Quantitative interpretation method and device for strategic metal ore layer logging curve in coal and medium

Through the quantitative interpretation method of the logging curve of the strategic metal ore layer in coal, the random forest model and the Bootstrap resampling method are used to solve the problems of high cost and low efficiency of the traditional exploration method, and high-precision prediction of the strategic metal ore layer in coal is achieved. Provides reliable predictions of layer thickness and content.

CN118938341AInactive Publication Date: 2024-11-12CHINA UNIV OF MINING & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410168206.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-02-06
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The traditional coal-based strategic metal exploration methods are costly and inefficient, and cannot adapt to the exploration needs of coal-based strategic metal ore layers. Logging and drilling cannot directly indicate the abundance of micro-strategic metal elements, resulting in insufficient prediction accuracy and operability.

Method used

The quantitative interpretation method of the strategic metal ore layer logging curve in coal is adopted. By determining the physical properties of the target sample, quality evaluation and histogram correction are carried out after the logging curve is collected, and quantitative interpretation is constructed using the random forest model and the Bootstrap resampling method. Model, predict ore layer thickness and content.

Benefits of technology

The thickness and content prediction of strategic metal ore layer with low cost, high operability and prediction accuracy is achieved, providing reliable prediction results for the evaluation of strategic metal ore in coal.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118938341A_ABST
    Figure CN118938341A_ABST
Patent Text Reader

Abstract

The invention discloses a quantitative interpretation method and device for a strategic metal ore layer logging curve in coal and a medium, and relates to the technical field of mineral exploration. The method comprises the following steps: determining physical property measurement data of a strategic metal target sample in coal; after a logging curve is collected, quality evaluation is carried out on a logging sequence in a research area, standardization processing is carried out on a drilled well through a histogram correction method, and core homing is carried out according to depth information, physical property measurement data and logging curve response matching degree recorded on a sample site; inputting physical property measurement data after core homing into a quantitative interpretation model, and determining a prediction result of the strategic metal content in the coal; the prediction result comprises ore bed thickness and content; the quantitative interpretation model is constructed according to a random forest model and a Bootstrap resampling method. According to the method, relatively high operability and prediction precision can be realized with a low-cost prediction scheme, and reliable prediction results of the thickness, the content and the like of the strategic metal ore layer are provided for evaluation of strategic metal minerals in coal.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of mineral exploration, and in particular to a quantitative interpretation method, equipment and medium for well logging curves of strategic metal ore layers in coal. Background Art

[0002] Strategic metals (such as gallium, lithium, etc.) are important resources widely used in national defense, military industry and high-tech fields. The reserves of traditional strategic metals are decreasing. Therefore, it is of great economic value and strategic significance to expand the strategic metal reserves by finding coal-based strategic metal minerals and proving the reserves of coal-based strategic metals.

[0003] Common geochemical exploration and research methods for the occurrence of strategic metals in coal can be divided into indirect methods and direct methods. The commonly used indirect method is to obtain sample data through sample collection and geochemical analysis, and use relevant statistical methods to study the occurrence form of strategic metals in coal. The direct method is to directly observe the occurrence state of strategic metals in samples using scanning electron microscopes (SEM) and electron probes (EMP). Since geochemical exploration is extremely dependent on field sample collection and laboratory testing, it has a large workload, limited spatial sampling density, high cost and low efficiency, and therefore cannot meet the exploration needs of strategic metal strata in coal-bearing areas.

[0004] Well logging drilling has low cost, high drilling density and rich underground stratigraphic information. In geophysical exploration, well logging drilling has the highest vertical accuracy, which can reach the centimeter level, but it cannot directly indicate the abundance of microscopic strategic metal elements. The relationship between macroscopic well logging response and microscopic strategic metals is relatively complex, and the simple use of correlation statistics has limitations. Summary of the invention

[0005] The purpose of the present invention is to provide a quantitative interpretation method, equipment and medium for logging curves of strategic metal ore layers in coal, which can achieve high operability and prediction accuracy with a low-cost prediction scheme, and provide reliable prediction results such as strategic metal ore layer thickness and content for the evaluation of strategic metal mineral resources in coal.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A quantitative interpretation method of well logging curves of strategic metal ore layers in coal, comprising:

[0008] Determine the physical property measurement data of the target sample; the target sample is a strategic metal in coal;

[0009] After collecting the logging curves, the quality of the logging sequence in the study area was evaluated, and the drilling was standardized using the histogram correction method. The core was located based on the depth information recorded on the sample site, the physical property measurement data and the matching degree of the logging curve response;

[0010] The physical property measurement data after core restoration is input into a quantitative interpretation model to determine the prediction results of the strategic metal content in the coal; the prediction results include the thickness and content of the ore layer; the quantitative interpretation model is constructed based on a random forest model and a Bootstrap resampling method.

