Construction method and application of intelligent interpretation model for mineral content logging

Abstract

The invention relates to the technical field of oilfield development, and discloses a construction method and application of a mineral content well logging intelligent interpretation model, reservoirs in a target rock area are classified by using a ternary comprehensive classification method composed of lithology, porosity and oil-bearing level, interference of the lithology, porosity and oil-bearing level on mineral content well logging interpretation can be reduced, and the interpretation accuracy is improved. The model learning precision is improved; multi-wavelet decomposition and reconstruction are carried out on logging curve data, all effective information of an original logging curve is reserved, the redundant degree between wavelet components is reduced, and the logging curve difference caused by the non-mineral content is effectively reduced; and based on the idea of an integration algorithm, training the intelligent mineral content logging interpretation model of each reservoir by using an optimal machine learning algorithm corresponding to different types of reservoirs, and then integrating the intelligent mineral content logging interpretation models of different types of reservoirs to obtain the intelligent mineral content logging interpretation model of the target rock area. The overall accurate mineral content identification of the target rock area can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of oilfield development, and in particular to a method for constructing a mineral content well logging intelligent interpretation model and its application. Background Art

[0002] Mineral quantitative interpretation is a method for quantitatively identifying minerals contained in rocks. Traditional mineral quantitative interpretation methods are quantitative interpretations based on core analysis. The mineral quantitative analysis of the core obtained from the coring well is completed by direct analysis (such as mineral thin section observation, electron microscope observation, etc.) and indirect analysis (such as XRD diffraction, spectral identification, etc.). In addition, some scholars have applied image recognition algorithms to perform image recognition on core images (such as microscopic thin section photos of the core and scanning electron microscope photos) to achieve quantitative identification of minerals in the core. However, the above methods are all for coring wells. Due to the small number of coring (only a small number of wells are cored), the high cost of experimental analysis, and the poor representativeness of experimental samples (local sampling cannot represent the overall reservoir characteristics), this type of method cannot be applied to the entire area and cannot predict the overall mineral content of the reservoir.

[0003] Therefore, there is an urgent need for a method for constructing an intelligent interpretation model of mineral content logging, so as to construct an intelligent interpretation model of mineral content logging that can be used to accurately predict the overall mineral content of the reservoir. Summary of the invention

[0004] The present invention provides a method for constructing a mineral content well logging intelligent interpretation model and its application, so as to solve the defect that the traditional mineral quantitative interpretation method cannot accurately predict the overall mineral content of the reservoir.

[0005] The present invention provides a method for constructing a mineral content well logging intelligent interpretation model, comprising:

[0006] Obtain lithology data, porosity data, and oil content grade data of the target rock area;

[0007] According to the lithology data, porosity data and oil content grade data of the target rock area, the reservoirs in the target rock area are classified to obtain several types of reservoirs;

[0008] Obtain mineral type data, mineral content data, and well logging curve data for each of several types of reservoirs;

[0009] According to the mineral type data and mineral content data of each type of reservoir in the several types of reservoirs, multi-wavelet decomposition and reconstruction are performed on the well logging curve data of each type of reservoir in the several types of reservoirs to obtain the reconstructed curve data of each type of reservoir in the several types of reservoirs;

[0010] According to the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in several types of reservoirs, the model is made to learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve, and the mineral content logging intelligent interpretation model is obtained.

[0011] In one embodiment, the lithology data includes data of any one of the following or any combination thereof: sandstone, siltstone, mudstone; the porosity data includes data of any one of the following or any combination thereof: high porosity, medium porosity, low porosity; the oil grade data includes data of any one of the following or any combination thereof: high oil saturation, medium oil saturation, low oil saturation.

[0012] In one embodiment, the reservoirs in the target rock area are classified according to the lithology data, porosity data, and oil content grade data of the target rock area to obtain several types of reservoirs, including:

[0013] A cluster analysis was performed on the lithology data, porosity data, and oil content grade data of the target rock area, and the reservoirs in the target rock area were classified into Class A reservoirs (mainly oil layers), Class B reservoirs (mainly water-flooded layers), Class C reservoirs (mainly poor oil layers), Class D reservoirs (mainly oil-dry layers), and Class E reservoirs (mainly tight layers).

[0014] In one embodiment, the method of classifying the reservoirs in the target rock area according to the lithology data, porosity data, and oil content grade data of the target rock area to obtain several types of reservoirs also includes:

[0015] According to the lithology data, porosity data and oil content grade data of the target rock area, the lithology data, porosity data and oil content grade data of Class A reservoir, Class B reservoir, Class C reservoir, Class D reservoir and Class E reservoir are obtained respectively;

[0016] A secondary clustering analysis was performed on the lithology data, porosity data, and oil content grade data of Class A reservoirs, Class B reservoirs, Class C reservoirs, Class D reservoirs, and Class E reservoirs. Class A reservoirs, Class B reservoirs, and Class C reservoirs were classified as Class I reservoirs, and Class D reservoirs and Class E reservoirs were classified as Class II reservoirs.

[0017] It should be noted that if there are enough core wells in the study area to provide enough training samples for Class A reservoirs, Class B reservoirs, Class C reservoirs, Class D reservoirs, and Class E reservoirs, secondary clustering may not be performed.

