A mineralization prediction method and device based on fluid parameters and fluid field modeling, medium and product

Abstract

The application discloses a metallogenic prediction method and device based on fluid parameters and fluid field modeling, a medium and a product, and relates to the field of metallogenic prediction. First, petrographic study is carried out on fluid inclusion samples of a target area to determine the metallogenic stage; the fluid inclusion samples of different metallogenic stages are analyzed and tested to obtain multiple fluid parameters; exploratory data analysis is performed on the multiple fluid parameters to screen out key fluid parameters closely related to the content of main ore-forming elements; a machine learning or deep learning algorithm is used to model the numerical relationship between the key fluid parameters and the main ore-forming elements to determine an optimal numerical model; the key fluid parameters are processed by using different spatial interpolation algorithms to establish a three-dimensional fluid field model; according to the three-dimensional fluid field model and the optimal numerical model, three-dimensional metallogenic prediction is carried out in combination with metallogenic geological elements to determine a prospecting target area; since the fluid parameters are introduced as prediction variables, the accuracy of metallogenic prediction is improved.

Description

Technical Field

[0001] This application relates to the field of mineralization prediction technology, and in particular to a mineralization prediction method, equipment, medium and product based on fluid parameters and fluid field modeling. Background Technology

[0002] Traditional mineralization prediction methods mostly employ quantitative geoscience methods (such as the weight of evidence method) or similar analogy methods, and process conventional geological, geophysical, and geochemical parameter data. They do not consider the contribution of fluids to mineralization, nor do they treat fluids as prediction variables. Therefore, the accuracy of mineralization prediction results is not high. Summary of the Invention

[0003] In response to the problems pointed out in the background section, this application proposes a mineralization prediction method, equipment, medium, and product based on fluid parameters and fluid field modeling. By introducing fluid parameters as prediction variables, this application considers mineralization prediction elements more comprehensively and improves the accuracy of mineralization prediction.

[0004] To achieve the above objectives, this application provides the following solution.

[0005] Firstly, this application provides a mineralization prediction method based on fluid parameters and fluid field modeling, including:

[0006] Collect fluid inclusion samples from the target area and conduct petrographic studies to determine the mineralization stage of each fluid inclusion sample;

[0007] Fluid inclusion samples from different mineralization stages were analyzed and tested to obtain multiple fluid parameters for different mineralization stages. These multiple fluid parameters include homogenization temperature, freezing point temperature, salinity, mineralization pressure, fluid composition, inclusion size, and inclusion distribution density.

[0008] Exploratory data analysis was conducted on multiple fluid parameters at different mineralization stages to identify key fluid parameters that are closely related to the content of main ore-forming elements at different mineralization stages.

[0009] Machine learning or deep learning algorithms are used to model the numerical relationship between key fluid parameters and main ore-forming elements in different mineralization stages, and to determine the optimal numerical model for different mineralization stages.

[0010] Different spatial interpolation algorithms were used to process the key fluid parameters of different mineralization stages to establish three-dimensional fluid field models of different mineralization stages.

[0011] Based on the three-dimensional fluid field models and optimal numerical models of different mineralization stages, and combined with mineralization geological elements, three-dimensional mineralization prediction is carried out to determine the prospecting target area.

[0012] Optionally, the collection of fluid inclusion samples from the target area and the conduct of petrographic studies to determine the mineralization stage of each fluid inclusion sample specifically include:

[0013] Fluid inclusion petrographic studies were conducted under a polarizing microscope. Based on the symbiotic and interpenetrating relationships between minerals in the fluid inclusion samples, fluid inclusion samples from different mineralization stages were identified.

[0014] Optionally, the analysis and testing of fluid inclusion samples from different mineralization stages to obtain multiple fluid parameters for different mineralization stages specifically includes:

[0015] Homogenization temperature measurement was performed on fluid inclusion samples from different mineralization stages to obtain their homogenization temperature and freezing point temperature, and their salinity and mineralization pressure were calculated.

[0016] LA-ICP-MS analysis was performed on fluid inclusion samples from different mineralization stages to determine their trace components;

[0017] ICP-MS analysis was conducted on fluid inclusion samples from different mineralization stages to determine the geochemical element content in the fluids.

[0018] The fluid ion composition in fluid inclusion samples from different mineralization stages was analyzed using liquid chromatography.

[0019] Fluid inclusion samples from different mineralization stages were observed using a polarizing microscope. The size and abundance of the inclusions were statistically analyzed, and the distribution density of the inclusions was determined based on the abundance.