[0011] Optionally, determining the physical property measurement result of the target sample specifically includes:

[0012] Core collection and sample preparation were carried out on the wells in the study area to obtain the collected samples;

[0013] Based on the collected samples, strategic metal and mineral components of the samples are determined, and target samples are selected;

[0014] The physical properties of the target sample are measured to obtain physical property measurement results.

[0015] Optionally, the core collection and sample preparation for the wells in the study area specifically includes:

[0016] Determine the sampling location, sample length and sample number of samples obtained from drilling in the study area;

[0017] The samples were processed into cylindrical samples with a diameter of 25 mm, a length of 50 mm, and a particle size less than 200 mesh by wire cutting.

[0018] Optionally, the determination of strategic metal and mineral components of the sample based on the collected sample specifically includes:

[0019] The strategic metal content in the collected samples was determined by inductively coupled plasma atomic emission spectrometry, and the mineral components in the collected samples were determined by X-ray diffraction to obtain the strategic metal content and mineral components of the collected samples.

[0020] Optionally, the measuring of the physical properties of the target sample specifically includes:

[0021] The size of the target sample was measured and the sample volume was calculated. The weight of the target sample was measured using a weighing balance, and the apparent density of the target sample was calculated using weight / volume. The longitudinal and transverse wave velocities of the target sample were measured using an HKN-B intelligent ultrasonic longitudinal and transverse wave automatic tester. The sample resistance was tested using a Shengli VC4091C digital bridge, and the sample resistivity was calculated based on the resistivity formula. The sample gamma value was determined using a JB4000 radiation meter.

[0022] Optionally, the training process of the quantitative interpretation model is:

[0023] Acquisition of training set: taking coal physical property experimental data as input and coal strategic metal data as output to form a sample data set, and using Bootstrap self-help method to divide the sample data set according to the set ratio to obtain training set and test set;

[0024] Random forest parameter optimization: assign a single decision tree node with the initial value of the original random forest, change the number of decision trees, select the number of decision tree nodes by observing the root mean square error, and after determining the number of decision tree nodes, gradually increase the number of decision trees until the root mean square error in the model is stable, and select the corresponding number of decision trees;

[0025] Random forest model evaluation: calculate the model out-of-bag error, importance and Gini coefficient parameters to evaluate the model accuracy;

[0026] Random forest model output: Use the test set as input to verify the prediction accuracy of the random forest model. If the accuracy is lower than the set value, return to step "Obtaining the training set" to adjust the Bootstrap division ratio until the prediction accuracy of the random forest model meets the set accuracy threshold and the model training is completed.

[0027] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the quantitative interpretation method of the logging curve of the strategic metal ore layer in coal mentioned above.

[0028] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the quantitative interpretation method of logging curves of strategic metal ore layers in coal as described above.

[0029] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0030] The present invention discloses a quantitative interpretation method, equipment and medium of a well logging curve of a strategic metal ore layer in coal, the method comprising determining the physical property measurement data of a target sample of strategic metals in coal; after collecting the well logging curve, performing a quality evaluation on the well logging sequence in the study area, and using a histogram correction method to standardize the drilling, and performing core homing according to the depth information, physical property measurement data and well logging curve response matching degree recorded on the spot of the sample; inputting the physical property measurement data after core homing into a quantitative interpretation model to determine the prediction result of the strategic metal content in coal; the prediction result includes the thickness and content of the ore layer; the quantitative interpretation model is constructed according to a random forest model and a Bootstrap resampling method. The present invention can achieve high operability and prediction accuracy with a low-cost prediction scheme, and provide reliable prediction results such as the thickness and content of strategic metal ore layers for the evaluation of strategic metal mineral resources in coal. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0032] Figure 1 Schematic diagram of the process of the non-destructive determination method of Young's modulus of bulk alloy of the present invention;