[0018] In one embodiment, the multi-wavelet decomposition and reconstruction of the well logging curve data of each type of reservoir in the several types of reservoirs is performed according to the mineral type data and mineral content data of each type of reservoir in the several types of reservoirs to obtain the reconstructed curve data of each type of reservoir in the several types of reservoirs, including:

[0019] Performing a first preprocessing on the well logging curve data of each type of reservoir in the plurality of types of reservoirs;

[0020] Using discrete wavelet transform method, multi-wavelet decomposition is performed on the pre-processed logging curve data of each type of reservoir in several types of reservoirs to obtain several wavelet components of each type of reservoir in several types of reservoirs;

[0021] According to the mineral type data and mineral content data of each type of reservoir in several types of reservoirs, the sensitivity difference between several sub-wave components of each type of reservoir in several types of reservoirs and the mineral content of different types of minerals is analyzed by correlation analysis method, and several sensitive sub-wave components of each type of reservoir in several types of reservoirs are screened out;

[0022] Well logging curve reconstruction is performed according to a number of sensitive sub-wave components of each type of reservoir in the number of types of reservoirs, so as to obtain the reconstructed curve data of each type of reservoir in the number of types of reservoirs.

[0023] In one embodiment, the first preprocessing of the well logging curve data of each type of reservoir in the plurality of types of reservoirs comprises:

[0024] normalizing well log data for each of several types of reservoirs; and / or,

[0025] Lithoelectric correction is performed on the well logging data of each of several types of reservoirs.

[0026] In one embodiment, the discrete wavelet transform method is used to perform multi-wavelet decomposition on the pre-processed logging curve data of each type of reservoir in the several types of reservoirs to obtain several wavelet components of each type of reservoir in the several types of reservoirs, including:

[0027] Using the discrete wavelet transform method, according to the multi-wavelet decomposition expression, the pre-processed logging curve data of each type of reservoir in several types of reservoirs are decomposed by multi-wavelet, and several wavelet components of each type of reservoir in several types of reservoirs are obtained. Among them, the multi-wavelet decomposition expression is:

[0028] f(i,j)=f1(i,j)+f2(i,j)+…+f n (i,j)

[0029]

[0030] In the multi-wavelet decomposition expression, f(i,j) represents the sum of the wavelet components under scale i and translation j, that is, the discrete wavelet transform result. Scale i is used to control the scale of the wavelet and affect the expansion or compression of the wavelet. A larger i corresponds to a wider wavelet (capturing low-frequency components), while a smaller i corresponds to a narrower wavelet (capturing high-frequency components). Translation j is used to determine the spatial position of the wavelet in the logging curve data, i,j∈Z, f1(i,j), f2(i,j), …, f n (i, j) represents several wavelet components of each type of reservoir, t represents the spatial variable of the logging curve, f(t) represents the input signal, represents the mother wavelet. Note that scale i and translation j are two variables, so f n (i,j) includes several wavelet components.

[0031] In one embodiment, the well logging curve reconstruction is performed according to several sensitive wavelet components of each type of reservoir in several types of reservoirs to obtain the reconstructed curve data of each type of reservoir in several types of reservoirs, including:

[0032] According to the mineral type data and mineral content data of each type of reservoir in several types of reservoirs, the sensitivity difference between several sub-wave components of each type of reservoir in several types of reservoirs and the mineral content of different types of minerals is analyzed by correlation analysis method, and several sensitive sub-wave components of each type of reservoir in several types of reservoirs are screened out;

[0033] A number of sensitive sub-wave components of each type of reservoir in a number of types of reservoirs are added together to reconstruct the logging curve, thereby obtaining the reconstructed curve data of each type of reservoir in the number of types of reservoirs.

[0034] In one embodiment, according to the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in several types of reservoirs, a machine learning algorithm is used to make the model learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve to obtain a mineral content logging intelligent interpretation model, including:

[0035] performing a second preprocessing on the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in the plurality of types of reservoirs;

[0036] For each of the several types of reservoirs, the mineral type data, mineral content data and reconstruction curve data after the second preprocessing are used, a plurality of machine learning algorithms are used to enable the model to learn the mapping relationship between the mineral content of different types of minerals in the reservoir and the reconstruction curve, and determine the optimal machine learning algorithm with the best model learning effect for each of the several types of reservoirs;

[0037] For each of the several types of reservoirs, the mineral type data, mineral content data and reconstruction curve data after the second preprocessing are performed, and an optimal machine learning algorithm with the best model learning effect corresponding to each of the several types of reservoirs is used to obtain a mineral content well logging intelligent interpretation model corresponding to each of the several types of reservoirs;

[0038] The mineral content logging intelligent interpretation model corresponding to each type of reservoir in several types of reservoirs is integrated to obtain the mineral content logging intelligent interpretation model of the target rock area.

[0039] In one embodiment, the method uses a machine learning algorithm to make the model learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve according to the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in the several types of reservoirs, so as to obtain a mineral content logging intelligent interpretation model. After the mineral type data, mineral content data and reconstruction curve data after the second preprocessing for each type of reservoir in the several types of reservoirs are obtained by using the optimal machine learning algorithm with the best model learning effect corresponding to each type of reservoir in the several types of reservoirs, the method further includes any of the following items or any combination thereof:

[0040] By using cross validation, the credibility of the intelligent well logging interpretation model for mineral content in each type of reservoir is evaluated;

[0041] Using Bayesian functions, hyperparameter optimization is performed on the intelligent interpretation model of mineral content logging for each type of reservoir corresponding to several types of reservoirs.

[0042] Using blind well test, the prediction accuracy and generalization ability of the intelligent interpretation model of mineral content logging corresponding to each type of reservoir are tested.

