Method, system and equipment for identifying stratum brine potassium salt enrichment layer and medium

Abstract

The invention belongs to the field of geophysical prospecting, and particularly relates to an identification method, system and equipment for a stratum brine potassium salt enrichment layer and a medium. The method comprises the following steps: comparing stratum brine layer identification results obtained by different methods, and determining a stratum with overlapped brine layer identification results as a final stratum brine layer; identifying a potassium-rich brine layer by using a natural gamma-ray spectrum logging potassium element content curve; calculating the shale content of the framework of the potassium-rich brine layer through the thorium curve so as to obtain the content of the potassium element in the potassium-rich brine; and determining the brine potassium salt enrichment layer according to the potassium element content in the potassium-rich brine. According to the method, the brine reservoir identification accuracy is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of geophysical prospecting, and particularly relates to a method, system, device and medium for identifying a stratum brine potash enrichment layer. BACKGROUND

[0002] In the field of mineral exploration, especially in the identification of stratum brine potash enrichment layer, accurate identification of potassium-rich brine reservoir is crucial. However, there are some problems in determining potassium-rich brine reservoir in the prior art, resulting in low reservoir identification accuracy.

[0003] Traditional brine reservoir identification methods may only rely on a single well logging curve or simple analysis methods, which cannot fully consider the complexity and diversity of the stratum. These methods may be disturbed by various factors, such as noise of logging data, heterogeneity of stratum, etc., thereby affecting the accuracy of identification.

[0004] In addition, in the determination of potassium element content, potassium and sodium are not measured separately in conventional water analysis, which brings difficulties to accurately determine the potassium content, thereby affecting the judgment of potassium-rich brine reservoir.

[0005] Since the identification accuracy of potassium-rich brine reservoir is directly related to the development and utilization of potassium resources, improving the reservoir identification accuracy is of great significance to meet the domestic demand for potassium resources and reduce dependence on imports. Therefore, it is an urgent problem to find a more accurate and reliable method for identifying stratum brine potash enrichment layer. SUMMARY

[0006] In view of the deficiencies of the prior art, the present application provides a method for identifying a brine potash enrichment layer, which solves the technical problem of low identification accuracy and reliability of potassium-rich brine reservoir.

[0007] The present application provides a method for identifying a stratum brine potash enrichment layer, comprising the following steps:

[0008] Comparing the stratum brine layer identification results obtained by different methods, the stratum with overlapping brine layer identification results is determined as the final stratum brine layer;

[0009] Identifying the potassium-rich brine layer by using the natural gamma-ray spectroscopy logging potassium element content curve;

[0010] Calculating the argillaceous content of the potassium-rich brine layer skeleton through the thorium curve, and then obtaining the potassium element content in the potassium-rich brine;

[0011] Determining the brine potash enrichment layer according to the potassium element content in the potassium-rich brine.

[0012] Preferably, the different methods comprise:

[0013] Preliminarily identifying the stratum brine layer through the well logging curve;

[0014] The stratum brine layer is identified by using the overlap method.

[0015] The well logging curve data is normalized, a deep learning model is established and trained to identify the stratum brine layer.

[0016] Preferably, the well logging curve data is normalized, a deep learning model is established and trained to identify the stratum brine layer, including the following steps:

[0017] The well logging curve data for identifying the stratum brine layer is normalized;

[0018] A deep learning model is established;

[0019] The stratum brine layer identified by the well logging curve is used as a data set composed of artificial annotation conclusions to train the deep learning model, and the training result is evaluated by accuracy rate;

[0020] The well logging curve data is input into the trained model to determine whether the stratum is a brine layer.

[0021] Preferably, the deep learning model is a long short-term memory network model.

[0022] Preferably, the loss function of the long short-term memory network model is:

[0023]

[0024] Wherein, y i is the actual value; y′ i is the predicted value;

[0025] The formula of the accuracy rate is:

[0026]

[0027] Wherein, n correct is the number of samples classified accurately by the deep learning model; n total is the total number of samples.

[0028] Preferably, the shale content of the potassium-rich brine layer skeleton is calculated through the thorium curve, and then the potassium element content in the potassium-rich brine is obtained, including the following steps:

[0029] The shale content of the potassium-rich brine layer skeleton is calculated through the thorium curve in the natural gamma ray spectrum curve;

[0030] The potassium element content of the potassium-rich brine layer skeleton is calculated according to the shale content;

[0031] The potassium element content in the potassium-rich brine is calculated according to the potassium element content in the natural gamma ray spectrum logging and the skeleton potassium element content.