[0033] Figure 2 This is a flow chart of the method for training the random forest model in this embodiment;

[0034] Figure 3 Schematic diagram of sample preparation, physical property testing, and geochemical testing in this embodiment; wherein, part (a) is a schematic diagram of sample preparation; part (b) is a schematic diagram of geochemical testing; and part (c) is a schematic diagram of physical property testing;

[0035] Figure 4 Schematic diagram of core return in this embodiment; wherein, part (a) is a schematic diagram of core position recorded on site; part (b) is a schematic diagram of core position returned to the core;

[0036] Figure 5 Schematic diagram of random forest parameter selection and input parameter importance ranking in this embodiment; wherein, part (a) is a schematic diagram of random forest parameter selection; part (b) is a schematic diagram of input parameter importance ranking;

[0037] Figure 6 Schematic diagram of random forest model prediction and stratum division of strategic metal strata in coal based on well logging curves.

[0038] Reference numerals:

[0039] 1-core sample; 2-cylindrical sample; 3-coal powder; 4-low-temperature ash; 5-ultrasonic excitation probe; 6-ultrasonic receiving probe; 7-collected longitudinal wave records; 8-collected shear wave records; 9-balance; 10-Victory VC4091C digital bridge; 11-stainless steel urn; 12-JB4000 radiation meter. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0041] The purpose of the present invention is to provide a quantitative interpretation method, equipment and medium for logging curves of strategic metal ore layers in coal, which can achieve high operability and prediction accuracy with a low-cost prediction scheme, and provide reliable prediction results such as strategic metal ore layer thickness and content for the evaluation of strategic metal mineral resources in coal.

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] like Figure 1 As shown, the present invention provides a quantitative interpretation method of well logging curves of strategic metal ore layers in coal, comprising:

[0044] Step 1: Determine the physical property measurement data of the target sample; the target sample is the strategic metal in coal;

[0045] Step 2: After collecting the logging curve, the quality of the logging sequence in the study area is evaluated, and the drilling is standardized using the histogram correction method. The core is located based on the depth information recorded on the sample site, the physical property measurement data and the matching degree of the logging curve response;

[0046] Step 3: Input the physical property measurement data after core restoration into the quantitative interpretation model to determine the predicted results of the strategic metal content in the coal; the predicted results include the thickness and content of the ore layer; the quantitative interpretation model is constructed based on the random forest model and the Bootstrap resampling method.

[0047] Based on the above technical solution, the following specific implementation is provided.

[0048] S10: Core collection and sample preparation: According to the "Core Analysis Method" (GB / T 29172-2012), the sampling position, sample length, and sample number of the samples obtained by drilling are determined, and the samples are sealed and preserved. Since the coal samples are brittle and soft, the samples are processed into cylindrical samples with a diameter of 25mm and a length of 50mm by wire cutting. According to the "Preparation Method of Coal Samples" (GB / T 474-2008), the coal samples are processed into coal powder samples with a particle size of less than 200 mesh.

[0049] Specifically, core samples 1 were obtained through drilling sampling, and the drilling point was located using RTK. The samples were numbered (e.g., HR-6#-1), the sample length (e.g., 15 cm), the sampling depth, and sealed with plastic wrap. Since the coal samples were brittle and soft, a cylindrical sample 2 with a diameter of 25 mm and a length of 50 mm was cut from the core using wire cutting. After the core was cut into cylindrical column samples, the remaining samples were crushed to a coal powder sample 3 with a particle size of less than 200 mesh.

[0050] S20: Determination of strategic metals and mineral components of samples shall refer to the standards of "Determination of Trace Elements in Coal by Inductively Coupled Plasma Atomic Emission Spectrometry" (SN / T 1600-2005), "Determination of Minerals in Coal" (GB / T 7560-2001), and "X-ray Diffractometer" (JB / T 11144-2011) respectively. The strategic metal content in the sample shall be determined by inductively coupled plasma atomic emission spectrometry, and the mineral components in the sample shall be determined by X-ray diffraction method to obtain the strategic metal content and mineral components of the sample.