[0020] Optionally, the exploratory data analysis of multiple fluid parameters at different mineralization stages to screen out key fluid parameters closely related to the content of main ore-forming elements at different mineralization stages specifically includes:

[0021] Exploratory data analysis of multiple fluid parameters at different mineralization stages was conducted using statistical analysis methods to study their statistical characteristics and distribution patterns, thereby identifying key fluid parameters closely related to the content of main ore-forming elements at different mineralization stages. The statistical analysis methods included correlation analysis, factor analysis, and principal component analysis.

[0022] Optionally, the step of using machine learning or deep learning algorithms to model the numerical relationship between key fluid parameters and main ore-forming elements in different mineralization stages, and determining the optimal numerical model for different mineralization stages, specifically includes:

[0023] Using key fluid parameters from different mineralization stages as independent variables and the content of major ore-forming elements as dependent variables, multiple quantitative numerical models between independent and dependent variables are established using various machine learning or deep learning algorithms. The machine learning algorithms include linear regression, decision tree, support vector machine, and random forest algorithms; the deep learning algorithms include convolutional neural network, recurrent neural network, long short-term memory network, and Transformer.

[0024] The optimal numerical model is selected from multiple quantitative numerical models based on the best fit.

[0025] Optionally, the key fluid parameters for different mineralization stages are processed using different spatial interpolation algorithms to establish three-dimensional fluid field models for different mineralization stages, specifically including:

[0026] Using the spatial coordinates of fluid inclusion samples from different mineralization stages as independent variables and the corresponding key fluid parameters as dependent variables, various spatial interpolation models between independent and dependent variables are constructed using the three-dimensional distance power ratio method, ordinary kriging method, machine learning or deep learning algorithms, and the optimal spatial interpolation model is determined.

[0027] The three-dimensional geological model of the target area is divided into multiple block models, and the center coordinates of each block model are input into the optimal spatial interpolation model to obtain the key fluid parameters corresponding to each block model.

[0028] The key fluid parameters corresponding to each block model are used as their attribute values ​​to jointly construct three-dimensional fluid field models for different mineralization stages.

[0029] Optionally, the step of conducting three-dimensional mineralization prediction based on three-dimensional fluid field models and optimal numerical models at different mineralization stages, combined with mineralization geological elements, to determine mineral exploration target areas specifically includes:

[0030] Based on the three-dimensional geological model of the target area, the metallogenic geological elements corresponding to each block model are extracted, including lithology, fault structure and fault curvature;

[0031] By inputting the key fluid parameters corresponding to each block model into the optimal numerical model, the content of the main ore-forming elements corresponding to each block model is obtained.

[0032] The classification label is determined based on the content of the main ore-forming elements corresponding to each block model;

[0033] The key fluid parameters and mineralized geological elements corresponding to each block model are used as inputs, and the corresponding classification labels are used as outputs. Different machine learning or deep learning algorithms are used to construct multiple classification models.

[0034] The model with the best prediction performance among multiple classification models is selected as the three-dimensional mineralization prediction model.

[0035] The three-dimensional mineralization prediction model was used to determine the mineral exploration target area.

[0036] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the mineralization prediction method based on fluid parameters and fluid field modeling.

[0037] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the mineralization prediction method based on fluid parameters and fluid field modeling.

[0038] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the mineralization prediction method based on fluid parameters and fluid field modeling.

[0039] Based on the specific embodiments provided in this application, the following technical effects are disclosed.

[0040] This application provides a mineralization prediction method, equipment, medium, and product based on fluid parameters and fluid field modeling. Both the optimal numerical model and the three-dimensional fluid field model are established based on key fluid parameters. On this basis, three-dimensional mineralization prediction is carried out in conjunction with mineralization geological elements to determine prospecting target areas. Because mineralization prediction elements such as fluid parameters are considered more comprehensively, the accuracy of mineralization prediction is significantly improved. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a schematic diagram of the process of a mineralization prediction method based on fluid parameters and fluid field modeling according to this application;

[0043] Figure 2 This is a longitudinal and transverse cross-sectional view of the ore body generated in the embodiments of this application, as well as a schematic diagram of the locations of boreholes and sampling points;

[0044] Figure 3 This is a schematic diagram of the structure of the computer device in the embodiments of this application. Detailed Implementation

[0045] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0046] This application proposes a mineralization prediction method, equipment, medium, and product based on fluid parameters and fluid field modeling. The aim is to introduce fluid parameters as prediction variables, to more comprehensively consider mineralization prediction factors, and thus improve the accuracy of mineralization prediction.