[0032] Preferably, the formula for calculating the mud content of the potassium-rich brine layer skeleton is:

[0033]

[0034]

[0035]

[0036] In the formula, V Th V is the thorium element coefficient; sh The mud content index; Th max The thorium curve value in mudstone; Th min Th represents the curve value of thorium in pure dolomite; Th represents the actual value of thorium in the natural gamma spectrum curve; c is a constant; V k The mud content of the potassium-rich brine layer framework;

[0037] The formula for calculating the potassium content of the potassium-rich brine layer framework is as follows:

[0038] K 骨架 =V k ×(K max -K min )+K min

[0039] K 骨架 Potassium content of the potassium-rich brine reservoir framework; K max This refers to the potassium content of the mudstone; K min The potassium content of pure dolomite;

[0040] The potassium content in potassium-rich brine is:

[0041] K 卤水 =K 总 -K 骨架

[0042] In the formula, K 卤水 This refers to the potassium content in potassium-rich brine; K 总 This represents the total potassium content in the formation; obtained through potassium curves from natural gamma-ray spectroscopy logging.

[0043] This invention provides a system for identifying potassium salt enrichment layers in formation brine, characterized in that the system comprises:

[0044] The well logging curve recognition module is used to identify formation brine layers through well logging curves;

[0045] Overlapping method identification module, used to identify formation brine layers using the overlapping method;

[0046] Deep learning model recognition module, used to identify formation brine layers using deep learning models;

[0047] The comparison module is used to compare the formation brine layer identification results of the well logging curve identification module, the overlap method identification module, and the deep learning model identification module, so as to obtain the final formation brine layer identification result.

[0048] A potassium-rich brine layer identification module is used to identify potassium-rich brine layers in the formation brine layer identification results.

[0049] A module for determining the potassium content in potassium-rich brine is used to obtain the potassium content in the potassium-rich brine.

[0050] A potassium salt enrichment layer identification module for brine is used to determine the potassium salt enrichment layer of brine based on the potassium content in the potassium-rich brine.

[0051] The present invention provides an apparatus including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the method for identifying potassium salt enrichment layers in formation brine.

[0052] The present invention provides a computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the method for identifying potassium salt enrichment layers in formation brine.

[0053] Compared with the prior art, the present invention has the following advantages:

[0054] 1. It greatly improves the accuracy of brine reservoir identification, and thus improves the accuracy of brine potassium salt enrichment layer identification.

[0055] 2. It can accurately obtain the potassium content in potassium-rich brines, providing accurate data support for potassium mining.

[0056] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1 A flowchart of a method for identifying potassium salt enrichment layers in formation brine provided by an embodiment of the present invention is shown;

[0059] Figure 2 The flowchart of step S1 of a method for identifying potassium salt enrichment layers in formation brine provided by an embodiment of the present invention is shown;

[0060] Figure 3 The flowchart of step S13 of a method for identifying potassium salt enrichment layers in formation brine provided by an embodiment of the present invention is shown.

[0061] Figure 4 The flowchart of step S3 of a method for identifying potassium salt enrichment layers in formation brine provided by an embodiment of the present invention is shown. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Example 1

[0064] See Figure 1 As shown in the figure, this embodiment provides a method for identifying potassium salt enrichment layers in formation brine, including:

[0065] S1: Compare the brine layer identification results obtained by different methods, and determine the strata with overlapping brine layer identification results as the final brine layer.

[0066] In formation brine layers, brine is not uniformly distributed but concentrated in specific areas, which are called brine reservoirs. The brine reservoir identification results obtained through different methods (S11, S12, S13) are compared to determine the final brine reservoir identification result. Specifically, if there is overlap in the brine reservoir results obtained by different methods, that reservoir is identified as the final brine reservoir.

[0067] See Figure 2 As shown, S1 is specifically:

[0068] S11: Preliminary determination of formation brine layer based on well logging curves.

[0069] The brine reservoir is rich in various halogen elements, with associated potassium isotopes. 40With higher K content, the radiation intensity of brine reservoirs is higher compared to other reservoirs, with natural gamma values ​​generally ranging from 15 to 48 API. In contrast, the clay content in anhydrite non-reservoir is extremely low, and the natural gamma value is generally below 15 KPI. Natural gamma logging curves can generally be used to accurately distinguish between brine reservoirs and mudstone layers. In addition, brine reservoirs generally have high porosity, and their sonic logging curves, density logging curves, and neutron logging curves are significantly different from those of other formations.