[0043] The present invention also provides a system for constructing a mineral content well logging intelligent interpretation model, comprising:

[0044] The first data acquisition module is used to obtain lithology data, porosity data, and oil content grade data of the target rock area;

[0045] The component module is used to classify the reservoirs in the target rock area according to the lithology data, porosity data and oil content grade data of the target rock area to obtain several types of reservoirs;

[0046] The second data acquisition module is used to: acquire mineral type data, mineral content data, and well logging curve data of each type of reservoir in several types of reservoirs;

[0047] The decomposition and reconstruction module is used to: perform multi-wavelet decomposition and reconstruction on the well logging curve data of each type of reservoir in the several types of reservoirs according to the mineral type data and mineral content data of each type of reservoir in the several types of reservoirs, so as to obtain the reconstructed curve data of each type of reservoir in the several types of reservoirs;

[0048] The training module is used to: based on the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in several types of reservoirs, use the machine learning algorithm to enable the model to learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve, so as to obtain the mineral content logging intelligent interpretation model.

[0049] The present invention also provides a mineral content well logging intelligent interpretation method, comprising:

[0050] Receive logging curve data of the rock area to be tested;

[0051] According to any of the above methods for constructing the intelligent interpretation model for mineral content logging, multi-wavelet decomposition and reconstruction are performed on the logging curve data of the rock area to be tested to obtain the reconstructed curve data of the rock area to be tested;

[0052] According to the reconstructed curve data of the rock area to be tested, the mineral content logging intelligent interpretation model of the target rock area is obtained by the construction method of the mineral content logging intelligent interpretation model described in any of the above items, and the mineral content logging interpretation results of the rock area to be tested are obtained. The mineral content logging interpretation results include logging and complete mineral content data.

[0053] The present invention also provides a mineral content well logging intelligent interpretation system, comprising:

[0054] The data receiving module is used to: receive the logging curve data of the rock area to be tested sent by at least one terminal;

[0055] A preprocessing module is used to: perform multi-wavelet decomposition and reconstruction on the logging curve data of the rock area to be measured according to any of the above-mentioned methods for constructing the intelligent interpretation model of mineral content logging, so as to obtain the reconstructed curve data of the rock area to be measured;

[0056] A prediction module is used to obtain the mineral content logging interpretation result of the rock area to be tested based on the reconstructed curve data of the rock area to be tested and the mineral content logging intelligent interpretation model of the target rock area obtained by the construction method of the mineral content logging intelligent interpretation model described in any one of the above items;

[0057] The data output module is used to send the logging interpretation results of the mineral content of the rock area to be tested back to at least one terminal.

[0058] The present invention also provides an electronic device, comprising a processor and a memory storing a computer program, wherein when the processor executes the computer program, the method for constructing a mineral content logging intelligent interpretation model and / or the mineral content logging intelligent interpretation method described above is implemented.

[0059] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for constructing a mineral content logging intelligent interpretation model and / or the method for intelligent interpretation of mineral content logging described above is implemented.

[0060] The present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute any of the above-mentioned methods for constructing a mineral content logging intelligent interpretation model and / or the mineral content logging intelligent interpretation method.

[0061] The method for constructing a mineral content well logging intelligent interpretation model and its application provided by the present invention at least bring the following beneficial effects:

[0062] The reservoirs in the target rock area are classified using a ternary comprehensive classification method consisting of lithology, porosity, and oil content level, which can reduce the interference of lithology, porosity, and oil content level on the interpretation of mineral content logging, improve the learning accuracy of the intelligent interpretation model of mineral content logging, and further achieve accurate mineral content prediction;

[0063] Multi-wavelet decomposition and reconstruction of logging curve data can not only retain all effective information of the original logging curve, but also reduce the redundancy between the wavelet components and effectively reduce the logging curve differences caused by non-mineral content;

[0064] Based on the idea of ​​integrated algorithm, the optimal machine learning algorithm corresponding to different types of reservoirs is first used to train the mineral content logging intelligent interpretation model of each reservoir, and then the mineral content logging intelligent interpretation models of different types of reservoirs are integrated to obtain the mineral content logging intelligent interpretation model of the target rock area, which can realize the accurate mineral content identification of the target rock area as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1This is one of the flow charts of a method for constructing an intelligent interpretation model for mineral content logging provided by the present invention.

[0067] Figure 2 The second flowchart of the method for constructing a mineral content well logging intelligent interpretation model provided by the present invention.

[0068] Figure 3 Schematic diagram of the process for reservoir classification.

[0069] Figure 4 Example of results for classifying a reservoir.

[0070] Figure 5 Schematic diagram of the process of multi-wavelet decomposition and reconstruction.

[0071] Figure 6 This is an example of the result of multi-wavelet decomposition.

[0072] Figure 7 Example of the training process for a smart interpretation model for mineral content well logs.

[0073] Figure 8 Example plot of interpreted results for mineral content well logs.

[0074] Fig. 9 A structural schematic diagram of a mineral content well logging intelligent interpretation system provided by the present invention.

[0075] Fig.10 This is a schematic structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0076] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments, and they should not be understood as limitations on the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In the description of the present invention, it should be understood that the terms used are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0077] Combine the following Figure 1-Figure 10 The method for constructing the intelligent interpretation model of mineral content logging provided by the present invention and its application are described. It should be noted that the execution subject of the method for constructing the intelligent interpretation model of mineral content logging and the intelligent interpretation method of mineral content logging provided by the present invention can be any network side device / terminal side device that meets the technical requirements, such as an intelligent interpretation device for mineral content logging, etc.