[0047] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] In one exemplary embodiment, such as Figure 1 As shown, a mineralization prediction method based on fluid parameters and fluid field modeling is provided, including the following steps 1 to 6.

[0049] Step 1: Collect fluid inclusion samples from the target area and conduct petrographic studies to determine the mineralization stage of each fluid inclusion sample.

[0050] Before collecting fluid inclusion samples from the target area (also known as the study area), it is necessary to first gather the latest geological data for the target area and update the three-dimensional geological model of the target area. The latest geological data includes data obtained from recent drilling and pitting projects, such as: basic geological data such as geological maps, stratigraphic data, and rock mass data; borehole data such as borehole data and borehole analysis data; structural data such as structural maps and structural descriptions; geophysical data such as gravity data, magnetic data, seismic data, and electromagnetic data; geochemical data such as soil geochemical data, water geochemical data, and rock geochemical data; remote sensing data such as satellite imagery and UAV aerial imagery; and other data such as topographic data, historical geological data, mining data, and hydrogeological data. Based on the latest geological data, the latest geological information, such as stratigraphic, rock mass, lithology, and structural information, is extracted. Based on the newly acquired geological information and combined with previously collected geological information, three-dimensional modeling technology is used. Using three-dimensional modeling software as a platform, either explicit (interpreting each section individually and connecting corresponding parts between sections) or implicit modeling techniques are employed to build a three-dimensional geological model of the target area. The ore body model is included in the 3D geological model, and the ore body model is a part of the 3D geological model.

[0051] Furthermore, using the sectioning function of 3D modeling software, longitudinal and transverse cross-sections of the ore body are automatically generated according to the section spacing. After the cross-sections are generated, technicians determine the sampling locations based on the geological information in the cross-sections.

[0052] Mineral-bearing samples are collected from sampling locations within the ore body of the target area, such as in drill cores or tunnels, using a geological hammer or other sampling tools. There are no restrictions on sample size, as long as it is large enough to prepare thin sections of fluid inclusions. Sampling intervals within the ore body or mineralized zone are 0.5–1 m. During collection, information such as the borehole number, location, depth, and lithology of the sample are recorded. The location of the sample can be determined using the recorded borehole number, location, and depth. Recording the lithology information is for statistical analysis of fluid inclusion parameters in different lithologies. After collection, the mineral-bearing samples are prepared into thin sections of fluid inclusions, which serve as the fluid inclusion samples for this application.

[0053] Furthermore, petrographic studies were conducted on each fluid inclusion sample under a polarizing microscope. Based on the symbiotic and interpenetrating relationships between minerals in the fluid inclusion samples, fluid inclusion samples from different mineralization stages were identified.

[0054] Petrographic studies of fluid inclusions (also referred to as inclusions) are fundamental to establishing three-dimensional fluid fields at different mineralization stages. Petrographic studies of collected fluid inclusion samples are conducted under a polarizing microscope. Detailed observation and statistical analysis of the abundance, size, and other information of the fluid inclusion samples are performed, and the types of fluid inclusions (primary or secondary inclusions) are identified. Using mineral interpenetration and assemblage relationships, the types of fluid inclusions and their differences at different stages before, during, and after mineralization are determined, identifying the mineralization stage corresponding to each fluid inclusion sample—pre-mineralization, during, or after mineralization. Different mineralization stages of fluid inclusions are also referred to as different phases of fluid inclusions.

[0055] Step 2: Analyze and test fluid inclusion samples from different mineralization stages to obtain multiple fluid parameters for different mineralization stages; the multiple fluid parameters include homogenization temperature, freezing point temperature, salinity, mineralization pressure, fluid composition, inclusion size, and inclusion distribution density.

[0056] The analytical methods used for fluid inclusion samples from different mineralization stages mainly include petrographic observation, homogenization thermometry, LA-ICP-MS, ICP-MS, and liquid chromatography. Among these, ICP-MS (Inductively Coupled Plasma Mass Spectrometry) is a highly sensitive and precise analytical technique widely used in earth sciences, environmental sciences, materials science, and biomedicine. It combines the high-temperature ionization characteristics of inductively coupled plasma (ICP) with the high-sensitivity detection capability of mass spectrometry (MS), enabling the simultaneous determination of the content and isotope ratios of multiple elements. LA-ICP-MS (Laser Ablation Inductively Coupled Plasma Mass Spectrometry) is a high-precision micro-area analysis technique capable of in-situ analysis of small samples (such as fluid inclusion thin sections).