[0070] Therefore, in this embodiment, underground brine reservoirs are identified using natural gamma logging curves, sonic logging curves, density logging curves, and neutron logging curves.

[0071] Furthermore, if the natural gamma ray curve value is within the range of 15-48 API, and the sonic logging curve shows a high value, the density logging curve shows a low value, and the neutron logging curve shows a high value, then a preliminary judgment can be made that the formation is a brine reservoir.

[0072] It should be noted that the high and low values ​​of sonic logging, density logging, and neutron logging curves are usually compared with empirical value ranges established through extensive research and practice in the same area. Within a specific geological region, a numerical range is determined by statistically analyzing sonic logging data from multiple known common aquifers or similar formations. If the measured value is significantly lower than the average value of this range, it can be considered a low value; conversely, it is considered a high value.

[0073] S12: The overlapping method is used to identify the formation brine layer.

[0074] The overlap method is a relatively accurate method for determining the type of fluid in a formation. Therefore, this embodiment uses the overlap method to further identify brine reservoirs.

[0075] Specifically, the deep resistivity logging curve and the compensated neutron logging curve are overlapped according to a preset lateral ratio, and the amplitude difference of the three curves is judged. If the amplitude difference is greater than a preset threshold, it is a further identified underground brine reservoir.

[0076] Among them, the preset horizontal scale of the deep resistivity logging curve is 1-100000; the preset horizontal scale of the compensated neutron logging curve is 30-65; and the preset threshold of the amplitude difference is twice the benchmark amplitude difference.

[0077] It should be noted that the baseline amplitude difference here is an amplitude standard determined based on previous data.

[0078] S13: Normalize the logging curves, establish and train a deep learning model to identify formation brine layers.

[0079] Well logging curves reflect the logging response characteristics of formations at different depths. In practical applications, directly establishing a brine reservoir identification model between the well logging curves and the brine reservoir using deep learning methods would significantly increase the model's complexity and introduce a large amount of noise, resulting in low identification accuracy. Therefore, before identifying brine reservoirs, it is necessary to select several reliable and representative curves from numerous well logging curves using sensitivity analysis.

[0080] Sensitivity analysis involves a thorough study of the characteristics of various logging curves, analyzing their correlation with brine reservoirs, stability, and response under different geological conditions. Data analysis, statistical methods, or relevant geophysical knowledge are used to assess the importance and sensitivity of each logging curve for brine reservoir identification, thereby selecting reliable and representative curves.

[0081] In this embodiment, the logging curves selected for identifying underground brine reservoirs using deep learning models are natural gamma logging curves, sonic logging curves, compensated neutron logging curves, density logging curves, and deep resistivity logging curves.

[0082] See Figure 3 As shown, S13 specifically refers to:

[0083] S131: Normalize the logging curve data for identifying formation brine layers.

[0084] Well logging data often uses different units, and these differences can affect the training and recognition results of deep learning models. To avoid this impact, the well logging curve data input into the deep learning model for identifying underground brine reservoirs needs to be normalized preprocessed. The purpose of normalization preprocessing is to transform the data to a uniform scale or range, making data from different units comparable, thereby improving the stability and accuracy of the model. This ensures that the model can better learn the features and patterns in the data without being affected by the units of data.

[0085] In this step, the maximum-minimum method is used for normalization. The normalization formula is:

[0086]

[0087] Wherein, x represents the normalized logging curve data; i x represents the data for the i-th logging curve; min The minimum value in the well logging curve data; the x max This refers to the maximum value in the well logging curve data.

[0088] S132: Build a deep learning model.

[0089] The deep learning model is a Long Short-Term Memory (LSTM) network model. The input to this model is normalized well logging data used to identify underground brine reservoirs. The output of the model is the formation identification conclusion; specifically, there are only two conclusions: brine reservoir and non-brine reservoir.

[0090] Specifically, this model only determines whether a formation is a brine reservoir, without classifying the specific lithology of non-brine reservoirs. This improves the model's focus on brine reservoirs, allowing the deep learning model to learn the characteristics of brine reservoirs more effectively.

[0091] The computation process of deep learning models involves matrix operations, and matrix operations, regardless of how many times they are superimposed, remain linear. Reservoir identification essentially does not involve purely linear problems. To enable the network model to handle nonlinear issues, activation functions with nonlinear factors need to be added to the deep learning model, thereby improving the neural network's nonlinear mapping capability, i.e., its approximation capability. In this embodiment, the sigmoid function is used as the activation function for the Long Short-Term Memory (LSTM) network model.

