low to high.

[0118] Therefore, the method for filling missing well logging data provided in the embodiment of the present application performs low-pass filtering in the wave number domain on multiple original well logging curves in the study area to obtain multiple low-frequency well logging curves that can display the low-frequency trend form, and then performs model training on multiple low-frequency well logging curves to predict multiple complete low-frequency well logging curves without missing data, and fill the part of the original well logging curve with missing data using the data on the corresponding low-frequency well logging curve without missing data to obtain multiple transition well logging curves. Based on the model training of multiple transition well logging curves, multiple well logging curves without missing data can be predicted, and these multiple well logging curves without missing data are well logging curves that have completed data missing filling. Compared with the prior art, this method makes full use of the existing effective well logging data, and on the basis of considering the overall trend of the missing data, combines with model training to predict well logging curves with complete data, so that the reliability and accuracy of filling missing well logging data are improved.

[0119] At the same time, this method does not depend on the type and correlation of logging curves, and has low requirements on the user's professional background. The prediction results ensure the integrity of the logging data, laying the foundation for subsequent reservoir prediction work such as logging interpretation, modeling, and inversion, and has a broad application market.

[0120] In a possible example, in order to verify the effectiveness of the method for filling the missing well logging data provided in the embodiment of the present application, taking the well logging data of a carbonate rock research area as an example, the data types include gamma, neutron, density, porosity, water saturation and P-wave velocity, and the curves are displayed from top to bottom as follows: Figure 2 ,at this time, Figure 2 is a real logging curve diagram without missing logging data. In each logging curve, several sections of logging data are randomly deleted, such as Figure 3 As shown, Figure 3 A logging curve diagram with missing logging data. In this case, multiple types of logging data are missing at the same time, and it is necessary to fill in the missing logging data. After the logging curve is predicted according to the method for filling in missing logging data provided in the embodiment of the present application, Figure 4 As shown, Figure 4 The well logging curve diagram after filling with the missing well logging data filling method shows that the predicted result is highly consistent with the actual result, indicating that the missing well logging data filling method provided in the embodiment of the present application is highly practical and the predicted result is reasonable.

[0121] Figure 5 This is a structural block diagram of a device for filling missing well logging data provided by an embodiment of the present application. Figure 5 , the device comprises:

[0122] A low-pass filtering module 501 is used to perform wave number domain low-pass filtering on multiple original well logging curves in the study area to obtain multiple low-frequency well logging curves, wherein each original well logging curve and each low-frequency well logging curve has missing data;

[0123] The first prediction module 502 is used to perform model training on multiple low-frequency logging curves to predict multiple low-frequency logging curves without missing data;

[0124] A filling module 503 is used to fill in the missing data portion of each original logging curve with the data at the same depth position on the corresponding low-frequency logging curve without missing data, so as to obtain multiple transition logging curves;

[0125] The second prediction module 504 is used to perform model training on multiple transition logging curves to predict multiple complete logging curves without missing data.

[0126] In some embodiments, the apparatus for filling missing well logging data provided in the embodiments of the present application further includes:

[0127] The acquisition module is used to acquire the logging data in the area to be studied, and the logging data includes multiple original logging curve data;

[0128] The preprocessing module is used to preprocess multiple original logging curve data to obtain multiple original logging curves.

[0129] In some embodiments, the pre-processing module includes:

[0130] A marking unit is used to detect each original logging curve data and mark the invalid value in each original logging curve, wherein the data corresponding to the invalid value represents missing data;

[0131] The deleting unit is used for deleting the data point at the current depth position of each original logging curve data in response to marking the data point at the current depth position as an invalid value.

[0132] In some embodiments, the labeling unit comprises:

[0133] The first marking subunit is used for marking each piece of original logging curve data as an invalid value in response to the data value being equal to a preset value; or,

[0134] The second marking subunit is used for marking the data value of each original logging curve data as an invalid value in response to the data value being greater than the first threshold or less than the second threshold.

[0135] In some embodiments, the pre-processing module further comprises:

[0136] A calculation unit, used for calculating the missing ratio of each original well logging curve based on the invalid value in each original well logging curve;

[0137] The sorting unit is used to obtain multiple original logging curves after missing sorting based on the missing ratio of each original logging curve. The missing sorting includes sorting from high to low missing ratio and sorting from low to high missing ratio.