[0057] On the one hand, for fluid inclusion samples from different mineralization stages (including pre-mineralization, mineralization period and post-mineralization stages), homogenization temperature measurement was carried out using a hot and cold stage to obtain parameters such as homogenization temperature and freezing point temperature, and to calculate their salinity and mineralization pressure.

[0058] On the other hand, LA-ICP-MS technology was used to conduct LA-ICP-MS tests on fluid inclusion samples from different mineralization stages to determine the trace components in the fluid inclusion samples, including the composition and content of trace elements (such as Li, B, Sr, Ba, Zn, Pb, U, etc.) and rare gases (such as He, Ne, Ar, etc.) in the fluid inclusions. Among them, Pb and Zn are considered to be the main ore-forming elements that are highly relevant to mineralization.

[0059] On the other hand, ICP-MS analysis was conducted on fluid inclusion samples from different mineralization stages to determine the geochemical element content of the fluids in the fluid inclusion samples. In fluid inclusion studies, geochemical element content refers to the concentration or mass fraction of various elements in the fluid inclusions. These elements include major elements, trace elements, and rare gases, which can provide important information about the source, evolution process, and mineralization environment of the ore-forming fluids.

[0060] On the other hand, liquid chromatography is used to analyze the liquid components, such as fluid ionic components, in fluid inclusion samples from different mineralization stages. In fluid inclusion studies, fluid ionic components refer to the various ions dissolved in the fluid inclusions. These ions can be metal ions, non-metal ions, anions, or cations, etc., and they exist in a dissolved state in the fluid, providing important information about the chemical properties, origin, and evolution of the mineralizing fluids.

[0061] On the other hand, fluid inclusion samples from different mineralization stages were observed using a polarizing microscope to statistically analyze parameters such as the abundance and size of fluid inclusions. The abundance of fluid inclusions refers to the number of fluid inclusions per unit volume or unit area, reflecting the distribution density of fluid inclusions in the mineral.

[0062] The analytical test results of fluid inclusions at different stages were compiled to establish a four-dimensional spatiotemporal database, which stores the mineralization stage corresponding to the fluid inclusion samples, as well as various fluid parameters such as homogenization temperature, freezing point temperature, salinity, mineralization pressure, fluid composition, inclusion size, and inclusion distribution density.

[0063] Step 3: Conduct exploratory data analysis on multiple fluid parameters at different mineralization stages to screen out key fluid parameters that are closely related to the content of main ore-forming elements at different mineralization stages.

[0064] Exploratory data analysis refers to the use of statistical analysis methods such as correlation analysis, factor analysis, and principal component analysis to analyze various fluid parameter variables, understand the data distribution characteristics, and obtain the relationship between various fluid parameters and the contents of main ore-forming elements (Pb, Zn) at different mineralization stages. Factor analysis and principal component analysis can create new combined variables, perform dimensionality reduction on the data, and screen out key fluid parameters that are closely related to the contents of main ore-forming elements.

[0065] Because the analysis and testing of fluid inclusions yielded a large variety of fluid parameters and a substantial amount of data, some parameters, such as those related to the mineralization stage of fluid inclusions, were not simple numerical values. Traditional statistical analysis methods were insufficient to reveal the relationships between these variables. Therefore, exploratory analysis techniques, such as correlation analysis, factor analysis, and principal component analysis, were employed to study the statistical characteristics and distribution patterns of each fluid parameter. These techniques can be implemented using statistical software. By studying the statistical characteristics and distribution patterns of each fluid parameter, parameter variables can be further optimized or simplified, or new parameter variables can be constructed. Ultimately, fluid parameters closely related to the content of the main ore-forming elements can be selected, termed key fluid parameters, which serve as the basis for mineralization prediction. These variables form the foundation for subsequent mineralization prediction. By analyzing the changes in key fluid parameters (temperature, salinity, pressure, trace element content, etc.) of fluid inclusions at different mineralization stages, the spatiotemporal variation patterns of different mineralization stages can be explored. These spatiotemporal variation patterns can guide mineralization prediction.

[0066] Step 4: Use machine learning or deep learning algorithms to model the numerical relationship between key fluid parameters and main ore-forming elements in different mineralization stages, and determine the optimal numerical model for different mineralization stages.

[0067] Specifically, using key fluid parameters selected from different mineralization stages as independent variables and the content of major ore-forming elements as dependent variables, a dataset for training the quantitative numerical model was constructed and divided into training, validation, and test sets. The parameter data in the training, validation, and test sets all come from a spatiotemporal database. Generally, the training and validation sets comprise 70% of the database data, which can be obtained through random sampling, while the remaining 30% is the test set data.