[0078] Figure 1-2 The figure is a flow chart of the method for constructing the intelligent interpretation model of mineral content logging provided by the present invention and its application. Figure 1 The present invention provides a method for constructing a mineral content well logging intelligent interpretation model, which may include:

[0079] Step S110, obtaining lithology data, porosity data, and oil grade data of the target rock area, wherein the lithology data includes data of any one of the following or any combination thereof: sandstone, siltstone, mudstone; the porosity data includes data of any one of the following or any combination thereof: high porosity, medium porosity, low porosity; the oil grade data includes data of any one of the following or any combination thereof: high oil saturation, medium oil saturation, low oil saturation.

[0080] Specifically, the lithology data, porosity data, and oil content grade data of the target rock area can be obtained from the relevant core data database of the target rock area, or from the historical exploration data of the target rock area. The core data include XRD diffraction data, scanning electron microscope data, displacement experiment data, and conventional rock thin section identification reports.

[0081] Step S120: classify the reservoirs in the target rock area according to the lithology data, porosity data, and oil content grade data of the target rock area to obtain several types of reservoirs.

[0082] In one embodiment, step S120 can perform cluster analysis on the lithology data, porosity data, and oil content grade data of the target rock area, and classify the reservoirs in the target rock area into Class A reservoirs (the example is mainly oil layers), Class B reservoirs (the example is mainly water-flooded layers), Class C reservoirs (the example is mainly poor oil layers), Class D reservoirs (the example is mainly oil-dry layers), and Class E reservoirs (the example is mainly dense layers).

[0083] In another embodiment, if the number of coring wells in the target rock area is small and cannot provide enough training samples for the Class A reservoir, the Class B reservoir, the Class C reservoir, the Class D reservoir, and the Class E reservoir, step S120 may include:

[0084] According to the lithology data, porosity data and oil content grade data of the target rock area, the lithology data, porosity data and oil content grade data of Class A reservoir, Class B reservoir, Class C reservoir, Class D reservoir and Class E reservoir are obtained respectively;

[0085] A secondary clustering analysis was performed on the lithology data, porosity data, and oil content grade data of Class A reservoirs, Class B reservoirs, Class C reservoirs, Class D reservoirs, and Class E reservoirs. Class A reservoirs, Class B reservoirs, and Class C reservoirs were classified as Class I reservoirs, and Class D reservoirs and Class E reservoirs were classified as Class II reservoirs.

[0086] See also Figure 3In this embodiment, the logging response characteristics of the logging data, core data, and dynamic data can be manually characterized, the main controlling geological factors can be analyzed, and the non-mineral influencing factors can be excluded to obtain the main controlling geological factors of the logging, namely, porosity, lithology, and oil content grade, and then the lithology data, porosity data, and oil content grade data of the target rock area can be obtained. Then, the lithology data, porosity data, and oil content grade data of the target rock area can be clustered once or twice by the K-Means clustering method. The result example of the secondary clustering analysis can be seen in Figure 4 .

[0087] Step S130: Obtain mineral type data, mineral content data, and well logging curve data for each of several types of reservoirs.

[0088] Specifically, step S130 can obtain mineral type data, mineral content data, and well logging curve data for each type of reservoir in several types of reservoirs from the relevant core data database, dynamic database, or historical exploration database of the target rock area. The conventional thin section identification report of the rock can be used as the main basis, combined with other core data, to further clarify the mineral types in the rocks of the target rock area, that is, mainly rock fragments, quartz, potassium feldspar and plagioclase, and also contain a small amount of kaolinite. At the same time, the contents of different types of minerals in the corresponding rock positions in the identification report can be sorted out to obtain mineral type data and mineral content data.

[0089] Step S140, based on the mineral type data and mineral content data of each of the several types of reservoirs, multi-wavelet decomposition and reconstruction are performed on the logging curve data of each of the several types of reservoirs to obtain the reconstructed curve data of each of the several types of reservoirs.

[0090] In one embodiment, see Figure 5 , step S140 may include:

[0091] Performing a first preprocessing on the logging curve data of each type of reservoir in several types of reservoirs, wherein the first preprocessing includes: standardizing the logging curve data of each type of reservoir in several types of reservoirs; and / or performing rock-electric correction on the logging curve data of each type of reservoir in several types of reservoirs, wherein, due to the different expansion coefficients of the drill pipe and the cable, there is a certain offset between the logging curve and the core depth, and at the same time, due to the systematic error of the logging instrument (caused by the systematic error or the logging instrument model), there are certain differences in the logging curves of different wells. Therefore, in this embodiment, the logging curve data of the target rock area is standardized and rock-electrically corrected. Specifically, by comprehensively analyzing the core description, logging, analysis and testing data, typical lithology or lithology combination, logging response "outlier" or curve combination, part of the mineral content (such as calcium) and the like can be selected as the homing mark or basis, based on the logging data, to ensure the one-to-one correspondence between the core data and the logging data;

[0092] Using the discrete wavelet transform method, according to the multi-wavelet decomposition expression, the pre-processed logging curve data of each type of reservoir in several types of reservoirs are decomposed by multi-wavelet, and several wavelet components of each type of reservoir in several types of reservoirs are obtained. Among them, the logging curve can be regarded as a composite wave, and can be decomposed by multi-wavelet using continuous wavelet transform (the mother wavelet can be selected from Marr, Morlet, Ricker, Meyer, Harr, etc.), and the multi-wavelet decomposition expression is:

[0093] f(i,j)=f1(i,j)+f2(i,j)+…+f n (i,j)

[0094]