[0092] Furthermore, to measure the degree of difference between the model's predicted values ​​and the true values, this embodiment introduces a loss function. The loss function is a non-negative function in deep learning models that measures the degree of difference between predicted and true values. It is the basis for parameter updates in deep learning models; the smaller the loss function, the more accurate the model's predictions. Training the network model essentially involves continuously calculating the loss function and updating the network weights and biases to minimize the loss.

[0093] In this embodiment, the loss function of the Long Short-Term Memory network model is loss(y′). i ,y i )for:

[0094]

[0095] Among them, y i This is the actual value; y′ i These are predicted values.

[0096] S133: Using the brine layer identified by well logging curves as a dataset composed of manually labeled conclusions, a deep learning model is trained, and the training results are evaluated by accuracy.

[0097] A dataset consisting of well logging curves with manually labeled conclusions is used to train the deep learning model. During training, the model learns the relationship between the features of these well logging curves and the manually labeled conclusions. After training, the training results of the deep learning model are evaluated using accuracy. Accuracy refers to the proportion of samples correctly classified by the model out of the total number of samples. By calculating the accuracy, we can understand how accurately the model predicts underground brine reservoirs, thereby judging the model's performance and effectiveness.

[0098] It should be noted that the logging curves included in the manually labeled dataset are: natural gamma logging curves, sonic logging curves, density logging curves, and neutron logging curves.

[0099] In this embodiment, the formula for the accuracy is:

[0100]

[0101] Where, n correct n is the number of samples that are accurately classified using a deep learning model. total This represents the total number of samples.

[0102] S134: Input the well logging curve data into the trained model to determine whether the formation is a brine layer.

[0103] The selected logging data (natural gamma ray logging, sonic logging, compensated neutron logging, density logging, and deep resistivity logging) used for identifying underground brine reservoirs using a deep learning model are input into the pre-trained deep learning model. This deep learning model analyzes and processes the input logging data, and through its internal calculations and logical judgments, ultimately determines whether the corresponding formation is an underground brine reservoir.

[0104] The following rules apply to determining the final depth range of the brine reservoir: If the brine reservoir identification result obtained in S13 overlaps with other brine reservoir identification results, then the final depth range of the brine reservoir shall be based on the depth range of the brine reservoir obtained in S13; if the brine reservoir identification results obtained in S11 and S12 overlap, then the final depth range of the brine reservoir shall be based on the depth range of the brine reservoir obtained in S12.

[0105] In this step, three methods—well logging curves, overlay method, and deep learning method—are used to identify underground brine reservoirs. The identification results obtained by these three methods are then compared and analyzed. In determining the final brine reservoir identification result, the results obtained by the deep learning method are given priority, followed by the results of the overlay method. By comprehensively using these two methods, combined with the well logging curve method, the accuracy of brine reservoir identification can be greatly improved.

[0106] S2: Identify potassium-rich brine layers using natural gamma spectroscopy logging of potassium content curves.

[0107] After identifying the reservoir as a brine layer, further determination is needed to determine if it is potassium-rich brine. The simplest method is usually to directly analyze the chemical composition and ion content of brine samples to obtain the potassium ion content. However, conventional water analysis does not measure potassium and sodium separately, making it difficult to determine the potassium content. Therefore, in this step, natural gamma ray spectroscopy logging is used to identify potassium-rich brine reservoirs within the final brine formation. Natural gamma ray spectroscopy logging is a logging method that records natural gamma rays in different energy ranges. The gamma rays released from the formation are mainly produced by the decay of three radioactive isotopes: potassium, thorium, and uranium. This method can provide the content of potassium, thorium, and uranium in the formation. The detector for natural gamma ray spectroscopy logging is basically the same as that for natural gamma ray logging, except that a multi-channel pulse amplitude analyzer is added. This analyzer can measure the number of pulses at different amplitudes to obtain gamma ray spectra at different energies, which can then be used to determine different radionuclides.

[0108] S3: The clay content of the potassium-rich brine layer skeleton is calculated by using the thorium curve, and then the potassium content in the potassium-rich brine is obtained.