[0068] Based on the training set data, various machine learning or deep learning algorithms are used to establish multiple quantitative numerical models between independent and dependent variables. The machine learning algorithms used may include Linear Regression, Decision Tree (CART), Support Vector Machine (SVM), and Random Forest. The deep learning algorithms used may include Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Transformer. During model training, validation set data is used to evaluate the model's performance, and the hyperparameters are adjusted based on the performance on the validation set to obtain the trained quantitative numerical model.

[0069] After model training, all quantitative numerical models are tested using test set data to evaluate their accuracy. Specifically, metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-squared Score can be used to evaluate model effectiveness. The model with the best fit (highest accuracy) from among multiple quantitative numerical models is selected as the optimal numerical model.

[0070] The following explanation uses a recurrent neural network (RNN) as an example.

[0071] The basic form of an RNN model is as follows:

[0072] (1)

[0073] (2)

[0074] in, In time step t Input; In time step t The output of . In time step t The hidden state; In time step t-1 is a hidden state. It is in a hidden state. To hide The weight matrix. It is input To hide The weight matrix. It is the bias vector of the hidden state. f ( ) is the activation function, usually tanh or ReLU. It is in a hidden state. To output The weight matrix. It is the output bias vector.

[0075] The process of constructing a quantitative numerical model based on a recurrent neural network (RNN) is as follows:

[0076] 4.1) Data preparation: Collect data on key fluid parameters and the content of main ore-forming elements at different mineralization stages, and divide the data into training set and validation set;

[0077] 4.2) Model initialization: Initialize the parameters of the RNN model, including the weight matrix and bias vector;

[0078] 4.3) Forward propagation: For each time step t Calculate the hidden state and output ;

[0079] 4.4) Loss Calculation: Calculate the loss function of the model, usually using mean squared error (MSE) or root mean square error (RMSE).

[0080] 4.5) Backpropagation: The gradient of the loss function with respect to the model parameters is calculated using the backpropagation algorithm. The model parameters are then updated using optimization algorithms such as gradient descent or Adam.

[0081] 4.6) Iterative training: Repeat steps 4.3) to 4.5) until the model converges or the maximum number of iterations is reached;

[0082] 4.7) Model Evaluation: Evaluate the model's performance using validation set data, and calculate the MSE or RMSE on the validation set;

[0083] 4.8) Parameter tuning: Adjust the model's hyperparameters, such as learning rate, hidden layer size, activation function, etc., based on the performance on the validation set;

[0084] 4.9) Final Model: Select the RNN model that performs best on the validation set as the final quantitative numerical model.

[0085] Step 5: Use different spatial interpolation algorithms to process the key fluid parameters of different mineralization stages and establish three-dimensional fluid field models for different mineralization stages.

[0086] Fluid field simulation is one of the key aspects of this application. A three-dimensional spatial dataset is constructed using the spatial coordinates of fluid inclusion samples from different mineralization stages as independent variables and the corresponding key fluid parameters as dependent variables. Based on this three-dimensional spatial dataset, fluid field simulations for different mineralization stages are conducted. Step 5 specifically includes steps 5.1 to 5.3.

[0087] Step 5.1: Using the spatial coordinates of fluid inclusion samples from different mineralization stages as independent variables and the corresponding key fluid parameters as dependent variables, construct various spatial interpolation models between the independent and dependent variables using the three-dimensional distance power ratio method, ordinary kriging method, machine learning or deep learning algorithms, and determine the optimal spatial interpolation model.

[0088] Specifically, the 3D spatial dataset is divided into training, validation, and test sets. Algorithms such as Inverse Distance Weight (IDW), Ordinary Kriging (OK), machine learning, or deep learning are used to construct various spatial interpolation models between independent and dependent variables in the training set. During model training, the validation set is used to evaluate the model's performance, and the hyperparameters are adjusted based on the performance on the validation set to obtain the trained spatial interpolation model. After model training is complete, the test set is used to test all spatial interpolation models, evaluate their accuracy, and determine the spatial interpolation model with the highest accuracy as the optimal spatial interpolation model.

[0089] The following explanation uses support vector machines as an example:

[0090] A) First, the three-dimensional spatial dataset is divided into two parts: a training set and a validation set. The model is trained using the training set first. The sample spatial coordinates (x, y, z) are used as independent variables, and the key fluid parameters of the sample are used as dependent variables. The kernel function (polynomial function or radial basis function) required by the support vector machine model is selected, and the support vector machine model is trained.