[0095] In the multi-wavelet decomposition expression, f(i,j) represents the sum of the wavelet components under scale i and translation j, that is, the discrete wavelet transform result. Scale i is used to control the scale of the wavelet and affect the expansion or compression of the wavelet. A larger i corresponds to a wider wavelet (capturing low-frequency components), while a smaller i corresponds to a narrower wavelet (capturing high-frequency components). Translation j is used to determine the spatial position of the wavelet in the logging curve data, i,j∈Z, f1(i,j), f2(i,j), …, f n (i, j) represents several wavelet components of each type of reservoir, t represents the spatial variable of the logging curve, f(t) represents the input signal, represents the mother wavelet; note that scale i and translation j are two variables, so f n (i,j) includes several wavelet components, and n is generally 2 to 4;

[0096] According to the mineral type data and mineral content data of each type of reservoir in several types of reservoirs, the sensitivity difference between several sub-wave components of each type of reservoir in several types of reservoirs and the mineral content of different types of minerals is analyzed by correlation analysis method, and several sensitive sub-wave components of each type of reservoir in several types of reservoirs are screened out;

[0097] A number of sensitive sub-wave components of each type of reservoir in a number of types of reservoirs are added together to reconstruct the logging curve, thereby obtaining the reconstructed curve data of each type of reservoir in the number of types of reservoirs.

[0098] Taking the natural gamma (GR) decomposition of the well logging curve as an example, the i value in the multi-wavelet decomposition expression is -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, and the n value is 2. Substituting the above i value into the multi-wavelet decomposition expression, the well logging curve will be decomposed into 20 wavelet components, namely f1(-6,j), f1(-5,j), f1(-4,j), f1(-3,j), f1(-2,j), f1(-1,j), f1(0,j), f1(1,j), f1(2,j), f1(3,j), f2(-6,j), f2(-5,j), f2(-4,j), f2(-3,j), f2(-2,j), f2(-1,j), f2(0,j), f2(1,j), f2(2,j), f2(3,j). The example diagram of multi-wavelet decomposition can be seen in Figure 6 .

[0099] Through correlation analysis, the sensitivity difference of each sub-wave component to the mineral content is determined. When analyzing the sensitivity of each sub-wave component of the logging curve to the mineral content, this embodiment takes into account the resolution difference between the two, that is, the mineral content is coarsened according to the main frequency of the sub-wave component; for example, the main frequency of the mother wavelet of the frequency division is 8Hz (there are 8 wavelengths per meter of the logging curve length), and the resolution is calculated according to 1 wavelength of 0.125m. The mineral content data obtained from the core needs to be calculated according to the average value of the mineral content at a spacing of 0.125m.

[0100] In this implementation scheme, taking the 10 wavelet components obtained by the above-mentioned natural gamma (GR) decomposition as an example, the correlation analysis results show that the wavelet components f1(-6,j), f1(-1,j), f1(3,j), f2(-6,j), f2(0,j), f2(2,j), f2(3,j) have no correlation with the carbonate mineral content, and the other wavelet components have varying degrees of correlation. Therefore, f1(-5,j), f1(-4,j), f1(-3,j), f1(-2,j), f1(0,j), f1(1,j), f1(2,j), f2(-5,j), f2(-4,j), f2(-3,j), f2(-2,j), f2(-1,j), f2(1,j) are selected for natural gamma (GR) curve reconstruction, and the reconstruction expression is as follows:

[0101] W(i,j)=

[0102] f1(-5,j)+f1(-4,j)+f1(-3,j)+f1(-2,j)+f1(0,j)+f1(1,j)+f1(2,j)+f2(-5,j)+f2(-4,j)+

[0103] f2(-3,j)+f2(-2,j)+f2(-1,j)+f2(1,j)

[0104] In the formula, W(i,j) represents the reconstructed logging curve.

[0105] Step S150: Based on the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in several types of reservoirs, a machine learning algorithm is used to enable the model to learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve, so as to obtain a mineral content logging intelligent interpretation model.

[0106] In one embodiment, see Figure 7 , step S150 may include:

[0107] Performing a second preprocessing on the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in the plurality of types of reservoirs, wherein the second preprocessing may include feature engineering, data dimension reduction and other processing;

[0108] For each of the several types of reservoirs, the mineral type data, mineral content data and reconstruction curve data after the second preprocessing are respectively used to make the model learn the mapping relationship between the mineral content of different types of minerals in the reservoir and the reconstruction curve, and determine the optimal machine learning algorithm with the best model learning effect for each of the several types of reservoirs. Specifically, the multiple machine learning algorithms may be, for example, out-of-bag random forest B-RF, gradient boosting decision tree XGBoost, etc., wherein B-RF can effectively reduce the risk of overfitting by integrating multiple decision tree models, and evaluate the model performance by out-of-bag data, significantly improving the stability and generalization ability of the model. XGBoost can effectively capture the complex nonlinear relationship in the data through the integration method of boosting trees, and has excellent prediction accuracy and the ability to process large-scale data. Furthermore, the optimal machine learning algorithm can be obtained by comparing the prediction effect of each model on each mineral, for example, B-RF is best for predicting quartz and rock fragments, and XGBoost is best for predicting potassium feldspar and plagioclase, etc.;

[0109] For each of the several types of reservoirs, the mineral type data, mineral content data and reconstruction curve data after the second preprocessing are performed, and an optimal machine learning algorithm with the best model learning effect corresponding to each of the several types of reservoirs is used to obtain a mineral content well logging intelligent interpretation model corresponding to each of the several types of reservoirs;

[0110] The mineral content logging intelligent interpretation model corresponding to each type of reservoir in several types of reservoirs is integrated to obtain the mineral content logging intelligent interpretation model of the target rock area.