[0051] Specifically: the ground 3-coal powder sample is added to the middle of the groove of the sample rack by positive pressure method. Then the sample is subjected to X-ray diffraction by an XRD instrument, and its diffraction pattern is analyzed by Jade software to obtain the mineral components of the coal sample. The high-frequency current generates a high-frequency electromagnetic field through the induction coil, which ionizes the working gas (Ar) to form a flame-like discharge high-temperature plasma. The sample interacts with the high-temperature plasma to generate characteristic radiation spectra. The characteristic radiation spectra generated are decomposed into monochromatic spectra representing each element by the grating spectrometer system. The spectral energy is detected by a semiconductor detector, and the content of the strategic metal elements to be tested in the test solution is calculated with reference to the standard solution measured at the same time. The content of strategic metals in coal samples (taking gallium as an example) is shown in Table 1.

[0052] S30: Determination of sample physical properties, measuring the regular sample size to calculate the volume, using a weighing balance to measure the sample weight, and using weight / volume to calculate the apparent density of the sample; using the HKN-B intelligent ultrasonic longitudinal and transverse wave automatic tester to measure the longitudinal and transverse wave velocities; using the Shengli VC4091C digital bridge to test the sample resistance, and calculating the sample resistivity based on the resistivity formula ρ=RS / L; using the JB4000 radiation measuring instrument to determine the sample gamma value.

[0053] Specifically: Figure 3 As shown, the physical properties of the cylindrical sample 2 are measured, the ultrasonic excitation probe 5 and the ultrasonic receiving probe 6 are placed at the two ends of the cylindrical sample 2, and the ultrasonic longitudinal and transverse wave automatic tester is used to obtain the collected longitudinal wave record 7 and the collected transverse wave record 8. The cylindrical sample 2 is placed on a balance 9 to record its weight, and a vernier caliper is used to record the diameter and length of the cylindrical sample 2, and the apparent density of the coal sample is calculated by weight / volume. The Shengli VC4091C digital bridge 10 test electrodes are placed at both ends of the cylindrical sample 2, the resistance is recorded, and the resistivity of the sample is calculated according to the resistivity formula ρ=RS / L. The cylindrical sample 2 is placed in a stainless steel urn 11, and the JB4000 radiation measuring instrument 12 is used to measure the gamma value of the sample. The physical properties of the sample are shown in Table 1.

[0054] Table 1 Physical properties of coal samples and strategic metal content in coal (partial)

[0055]

[0056]

[0057] S40: Well logging curve standardization and core placement, evaluate the quality of the well logging sequence in the study area, select wells with higher logging quality as standard wells; standardize the remaining wells using the histogram correction method. Core placement is performed based on the depth information of the sample field records, the matching degree between the sample physical property measurement results and the well logging curve response.

[0058] Specifically, after collecting the logging curves, the quality of the logging sequence in the study area is evaluated, and the wells with higher logging quality are selected as standard wells based on the amplitude range, signal-to-noise ratio and other characteristics of the logging curves; the amplitude range and median of the same type of curves of other wells are corrected to the same range as the standard well using the histogram correction method. Figure 4 As shown in the figure, the sampling depth of the coal sample recorded on site, the density and velocity values ​​of the experimental test are projected onto the logging curve to determine whether the laboratory test results match the logging test results. If not, the coal sample sampling position is moved up in sequence according to the location of the gangue until the laboratory test results match the logging measurement results.

[0059] S50: Establishment of a random forest model for strategic metal ore layers in coal based on experimental data, including: taking the experimental data set as the original sample, and using the Bootstrap resampling method to extract multiple samples from the original sample; making a decision tree model for each Bootstrap sample; combining multiple decision trees for prediction, and obtaining the final prediction result through voting; calculating the accuracy of the random forest model based on the content of multiple predicted strategic metals and the experimental strategic content; and optimizing the established random forest model based on the model prediction accuracy to obtain a tuned random forest model.