[0109] In the current step, the potassium content measured by well logging consists of two parts. One part is the potassium content contained in the reservoir framework (i.e., the solid structures such as rocks that make up the reservoir); the other part is the potassium content in the brine of the reservoir. However, in potash mining, the primary focus is on the potassium content in the brine, as this potassium can be extracted using specific processes for potash production. Therefore, it is necessary to calculate the potassium content in the brine separately to determine its specific level and provide accurate data support for potash mining.

[0110] refer to Figure 4 As shown, S3 specifically refers to:

[0111] S31: Calculate the clay content of the potassium-rich brine layer skeleton using the thorium curve in the natural gamma spectrum curve.

[0112] Typically, in brine reservoirs, the framework portion of the reservoir has the greatest impact on thorium logging curves. However, the thorium content in potassium-rich brines is very low. Considering this characteristic, this embodiment uses thorium-related data to inversely deduce the clay content in the framework of potassium-rich brine reservoirs.

[0113] Since thorium mainly comes from the reservoir framework, and is extremely rare in potassium-rich brines, analyzing the thorium content in logging curves can help infer the content of clay in the reservoir framework, thus providing important information for further research on the characteristics of potassium-rich brine reservoirs.

[0114]

[0115]

[0116]

[0117] In the formula, V Th V is the thorium element coefficient; sh The mud content index; Th max The thorium curve value in mudstone; Th min Th represents the curve value of thorium in pure dolomite; Th represents the actual value of thorium in the natural gamma spectrum curve; c is a constant; V k The mud content of the potassium-rich brine layer framework.

[0118] S32: The potassium content of the potassium-rich brine layer skeleton is calculated based on the mud content.

[0119] The formula for calculating the potassium content of the potassium-rich brine layer framework is as follows:

[0120] K 骨架 =V k ×(K max -K min )+K min

[0121] K 骨架 Potassium content of the potassium-rich brine reservoir framework; K max This refers to the potassium content of the mudstone; K min This represents the potassium content of pure dolomite.

[0122] S33: Calculate the potassium content in potassium-rich brine based on the potassium content in natural gamma spectroscopy logging and the potassium content in the skeletal framework.

[0123] The potassium content in the potassium-rich brine is as follows:

[0124] K 卤水 =K 总 -K 骨架

[0125] In the formula, K 卤水 This refers to the potassium content in potassium-rich brine; K 总 The total potassium content in the formation is obtained from the potassium curve in natural gamma ray spectroscopy logging.

[0126] S4: Determine the potassium salt enrichment layer of the brine based on the potassium content in the potassium-rich brine.

[0127] Among them, when the potassium content in potassium-rich brine exceeds 100 g / L, it is identified as a potassium salt enrichment layer in the brine.

[0128] Example 2

[0129] This embodiment provides a system for identifying potassium-rich layers in formation brine. The system employs a method for identifying potassium-rich layers in formation brine, and includes:

[0130] The well logging curve recognition module is used to identify formation brine layers through well logging curves.

[0131] The overlapping method identification module is used to identify formation brine layers using the overlapping method.

[0132] The deep learning model recognition module is used to identify formation brine layers using deep learning models.

[0133] The comparison module is used to compare the formation brine layer identification results of the well logging curve identification module, the overlap method identification module, and the deep learning model identification module, so as to obtain the final formation brine layer identification result.

[0134] A potassium-rich brine layer identification module is used to identify potassium-rich brine layers in the formation brine layer identification results.

[0135] A module for determining the potassium content in potassium-rich brine is used to obtain the potassium content in the potassium-rich brine.

[0136] A potassium salt enrichment layer identification module for brine is used to determine the potassium salt enrichment layer of brine based on the potassium content in the potassium-rich brine.

[0137] Example 3

[0138] This embodiment provides an electronic device, including a memory and a processor. The memory is used to store computer programs, and the processor runs the computer programs to enable the electronic device to perform the identification method of formation brine potassium salt enrichment layer of Embodiment 1.

[0139] This embodiment includes a computer-readable storage medium storing a data processing program, which is executed by a processor to identify a potassium salt enrichment layer in a formation brine.

[0140] Those skilled in the art will understand that the embodiments described herein can be provided as methods, apparatus (devices), or computer program products. Therefore, this document may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. This includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0141] This document is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (devices), and computer program products according to embodiments herein. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0142] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0143] The embodiments and / or implementation methods described above are merely preferred embodiments and / or implementation methods for implementing the technology of the present invention, and are not intended to limit the implementation methods of the technology of the present invention in any way. Any person skilled in the art can make some modifications or alterations to other equivalent embodiments without departing from the scope of the technical means disclosed in the content of the present invention, but they should still be regarded as the technology or embodiments that are substantially the same as the present invention.