[0111] Among them, after the mineral content logging intelligent interpretation model corresponding to each type of reservoir in several types is obtained through training, any one of the following items or any combination thereof may be performed first, and then the mineral content logging intelligent interpretation model corresponding to each type of reservoir in several types may be integrated:

[0112] By using cross-validation, the credibility of the intelligent well logging interpretation model for mineral content in each type of reservoir corresponding to several types of reservoirs is evaluated to ensure the reliability of the model;

[0113] Using the Bayesian function, the hyperparameters of the mineral content logging intelligent interpretation model corresponding to each type of reservoir in several types of reservoirs are optimized. In this embodiment, if the credibility on the test set is less than 90%, the model is qualified, otherwise the cross-validation and hyperparameter optimization are repeated;

[0114] The blind well test is used to test the prediction accuracy and generalization ability of the mineral content logging intelligent interpretation model corresponding to each type of reservoir in several types of reservoirs. In this embodiment, if the blind well accuracy reaches 85%, the model accuracy is qualified, otherwise the cross-validation, hyperparameter optimization and blind well test are repeated.

[0115] After training the intelligent interpretation model of mineral content logging, the logging curve data of the rock area to be tested can be first decomposed and reconstructed by multi-wavelet according to the above method to obtain the reconstructed curve data of the rock area to be tested, and then the mineral content logging interpretation results of the rock area to be tested can be obtained according to the reconstructed curve data of the rock area to be tested by the intelligent interpretation model of mineral content logging. For examples of mineral content logging interpretation results, see Figure 8 .

[0116] It should be noted that the present invention does not limit the order of the steps of classifying the reservoirs in the target rock area and performing multi-wavelet decomposition and reconstruction of the logging curve in the method for constructing the intelligent interpretation model of mineral content logging. When the present invention is actually applied, the lithology data, porosity data, and oil content grade data of the target rock area can be first obtained to classify the reservoirs in the target rock area, and then the mineral type data, mineral content data, and logging curve data of each type of reservoir in several types of reservoirs can be obtained to perform multi-wavelet decomposition and reconstruction of the logging curve. Alternatively, the mineral type data, mineral content data, and logging curve data of the target rock area can be first obtained to perform multi-wavelet decomposition and reconstruction of the logging curve, and then the lithology data, porosity data, and oil content grade data of the target rock area can be obtained to classify the reservoirs in the target rock area. The purpose is to obtain the relationship between the reconstruction curves and mineral contents corresponding to different types of reservoirs, so that the model can learn the mapping relationship between the reconstruction curves under different types of reservoirs and the contents of different types of minerals.

[0117] The method for constructing a mineral content well logging intelligent interpretation model and its application provided by the present invention at least bring the following beneficial effects:

[0118] The reservoirs in the target rock area are classified using a ternary comprehensive classification method consisting of lithology, porosity, and oil content level, which can reduce the interference of lithology, porosity, and oil content level on the interpretation of mineral content logging, improve the learning accuracy of the intelligent interpretation model of mineral content logging, and further achieve accurate mineral content prediction;

[0119] Multi-wavelet decomposition and reconstruction of logging curve data can not only retain all effective information of the original logging curve, but also reduce the redundancy between the wavelet components and effectively reduce the logging curve differences caused by non-mineral content;

[0120] Based on the idea of ​​integrated algorithm, the optimal machine learning algorithm corresponding to different types of reservoirs is first used to train the mineral content logging intelligent interpretation model of each reservoir, and then the mineral content logging intelligent interpretation models of different types of reservoirs are integrated to obtain the mineral content logging intelligent interpretation model of the target rock area, which can realize the accurate mineral content identification of the target rock area as a whole.

[0121] The following describes a system for constructing a mineral content logging intelligent interpretation model and a mineral content logging intelligent interpretation system provided by the present invention. The system for constructing a mineral content logging intelligent interpretation model described below and the method for constructing a mineral content logging intelligent interpretation model described above can refer to each other.

[0122] The present invention provides a system for constructing a mineral content well logging intelligent interpretation model, which may include:

[0123] The first data acquisition module is used to obtain lithology data, porosity data, and oil content grade data of the target rock area;

[0124] The component module is used to classify the reservoirs in the target rock area according to the lithology data, porosity data and oil content grade data of the target rock area to obtain several types of reservoirs;

[0125] The second data acquisition module is used to: acquire mineral type data, mineral content data, and well logging curve data of each type of reservoir in several types of reservoirs;

[0126] The decomposition and reconstruction module is used to: perform multi-wavelet decomposition and reconstruction on the well logging curve data of each type of reservoir in the several types of reservoirs according to the mineral type data and mineral content data of each type of reservoir in the several types of reservoirs, so as to obtain the reconstructed curve data of each type of reservoir in the several types of reservoirs;

[0127] The training module is used to: based on the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in several types of reservoirs, use the machine learning algorithm to enable the model to learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve, so as to obtain the mineral content logging intelligent interpretation model.

[0128] Reference Fig. 9 The present invention provides a mineral content logging intelligent interpretation system, which may include:

[0129] The data receiving module is used to: receive the logging curve data of the rock area to be tested sent by at least one terminal;

[0130] A preprocessing module is used to: perform multi-wavelet decomposition and reconstruction on the logging curve data of the rock area to be measured according to any of the above-mentioned methods for constructing the intelligent interpretation model of mineral content logging, so as to obtain the reconstructed curve data of the rock area to be measured;

[0131] A prediction module is used to obtain the mineral content logging interpretation result of the rock area to be tested based on the reconstructed curve data of the rock area to be tested and the mineral content logging intelligent interpretation model of the target rock area obtained by the construction method of the mineral content logging intelligent interpretation model described in any one of the above items;

[0132] The data output module is used to send the logging interpretation results of the mineral content of the rock area to be tested back to at least one terminal.