[0091] B) The parameters of the kernel function of the support vector machine are optimized using a grid search method to obtain the optimal parameters;

[0092] C) Retrain the support vector machine model using the optimal kernel function parameters;

[0093] D) Use a validation set to validate the support vector machine model and evaluate its accuracy;

[0094] The support vector machine model obtained through the above training is the spatial interpolation model, which forms the basis for deriving key fluid parameters from spatial coordinates later. Only with this spatial interpolation model can the key fluid parameters at each spatial location be estimated.

[0095] The spatial interpolation model of this application is obtained by interpolating the relationship between key fluid parameters and spatial coordinates of the fluid inclusion body through machine learning or deep learning. The resulting model is used to construct the relationship between each spatial coordinate (x, y, z) and the aforementioned key fluid parameters, thereby constructing a three-dimensional fluid field.

[0096] Step 5.2: Divide the three-dimensional geological model of the target area into multiple block models, and input the center coordinates of each block model into the optimal spatial interpolation model to obtain the key fluid parameters corresponding to each block model.

[0097] In 3D geological software, the size of the block is determined based on the distribution range of the samples and the distribution of boreholes. The 3D geological model of the target area is further divided into multiple block models, and the center coordinates (x, y, z) of each block model are exported to a .csv file to obtain a spatial coordinate file.

[0098] Input the spatial coordinate file into the optimal spatial interpolation model obtained in step 5.1 to obtain the predicted dependent variable value, which is the key fluid parameter at that spatial coordinate location. Save all the key fluid parameter values ​​with spatial coordinates obtained in the above steps into a new .csv file.

[0099] Step 5.3: Use the key fluid parameters corresponding to each block model as its attribute values ​​to jointly construct a three-dimensional fluid field model for different mineralization stages.

[0100] Import the new .csv file from step 5.2 into the 3D geological software, and use the key fluid parameters corresponding to each block model as its attribute values ​​to form a numerical 3D model of fluid inclusions. The numerical 3D model of fluid inclusions in a certain mineralization stage is the 3D fluid field model of that mineralization stage.

[0101] Step 6: Based on the three-dimensional fluid field model and optimal numerical model of different mineralization stages, and combined with mineralization geological elements, carry out three-dimensional mineralization prediction and determine the prospecting target area.

[0102] The three-dimensional geological model of the target area, as well as the three-dimensional fluid field model and optimal numerical model of different mineralization stages constructed in the above steps, are the foundation and key to carrying out mineralization prediction. Step 6 specifically includes the following steps 6.1 to 6.6.

[0103] Step 6.1: Extract the mineralized geological elements corresponding to each block model based on the three-dimensional geological model of the target area, including lithology, fault structure and fault curvature.

[0104] Based on geologists' research on mineral deposit geology, metallogenic geological elements related to mineralization are selected. Then, based on the updated three-dimensional geological model and three-dimensional fluid field model, metallogenic geological elements corresponding to each block model are extracted, such as lithology, fault structure, and fault curvature.

[0105] Step 6.2: Input the key fluid parameters corresponding to each block model into the optimal numerical model to obtain the content of the main ore-forming elements corresponding to each block model.

[0106] During the sampling phase, only the content of principal ore-forming elements at the location of fluid inclusion samples can be obtained. However, when training the classification model, relevant data from all block models in the target area are needed as training samples. Therefore, the optimal numerical models for different ore-forming stages obtained in step 4 can be used to generate the corresponding principal ore-forming element contents based on the key fluid parameters of each block model.

[0107] Step 6.3: Determine the classification label based on the content of the main ore-forming elements corresponding to each block model.

[0108] For example, samples where the sum of the contents of the main ore-forming elements Pb and Zn (Pb+Zn) is greater than the boundary grade are assigned a value of 1, and samples where the content is less than the boundary grade are assigned a value of 0, thus generating their classification label (named Ore). Here, the classification label 1 represents ore and 0 represents non-ore.

[0109] Step 6.4: Using the key fluid parameters and mineralized geological elements corresponding to each block model as input and the corresponding classification labels as output, construct multiple classification models using different machine learning or deep learning algorithms.

[0110] For each block model, its metallogenic geological elements and key fluid parameters are exported as data with spatial coordinates, which can be exported as a .csv file. The key fluid parameters and metallogenic geological elements corresponding to each block model are used as input variables, and the corresponding classification labels (0 or 1) are used as output variables to form a new dataset. The new dataset is then divided into training, validation, and test sets.