Claims

1. A method for identifying potassium salt enrichment layers in formation brine, characterized in that, Includes the following steps: By comparing the brine layer identification results obtained by different methods, the strata with overlapping brine layer identification results are identified as the final brine layers. Potassium-rich brine layers were identified using natural gamma ray spectroscopy logging of potassium content curves. The clay content of the potassium-rich brine layer skeleton is calculated by thorium curve, and then the potassium content in the potassium-rich brine is obtained. The potassium salt enrichment layer of the brine is determined based on the potassium content in the potassium-rich brine.

2. The method for identifying potassium salt enrichment layers in formation brine according to claim 1, characterized in that, The different methods include: Preliminary identification of formation brine layers using well logging curves; The overlapping method was used to identify the formation brine layer; The logging curves were normalized, and a deep learning model was established and trained to identify formation brine layers.

3. The method for identifying potassium salt enrichment layers in formation brine according to claim 2, characterized in that, The process of normalizing well logging curves, establishing and training a deep learning model to identify formation brine layers includes the following steps: Normalize the logging curve data used to identify formation brine layers; Build deep learning models; A deep learning model was trained using a dataset composed of manually labeled conclusions from formation brine layers identified by well logging curves, and the training results were evaluated by accuracy. The well logging curve data is input into the trained model to determine whether the formation is a brine layer.

4. The method for identifying potassium salt enrichment layers in formation brine according to claim 3, characterized in that, The deep learning model is a long short-term memory network model.

5. The method for identifying potassium salt enrichment layers in formation brine according to claim 4, characterized in that, The loss function of the Long Short-Term Memory network model is: Among them, y i This is the actual value; y′ i This is a predicted value; The formula for the accuracy rate is: Where, n correct n is the number of samples that are accurately classified using a deep learning model. total This represents the total number of samples.

6. The method for identifying potassium salt enrichment layers in formation brine according to claim 1, characterized in that, The method of calculating the mud content of the potassium-rich brine layer skeleton through thorium curves, and then obtaining the potassium content in the potassium-rich brine, includes the following steps: The clay content of the potassium-rich brine layer skeleton was calculated using the thorium curve in the natural gamma spectrum curve. The potassium content of the potassium-rich brine layer skeleton was calculated based on the mud content. The potassium content in potassium-rich brine was calculated based on the potassium content in natural gamma spectroscopy logging and the potassium content in the skeletal framework.

7. The method for identifying potassium salt enrichment layers in formation brine according to claim 6, characterized in that, The formula for calculating the mud content of the potassium-rich brine layer skeleton is as follows: In the formula, V Th V is the thorium element coefficient; sh The mud content index; Th max The thorium curve value in mudstone; Th min Th represents the curve value of thorium in pure dolomite; Th represents the actual value of thorium in the natural gamma spectrum curve; c is a constant; V k The mud content of the potassium-rich brine layer framework; The formula for calculating the potassium content of the potassium-rich brine layer framework is as follows: K 骨架 =V k ×(K max -K min )+K min K 骨架 The potassium content of the potassium-rich brine layer framework; K max This refers to the potassium content of the mudstone; K min The potassium content of pure dolomite; The potassium content in potassium-rich brine is: K 卤水 =K 总 -K 骨架 In the formula, K 卤水 This refers to the potassium content in potassium-rich brine; K 总 The total potassium content in the formation is obtained from the potassium curve in natural gamma ray spectroscopy logging.

8. A system for identifying potassium salt enrichment layers in formation brine, characterized in that, The system includes: The well logging curve recognition module is used to identify formation brine layers through well logging curves; Overlapping method identification module, used to identify formation brine layers using the overlapping method; Deep learning model recognition module, used to identify formation brine layers using deep learning models; The comparison module is used to compare the formation brine layer identification results of the well logging curve identification module, the overlap method identification module, and the deep learning model identification module, so as to obtain the final formation brine layer identification result. A potassium-rich brine layer identification module is used to identify potassium-rich brine layers in the formation brine layer identification results. A module for determining the potassium content in potassium-rich brine is used to obtain the potassium content in the potassium-rich brine. A potassium salt enrichment layer identification module for brine is used to determine the potassium salt enrichment layer of brine based on the potassium content in the potassium-rich brine.

9. An apparatus comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes a computer program, it implements the method described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When executed by a processor, the computer program implements the method described in any one of claims 1 to 7.

Images