[0133] Fig.10 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820 and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the mineral content logging intelligent interpretation method, which includes:

[0134] Receive logging curve data of the rock area to be tested;

[0135] According to any of the above methods for constructing the intelligent interpretation model for mineral content logging, multi-wavelet decomposition and reconstruction are performed on the logging curve data of the rock area to be tested to obtain the reconstructed curve data of the rock area to be tested;

[0136] According to the reconstructed curve data of the rock area to be tested, the mineral content logging intelligent interpretation model of the target rock area is obtained by any of the above-mentioned methods for constructing the mineral content logging intelligent interpretation model, and the mineral content logging interpretation results of the rock area to be tested are obtained.

[0137] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0138] On the other hand, the present invention further provides a computer program product, the computer program product comprising a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute a mineral content well logging intelligent interpretation method, the method comprising:

[0139] Receive logging curve data of the rock area to be tested;

[0140] According to any of the above methods for constructing the intelligent interpretation model for mineral content logging, multi-wavelet decomposition and reconstruction are performed on the logging curve data of the rock area to be tested to obtain the reconstructed curve data of the rock area to be tested;

[0141] According to the reconstructed curve data of the rock area to be tested, the mineral content logging intelligent interpretation model of the target rock area is obtained by any of the above-mentioned methods for constructing the mineral content logging intelligent interpretation model, and the mineral content logging interpretation results of the rock area to be tested are obtained.

[0142] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform a mineral content logging intelligent interpretation method, the method comprising:

[0143] Receive logging curve data of the rock area to be tested;

[0144] According to any of the above methods for constructing the intelligent interpretation model for mineral content logging, multi-wavelet decomposition and reconstruction are performed on the logging curve data of the rock area to be tested to obtain the reconstructed curve data of the rock area to be tested;

[0145] According to the reconstructed curve data of the rock area to be tested, the mineral content logging intelligent interpretation model of the target rock area is obtained by any of the above-mentioned methods for constructing the mineral content logging intelligent interpretation model, and the mineral content logging interpretation results of the rock area to be tested are obtained.

[0146] The device embodiments described above are only illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0147] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for constructing an intelligent interpretation model for mineral content logging, characterized in that: include: Obtain lithology data, porosity data, and oil content grade data of the target rock area; According to the lithology data, porosity data and oil content grade data of the target rock area, the reservoirs in the target rock area are classified to obtain several types of reservoirs; Obtain mineral type data, mineral content data, and well logging curve data for each of several types of reservoirs; According to the mineral type data and mineral content data of each type of reservoir in the several types of reservoirs, multi-wavelet decomposition and reconstruction are performed on the well logging curve data of each type of reservoir in the several types of reservoirs to obtain the reconstructed curve data of each type of reservoir in the several types of reservoirs; According to the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in several types of reservoirs, the model is made to learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve, and the mineral content logging intelligent interpretation model is obtained.

2. The method for constructing a mineral content well logging intelligent interpretation model according to claim 1, characterized in that: Lithology data include any one of the following or any combination of data: sandstone, siltstone, mudstone; porosity data include any one of the following or any combination of data: high porosity, medium porosity, low porosity; oil grade data include any one of the following or any combination of data: high oil saturation, medium oil saturation, low oil saturation.

3. The method for constructing a mineral content logging intelligent interpretation model according to claim 1, characterized in that: According to the lithology data, porosity data and oil content grade data of the target rock area, the reservoirs in the target rock area are classified to obtain several types of reservoirs, including: Cluster analysis is performed on the lithology data, porosity data, and oil content grade data of the target rock area, and the reservoirs in the target rock area are classified into Class A reservoirs, Class B reservoirs, Class C reservoirs, Class D reservoirs, and Class E reservoirs.

4. The method for constructing a mineral content well logging intelligent interpretation model according to any one of claims 1 to 3, characterized in that: The method of performing multi-wavelet decomposition and reconstruction on the well logging curve data of each type of reservoir in the several types of reservoirs according to the mineral type data and mineral content data of each type of reservoir in the several types of reservoirs, and obtaining the reconstructed curve data of each type of reservoir in the several types of reservoirs, includes: Performing a first preprocessing on the well logging curve data of each type of reservoir in the plurality of types of reservoirs; Using discrete wavelet transform method, multi-wavelet decomposition is performed on the pre-processed logging curve data of each type of reservoir in several types of reservoirs to obtain several wavelet components of each type of reservoir in several types of reservoirs; According to the mineral type data and mineral content data of each type of reservoir in several types of reservoirs, the sensitivity difference between several sub-wave components of each type of reservoir in several types of reservoirs and the mineral content of different types of minerals is analyzed by correlation analysis method, and several sensitive sub-wave components of each type of reservoir in several types of reservoirs are screened out; Well logging curve reconstruction is performed according to a number of sensitive sub-wave components of each type of reservoir in the number of types of reservoirs, so as to obtain the reconstructed curve data of each type of reservoir in the number of types of reservoirs.