[0138] In some embodiments, the plurality of low-frequency logging curves include a target low-frequency logging curve, the target low-frequency logging curve includes a first layer segment and a second layer segment, wherein the first layer segment is a layer segment where a valid value is located, and the second layer segment is a layer segment where an invalid value is located, and the first prediction module 502 includes:

[0139] The first processing unit is used to perform the following first processing process for the target low-frequency logging curve: using the data of the first layer segment of the target low-frequency logging curve and the data of the corresponding layer segments of other low-frequency logging curves other than the target low-frequency logging curve at the same depth position as the first layer segment as the output and input of the model training respectively to perform model training to obtain a first training model, wherein the missing data of the corresponding layer segments of other low-frequency logging curves at the same depth position as the first layer segment are obtained by polynomial interpolation; substituting the data of the corresponding layer segments of other low-frequency logging curves other than the target low-frequency logging curve at the same depth position as the second layer segment into the first training model to obtain the predicted data of the second layer segment of the target low-frequency logging curve; filling the predicted data of the second layer segment of the target low-frequency logging curve into the second layer segment of the target low-frequency logging curve to obtain a low-frequency logging curve without missing data;

[0140] The first execution unit is used to take each low-frequency logging curve as a target low-frequency logging curve one by one, execute the above first processing process, and obtain multiple low-frequency logging curves without missing data.

[0141] In some embodiments, the first execution unit includes:

[0142] A first execution subunit is used to sort by missing values, take each low-frequency logging curve as a target low-frequency logging curve, and execute the first processing as above, wherein the missing order is obtained according to the missing ratio of the original logging curve, and the missing ratio of the original logging curve is obtained according to the invalid value in the original logging curve;

[0143] Among them, the missingness ranking is sorted from low to high missingness ratio.

[0144] In some embodiments, the plurality of transition logging curves include a target transition logging curve, the target transition logging curve includes a third layer segment and a fourth layer segment, wherein the third layer segment is a layer segment where a valid value is located, and the fourth layer segment is a layer segment where an invalid value is located, and the second prediction module 504 includes:

[0145] The second processing unit is used to perform the following second processing process on the target transition logging curve: using the data of the third layer segment of the target transition logging curve and the data of the corresponding layer segments of other transition logging curves other than the target transition logging curve at the same depth position as the third layer segment as the output and input of the model training respectively to perform model training to obtain a second training model; substituting the data of the corresponding layer segments of other transition logging curves other than the target transition logging curve at the same depth position as the fourth layer segment into the second training model to obtain the predicted data of the fourth layer segment of the target transition logging curve; filling the predicted data of the fourth layer segment of the target transition logging curve into the fourth layer segment of the target transition logging curve to obtain a complete logging curve without missing data;

[0146] The second execution unit is used to take each transition logging curve as a target transition logging curve one by one, and execute the above second processing process to obtain multiple complete logging curves without missing data.

[0147] In some embodiments, the second execution unit includes:

[0148] The second execution subunit is used to sort by missing values, take each transition logging curve as a target transition logging curve, and execute the second processing as above, wherein the missing sort is obtained according to the missing ratio of the original logging curve, and the missing ratio of the original logging curve is obtained according to the invalid value in the original logging curve;

[0149] Among them, the missingness ranking is sorted from low to high missingness ratio.

[0150] Therefore, the device for filling missing well logging data provided by the embodiment of the present application makes full use of the existing effective well logging data, considers the overall trend of the missing data, combines with model training, and predicts the well logging curve with complete data, so that the reliability and accuracy of filling missing well logging data are improved.

[0151] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory including a program code, and the program code can be executed by a processor in an electronic device to complete the method for filling the missing well logging data in the above embodiment. For example, the non-volatile computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0152] A person skilled in the art can understand that all or part of the steps of implementing the above embodiments can be completed by hardware or by instructing related hardware through a program. The above program can be stored in a computer non-volatile computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.

[0153] In the present application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. The term "plurality" refers to two or more than two, unless otherwise clearly defined.

[0154] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the present application disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only.

[0155] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for filling missing well logging data, characterized in that: The method comprises: The multiple original well logging curves in the study area are subjected to wave number domain low-pass filtering to obtain multiple low-frequency well logging curves, in which each original well logging curve and each low-frequency well logging curve has missing data; Performing model training on the plurality of low-frequency logging curves to predict and obtain a plurality of low-frequency logging curves without missing data; For each part of the original well logging curve with missing data, the part is filled with data at the same depth position on the corresponding low-frequency well logging curve without missing data, so as to obtain multiple transition well logging curves; Model training is performed on the multiple transition logging curves to predict and obtain multiple complete logging curves without missing data.

2. The method for filling missing well logging data according to claim 1, characterized in that: Before performing wavenumber domain low-pass filtering on the multiple original well logging curves in the study area to obtain multiple low-frequency well logging curves, the method further includes: Acquire well logging data in the area to be studied, wherein the well logging data includes a plurality of original well logging curve data; The multiple original logging curve data are preprocessed to obtain multiple original logging curves.

3. The method for filling missing well logging data according to claim 2, characterized in that: The preprocessing of the plurality of original well logging curve data to obtain the plurality of original well logging curves comprises: Detecting each original logging curve data, marking invalid values ​​in each original logging curve, wherein the data corresponding to the invalid values ​​represent missing data; For each original logging curve data, the data point at the current depth position is marked as an invalid value, and the data point at the current depth position of each original logging curve is deleted.