[0060] like Figure 2 As shown, in step S50, obtaining the random forest model of the strategic metal ore layer in coal of the random forest includes the following steps:

[0061] S501: Acquisition of training data; Taking the experimental data of coal properties as input and the strategic metal data of coal as output to form a sample data set, the Bootstrap method is used to divide the training set and the test set in proportion. For example: the physical property parameters of coal samples tested in the laboratory and the strategic metal content of coal samples tested in geochemical tests are summarized, as shown in Table 1. Taking the experimental data of coal properties as input and the strategic metal data of coal as output to form a sample data set. After randomly shuffling the sample data set, the Bootstrap method is used to divide the training set and the test set during the random forest simulation in a ratio of 7:3. The training set is used to train the random forest model, and the test set is used to test the prediction accuracy of the random forest model.

[0062] S502: Random forest parameter optimization; assign the original random forest initial value to a single decision tree node, change the number of decision trees, and determine the appropriate number of decision tree nodes by observing the root mean square error. After determining the number of decision tree nodes, gradually increase the number of decision trees until the root mean square error in the model is stable, and select the corresponding number of decision trees. For example: set the number of cycles to 500, assign the original random forest initial value to a single decision tree node (2, 3, 4, 10, 20), change the number of decision trees, and observe the root mean square error to determine the appropriate number of decision tree nodes, such as Figure 5 As shown in (a), the number of nodes in a single decision tree corresponding to the minimum root mean square error is 4; after determining that the number of decision tree nodes is 4, the number of decision trees is gradually increased until the root mean square error in the model is stable, and the corresponding number of decision trees is selected as 250.

[0063] S503: Random forest model evaluation; calculate the model out-of-bag error, importance, Gini coefficient and other parameters to evaluate the model accuracy. For example: after determining that the number of decision tree nodes of the random forest is 4 and the number of decision trees is 250, first use the physical property parameters of the coal samples in the training set as input and the strategic metal content of the coal samples in the training set as output to train the model. Increase the number of input parameters in turn, calculate the prediction accuracy of the random forest model, and obtain the importance ranking of the input parameters, such as Figure 5 As shown in (b).

[0064] S504: Random forest model output; use the test set as input to verify the prediction accuracy of the random forest model. If the accuracy is low, return to S501 to adjust the Bootstrap division ratio and repeat steps S501-S504. For example: if the prediction accuracy is high, output the model. For example: first use the physical properties of the coal samples in the test set as input and the strategic metal content of the coal samples in the test set as output to test the model. Calculate the error between the strategic metal content of the coal samples in the model test set and the strategic metal content of the coal samples predicted by the random forest model, as shown in Table 2. If the accuracy is low, return to S501 to adjust the Bootstrap division ratio and repeat steps S501-S504.

[0065] Table 2 Error statistics between true value and predicted value of coal sample test set

[0066]

[0067] S60: Prediction of strategic metal ore layers in coal based on random forest model of well logging curves. Taking well logging curves as input, the optimized random forest model is used to predict the strategic metal content of coal seam sections; cluster analysis is performed on the prediction results, and the ore layers are divided according to coal thickness and strategic metal content in coal.

[0068] Specifically, the well logging curve of the coal seam section is selected, and the well logging curve is used as input. The optimized random forest model is used to predict the strategic metal content of the coal seam section; the prediction results are clustered and analyzed, and the prediction results are divided into two categories, one is gallium-poor coal and the other is gallium-rich coal. Taking gallium among strategic metals as an example, the boundary grade of gallium in coal is 30ug / g, the lowest industrial grade is 50ug / g, and the minimum mineable thickness is 0.7m. Gallium-rich ore layers are divided according to coal thickness and strategic metal content in coal, such as Figure 6 shown.

[0069] In this embodiment, random forest is a classifier based on statistical theory, which uses multiple trees to train and predict samples. It has high prediction accuracy, good tolerance for outliers and noise, and is not prone to overfitting. At present, the collection of training data sets for the random forest strategic metal prediction model can be completed by sampling the drilling cores and preparing the samples for geochemistry (main and trace elements, mineral components, etc.) and geophysics (density, velocity, resistivity, etc.) testing of in-situ samples. Standardize the logging curves and return the cores to improve the consistency between logging data and between logging data and experimental data. The Bootstrap resampling method is used to divide the laboratory test results into training sets and test sets, and train the random forest prediction model for predicting coal-based strategic metals with physical property parameters. Using the logging curve as input, the random forest model predicts the content of strategic metals in coal; then, the coal seams are stratified by cluster analysis, and the ore layers are divided according to the thickness of the coal and the content of strategic metals in the coal.