5. The method for constructing a mineral content well logging intelligent interpretation model according to claim 4, characterized in that: The first preprocessing of the well logging curve data of each type of reservoir in the plurality of types of reservoirs includes: normalizing well log data for each of several types of reservoirs; and / or, Performing rock electrical correction on well log data for each of several types of reservoirs; Preferably, the method of using discrete wavelet transform to perform multi-wavelet decomposition on the pre-processed logging curve data of each type of reservoir in the several types of reservoirs to obtain several wavelet components of each type of reservoir in the several types of reservoirs includes: Using the discrete wavelet transform method, according to the multi-wavelet decomposition expression, the pre-processed logging curve data of each type of reservoir in several types of reservoirs are decomposed by multi-wavelet, and several wavelet components of each type of reservoir in several types of reservoirs are obtained. Among them, the multi-wavelet decomposition expression is: f(i,j)=f1(i,j)+f2(i,j)+…+f n (i,j) In the multi-wavelet decomposition expression, f(i,j) represents the sum of the wavelet components under scale i and translation j, that is, the discrete wavelet transform result. Scale i is used to control the scale of the wavelet and affect the expansion or compression of the wavelet. A larger i corresponds to a wider wavelet (capturing low-frequency components), while a smaller i corresponds to a narrower wavelet (capturing high-frequency components). Translation j is used to determine the spatial position of the wavelet in the logging curve data, i,j∈Z, f1(i,j), f2(i,j), …, f n (i, j) represents several wavelet components of each type of reservoir, t represents the spatial variable of the logging curve, f(t) represents the input signal, represents the mother wavelet; Preferably, the well logging curve reconstruction is performed according to several sensitive wavelet components of each type of reservoir in several types of reservoirs to obtain the reconstructed curve data of each type of reservoir in several types of reservoirs, including: According to the mineral type data and mineral content data of each type of reservoir in several types of reservoirs, the sensitivity difference between several sub-wave components of each type of reservoir in several types of reservoirs and the mineral content of different types of minerals is analyzed by correlation analysis method, and several sensitive sub-wave components of each type of reservoir in several types of reservoirs are screened out; A number of sensitive sub-wave components of each type of reservoir in a number of types of reservoirs are added together to reconstruct the logging curve, thereby obtaining the reconstructed curve data of each type of reservoir in the number of types of reservoirs.

6. The method for constructing a mineral content well logging intelligent interpretation model according to claim 5, characterized in that: According to the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in the several types of reservoirs, a machine learning algorithm is used to make the model learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve, so as to obtain a mineral content logging intelligent interpretation model, including: performing a second preprocessing on the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in the plurality of types of reservoirs; For each of the several types of reservoirs, the mineral type data, mineral content data and reconstruction curve data after the second preprocessing are used, a plurality of machine learning algorithms are used to enable the model to learn the mapping relationship between the mineral content of different types of minerals in the reservoir and the reconstruction curve, and determine the optimal machine learning algorithm with the best model learning effect for each of the several types of reservoirs; For each of the several types of reservoirs, the mineral type data, mineral content data and reconstruction curve data after the second preprocessing are performed, and an optimal machine learning algorithm with the best model learning effect corresponding to each of the several types of reservoirs is used to obtain a mineral content well logging intelligent interpretation model corresponding to each of the several types of reservoirs; Integrate the mineral content logging intelligent interpretation model corresponding to each type of reservoir in several types of reservoirs to obtain the mineral content logging intelligent interpretation model of the target rock area; Preferably, the method uses a machine learning algorithm to make the model learn the mapping relationship between the mineral content of different types of minerals in each type of reservoir in the target rock area and the reconstruction curve according to the mineral type data, mineral content data and reconstruction curve data of each type of reservoir in the several types of reservoirs, so as to obtain the mineral content logging intelligent interpretation model; after the mineral type data, mineral content data and reconstruction curve data after the second preprocessing for each type of reservoir in the several types of reservoirs are obtained by using the optimal machine learning algorithm with the best model learning effect corresponding to each type of reservoir in the several types of reservoirs, the method further includes any of the following items or any combination thereof: By using cross validation, the credibility of the intelligent well logging interpretation model for mineral content in each type of reservoir is evaluated; Using Bayesian functions, hyperparameter optimization is performed on the intelligent interpretation model of mineral content logging for each type of reservoir corresponding to several types of reservoirs. Using blind well test, the prediction accuracy and generalization ability of the intelligent interpretation model of mineral content logging corresponding to each type of reservoir are tested.

7. A mineral content logging intelligent interpretation method, characterized in that: include: Receive logging curve data of the rock area to be tested; According to the method for constructing the intelligent interpretation model of mineral content logging according to any one of claims 1 to 6, multi-wavelet decomposition and reconstruction are performed on the logging curve data of the rock area to be measured to obtain the reconstructed curve data of the rock area to be measured; According to the reconstructed curve data of the rock area to be tested, the mineral content logging intelligent interpretation model of the target rock area is obtained by the method for constructing the mineral content logging intelligent interpretation model described in any one of claims 1 to 6, and the mineral content logging interpretation result of the rock area to be tested is obtained.

8. A mineral content logging intelligent interpretation system, characterized in that: include: The data receiving module is used to: receive the logging curve data of the rock area to be tested sent by at least one terminal; A preprocessing module, used for: performing multi-wavelet decomposition and reconstruction on the logging curve data of the rock area to be measured according to the method for constructing the intelligent interpretation model of mineral content logging according to any one of claims 1 to 6, to obtain the reconstructed curve data of the rock area to be measured; A prediction module, used to obtain the mineral content logging interpretation result of the rock area to be tested based on the reconstructed curve data of the rock area to be tested and the mineral content logging intelligent interpretation model of the target rock area obtained by the method for constructing the mineral content logging intelligent interpretation model according to any one of claims 1 to 6; The data output module is used to send the logging interpretation results of the mineral content of the rock area to be tested back to at least one terminal.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method for constructing a mineral content well logging intelligent interpretation model as described in any one of claims 1 to 6 and / or the method for intelligent interpretation of mineral content well logging as described in claim 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for constructing a mineral content well logging intelligent interpretation model as described in any one of claims 1 to 6 and / or the method for intelligent interpretation of mineral content well logging as described in claim 7 are implemented.

Images