[0111] Based on the training set data, various machine learning or deep learning algorithms are used to build and train multiple classification models between input and output variables. The machine learning algorithms used may include Logistic Regression, Decision Tree (CART), Support Vector Machine (SVM), and Random Forest. The deep learning algorithms used may include Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Transformer. During model training, validation set data is used to evaluate the model's performance, and the hyperparameters are adjusted based on the performance on the validation set to obtain the trained classification model.

[0112] After model training, all classification models are tested using test set data to evaluate their accuracy. Specifically, metrics such as ROC curves and AUC values ​​can be used to evaluate model effectiveness. The model with the highest accuracy from multiple classification models is then selected as the 3D mineralization prediction model.

[0113] The following explanation uses the random forest algorithm as an example:

[0114] a) Divide the sample dataset into two subsets: training samples (70% of all samples) and validation samples (30% of all samples);

[0115] b) A random forest algorithm is used to train a subset of training samples. Key fluid parameters and ore-forming geological elements are used as input variables, and the corresponding classification labels (0 or 1) are used as output variables to create a random forest classification model. Debugging tools are used to optimize parameters such as ntree and split.

[0116] c) The above classification model is validated using a validation set to evaluate its accuracy. ROC curves and AUC values ​​are used to evaluate the model's effectiveness.

[0117] Step 6.5: Select the model with the best prediction performance from multiple classification models as the three-dimensional mineralization prediction model.

[0118] Following the steps outlined above, different algorithms are applied to create classification models, and the ROC curves and AUC values ​​of each classification model are obtained. The classification model with the largest AUC value is then selected as the final three-dimensional mineralization prediction model for mineralization prediction.

[0119] Step 6.6: Use the three-dimensional mineralization prediction model to determine the mineral exploration target area.

[0120] When performing mineralization prediction, simply divide the three-dimensional geological model of any study area into block models, and input the key fluid parameters and mineralization geological elements of each block model into the three-dimensional mineralization prediction model. The corresponding classification label (0 or 1) will be output, and the area corresponding to the block model with the classification label of 1 will be identified as the mineral exploration target area.

[0121] Guided by modern metallogenic geology theory and supported by modern fluid inclusion analysis and testing techniques, 3D modeling techniques, machine learning (deep learning) algorithms, and spatial interpolation algorithms, this application investigates methods for constructing 3D fluid fields at different metallogenic stages, based on a research approach that combines theoretical research and practical analysis, thereby guiding mineralization prediction. Unlike other mineralization prediction methods, this application's method is based on fluid parameters of fluid inclusions. It uses key fluid parameters of fluid inclusions, along with metallogenic geological elements of the study area, as prediction variables, thus comprehensively considering mineralization prediction factors, possessing greater geological significance, and providing more reliable and accurate predictions for hydrothermal deposits.

[0122] This application addresses two major scientific questions: the characteristics of fluid evolution over time in different mineralization stages of hydrothermal deposits and their quantitative relationship with ore-forming elements; and mineralization prediction based on ore-forming fluid models. It designs four key technical points: ① determination of fluid inclusions in different mineralization stages; ② establishment of quantitative numerical models based on key fluid parameters of fluid inclusions; ③ construction of three-dimensional fluid field models for different mineralization stages; and ④ construction of a three-dimensional mineralization prediction model. Applying these key technical points and technical routes, the following section details the specific implementation process of the proposed method, using the Caixiashan lead-zinc deposit as the target area, including S1 to S6.

[0123] S1: Collect fluid inclusion samples from the target area and conduct petrographic studies to determine the mineralization stage of each fluid inclusion sample.

[0124] The latest geological data was collected to update the 3D geological model of the Caixia Mountain area. Using the cross-section function of 3D geological software, longitudinal and transverse profiles of the ore body were generated based on the updated 3D geological model to determine sampling locations. The longitudinal and transverse profiles of the ore body, as well as the locations of boreholes and sampling points, are shown below. Figure 2 As shown, the labels in the cross-sectional diagram are the numbers of the ore bodies.

[0125] Sampling will be conducted on Orebody II of the Caixia Mountain main orebody. The hanging wall of this orebody is composed of marble, and the footwall is composed of siltstone. Samples will be collected from borehole cores or tunnels. Specifically, sphalerite-bearing samples will be collected within the orebody or mineralized zone, with sampling intervals of 0.5–1 m. During collection, the borehole number, location, depth, and lithology of the sample will be recorded. After sample collection, fluid inclusion thin sections will be prepared. Since samples have already been collected, an additional 120 fluid inclusion samples are expected to be collected this time.