4. The method for filling missing well logging data according to claim 3, characterized in that: The detecting of each original logging curve data and marking of invalid values ​​in each original logging curve includes: For each piece of original logging curve data, in response to the data value being equal to a preset value, marking the data value as an invalid value; or, For each piece of original logging curve data, in response to a data value being greater than a first threshold or less than a second threshold, the data value is marked as an invalid value.

5. The method for filling missing well logging data according to claim 2, characterized in that: The preprocessing of the plurality of original well logging curve data to obtain the plurality of original well logging curves further comprises: Based on the invalid values ​​in each original logging curve, the missing ratio of each original logging curve is calculated; Based on the missing ratio of each original well logging curve, a plurality of original well logging curves after missing ratio sorting are obtained, wherein the missing ratio sorting includes sorting from high to low missing ratio and sorting from low to high missing ratio.

6. The method for filling missing well logging data according to claim 1, characterized in that: The multiple low-frequency logging curves include a target low-frequency logging curve, the target low-frequency logging curve includes a first layer segment and a second layer segment, wherein the first layer segment is a layer segment where a valid value is located, and the second layer segment is a layer segment where an invalid value is located, and the model training is performed on the multiple low-frequency logging curves to predict multiple low-frequency logging curves without missing data, including: For the target low-frequency logging curve, the following first processing process is performed: The data of the first layer segment of the target low-frequency logging curve and the data of the corresponding layer segments of other low-frequency logging curves other than the target low-frequency logging curve at the same depth as the first layer segment are respectively used as the output and input of the model training to obtain a first training model, wherein the missing data of the corresponding layer segments of other low-frequency logging curves at the same depth as the first layer segment are obtained by using polynomial interpolation; Substituting data of corresponding layers of other low-frequency logging curves, except the target low-frequency logging curve, at the same depth as the second layer segment into the first training model to obtain predicted data of the second layer segment of the target low-frequency logging curve; Filling the predicted data of the second layer segment of the target low-frequency logging curve into the second layer segment of the target low-frequency logging curve to obtain a low-frequency logging curve without missing data; Each low-frequency logging curve is taken as a target low-frequency logging curve one by one, and the first processing process as described above is performed to obtain the plurality of low-frequency logging curves without missing data.

7. The method for filling missing well logging data according to claim 6, characterized in that: The step of taking each low-frequency logging curve as a target low-frequency logging curve and executing the first processing as described above comprises: According to the missing order, each low-frequency logging curve is used as a target low-frequency logging curve, and the first processing process as described above is performed, wherein the missing order is obtained according to the missing ratio of the original logging curve, and the missing ratio of the original logging curve is obtained according to the invalid value in the original logging curve; The missingness ranking is a ranking from low to high missingness ratio.

8. The method for filling missing well logging data according to claim 1, characterized in that: The multiple transition logging curves include a target transition logging curve, the target transition logging curve includes a third layer segment and a fourth layer segment, wherein the third layer segment is a layer segment where a valid value is located, and the fourth layer segment is a layer segment where an invalid value is located, and the model training is performed on the multiple transition logging curves to predict multiple complete logging curves without missing data, including: For the target transition logging curve, the following second processing process is performed: The model training is performed using the data of the third layer segment of the target transition logging curve and the data of the corresponding layer segments of other transition logging curves other than the target transition logging curve at the same depth as the third layer segment as the output and input of the model training, respectively, to obtain a second training model; Substituting data of corresponding layers of other transition logging curves except the target transition logging curve at the same depth as the fourth layer segment into the second training model to obtain predicted data of the fourth layer segment of the target transition logging curve; Filling the predicted data of the fourth layer segment of the target transition logging curve into the fourth layer segment of the target transition logging curve to obtain a complete logging curve without missing data; Each transition logging curve is taken as a target transition logging curve one by one, and the second processing process as described above is performed to obtain the plurality of complete logging curves without missing data.

9. The method for filling missing well logging data according to claim 8, characterized in that: The second processing step of taking each transition logging curve as the target transition logging curve one by one and performing the second processing step as described above comprises: According to the missing order, each transition logging curve is used as a target transition logging curve, and the second processing process as described above is performed, wherein the missing order is obtained according to the missing ratio of the original logging curve, and the missing ratio of the original logging curve is obtained according to the invalid value in the original logging curve; The missingness ranking is a ranking from low to high missingness ratio.

10. A device for filling missing well logging data, characterized in that: The device comprises: A low-pass filtering module is used to perform low-pass filtering in the wave number domain on multiple original well logging curves in the study area to obtain multiple low-frequency well logging curves, wherein each original well logging curve and each low-frequency well logging curve has missing data; A first prediction module is used to perform model training on the plurality of low-frequency logging curves to predict a plurality of low-frequency logging curves without missing data; A filling module is used to fill in the missing data of each original logging curve with the data at the same depth position on the corresponding low-frequency logging curve without missing data, so as to obtain multiple transition logging curves; The second prediction module is used to perform model training on the plurality of transition logging curves to predict a plurality of complete logging curves without missing data.