[0070] In addition, the present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the quantitative interpretation method of the logging curve of the strategic metal ore layer in coal mentioned above.

[0071] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the quantitative interpretation method of logging curves of strategic metal ore layers in coal as described above.

[0072] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0073] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A quantitative interpretation method of well logging curves of strategic metal ore layers in coal, characterized in that: include: Determine the physical property measurement data of the target sample; The target samples are strategic metals in coal; After collecting the logging curves, the quality of the logging sequence in the study area was evaluated, and the drilling was standardized using the histogram correction method. The core was located based on the depth information recorded on the sample site, the physical property measurement data and the matching degree of the logging curve response; The physical property measurement data after core restoration is input into a quantitative interpretation model to determine the prediction results of the strategic metal content in the coal; the prediction results include the thickness and content of the ore layer; the quantitative interpretation model is constructed based on a random forest model and a Bootstrap resampling method.

2. The quantitative interpretation method of well logging curves of strategic metal ore layers in coal according to claim 1 is characterized in that: Determining the physical property measurement results of the target sample specifically includes: Core collection and sample preparation were carried out on the wells in the study area to obtain the collected samples; Based on the collected samples, strategic metal and mineral components of the samples are determined, and target samples are selected; The physical properties of the target sample are measured to obtain physical property measurement results.

3. The quantitative interpretation method of well logging curves of strategic metal ore layers in coal according to claim 2 is characterized in that: The core collection and sample preparation for the wells in the study area specifically include: Determine the sampling location, sample length and sample number of samples obtained from drilling in the study area; The samples were processed into cylindrical samples with a diameter of 25 mm, a length of 50 mm, and a particle size less than 200 mesh by wire cutting.

4. The quantitative interpretation method of well logging curves of strategic metal ore layers in coal according to claim 2 is characterized in that: The determination of strategic metal and mineral components of the samples based on the collected samples specifically includes: The strategic metal content in the collected samples was determined by inductively coupled plasma atomic emission spectrometry, and the mineral components in the collected samples were determined by X-ray diffraction to obtain the strategic metal content and mineral components of the collected samples.

5. The quantitative interpretation method of well logging curves of strategic metal ore layers in coal according to claim 2 is characterized in that: The measuring of the physical properties of the target sample specifically includes: The size of the target sample was measured and the sample volume was calculated. The weight of the target sample was measured using a weighing balance, and the apparent density of the target sample was calculated using weight / volume. The longitudinal and transverse wave velocities of the target sample were measured using an HKN-B intelligent ultrasonic longitudinal and transverse wave automatic tester. The sample resistance was tested using a Shengli VC4091C digital bridge, and the sample resistivity was calculated based on the resistivity formula. The sample gamma value was determined using a JB4000 radiation meter.

6. The quantitative interpretation method of well logging curves of strategic metal ore layers in coal according to claim 1, characterized in that: The training process of the quantitative interpretation model is: Acquisition of training set: taking coal physical property experimental data as input and coal strategic metal data as output to form a sample data set, and using Bootstrap self-help method to divide the sample data set according to the set ratio to obtain training set and test set; Random forest parameter optimization: assign a single decision tree node with the initial value of the original random forest, change the number of decision trees, select the number of decision tree nodes by observing the root mean square error, and after determining the number of decision tree nodes, gradually increase the number of decision trees until the root mean square error in the model is stable, and select the corresponding number of decision trees; Random forest model evaluation: calculate the model out-of-bag error, importance and Gini coefficient parameters to evaluate the model accuracy; Random forest model output: Use the test set as input to verify the prediction accuracy of the random forest model. If the accuracy is lower than the set value, return to step "Getting the training set" to adjust the Bootstrap division ratio until the prediction accuracy of the random forest model meets the set accuracy threshold, completing the model training.

7. An electronic device, characterized in that: It comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the quantitative interpretation method of logging curves of strategic metal ore layers in coal according to any one of claims 1-6.

8. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed by a processor, implements a quantitative interpretation method for logging curves of strategic metal ore layers in coal as described in any one of claims 1-6.

Citation Information

Patent Citations

  • Rock core homing method and device and computing system

    CN116068646A

  • Clastic rock reservoir porosity evaluation method and application

    CN116792091A