[0126] Furthermore, petrographic studies of fluid inclusions will be conducted to define the mineralization stages of fluid inclusion samples and to identify fluid inclusions at different stages.

[0127] Specifically, detailed and systematic observations are conducted under a microscope to identify fluid inclusions at different mineralization stages (i.e., different phases) based on the symbiotic and interpenetrating relationships between minerals. For example, inclusions in sphalerite or primary inclusions in quartz and calcite associated with sphalerite are identified as mineralization-stage inclusions; primary inclusions in carbonate minerals interpenetrated by metallic sulfides such as sphalerite are identified as pre-mineralization inclusions; and inclusions in carbonate minerals or quartz veins interpenetrating metallic sulfides are identified as post-mineralization inclusions.

[0128] S2: Analyze and test fluid inclusion samples from different mineralization stages to obtain multiple fluid parameters for different mineralization stages; the multiple fluid parameters include homogenization temperature, freezing point temperature, salinity, mineralization pressure, fluid composition, inclusion size, and inclusion distribution density.

[0129] Fluid inclusion thin sections were observed using a polarizing microscope. Abundance, size, and other parameters of the fluid inclusion samples were statistically analyzed to determine their properties and classify them into pre-mineralization, mineralization, and post-mineralization stages. Homogenization temperature measurements were performed using a hot-cold stage to obtain parameters such as homogenization temperature and freezing point temperature, and to calculate salinity and mineralization pressure. After homogenization temperature measurements of the fluid inclusion samples, samples with larger diameters (>20 μm) and from different stages were selected for LA-ICP-MS analysis to determine the trace element composition of the fluids at different stages. Single minerals from the mineralization stages were selected, and the ionic composition of the mineralization fluids was analyzed using liquid chromatography. The geochemical element content of the fluids in the fluid inclusion samples was determined using ICP-MS.

[0130] S3: Exploratory data analysis of multiple fluid parameters at different mineralization stages was conducted to screen out key fluid parameters that are closely related to the content of main ore-forming elements at different mineralization stages.

[0131] By using statistical software and employing exploratory analysis methods such as correlation analysis, factor analysis, and principal component analysis, we can study the statistical characteristics and distribution patterns of various fluid parameters, further optimize and simplify parameter variables, or construct new parameter variables, and finally screen out the fluid parameters that are closely related to the content of the main ore-forming elements as key fluid parameters.

[0132] S4: Use machine learning or deep learning algorithms to model the numerical relationship between key fluid parameters and main ore-forming elements in different mineralization stages, and determine the optimal numerical model for different mineralization stages.

[0133] This step involves modeling the quantitative relationship between key fluid parameters and the content of major ore-forming elements. For key fluid parameters (such as homogenization temperature, salinity, freezing point temperature, fluid composition, inclusion size, and inclusion distribution density) of samples from different periods, machine learning or deep learning algorithms, such as decision trees, linear regression, random forests, support vector machines, and convolutional neural networks, are used to reveal the numerical relationship between key fluid parameters and major ore-forming elements in samples from different periods. The quantitative relationship between the content of major ore-forming elements (Pb, Zn, etc.) and key fluid parameters from different periods is studied, and the optimal numerical model for different mineralization stages is determined.

[0134] S5: Different spatial interpolation algorithms are used to process the key fluid parameters of different mineralization stages to establish three-dimensional fluid field models of different mineralization stages.

[0135] This step involves establishing three-dimensional fluid fields for different mineralization stages. By processing key fluid parameters at different mineralization stages using different spatial interpolation algorithms, three-dimensional fluid fields for each stage are established. Cross-validation, sampling with replacement, and hold-out methods are employed to compare the spatial interpolation models and three-dimensional fluid field models established by different algorithms. Performance metrics such as MSE, RMSE, and MAE are used to evaluate the performance of various algorithms and models, ultimately selecting the optimal spatial interpolation algorithm and the three-dimensional fluid field model with the highest accuracy.

[0136] S6: Based on the three-dimensional fluid field model and optimal numerical model of different mineralization stages, and combined with mineralization geological elements, three-dimensional mineralization prediction is carried out to determine the prospecting target area.

[0137] This step involves three-dimensional mineralization prediction based on fluid field information. Based on three-dimensional fluid field models of different mineralization stages and the optimal numerical model relating key fluid parameters to the content of major ore-forming elements, combined with information on mineralization geological elements such as deposit geology, structure, alteration, and magmatic intrusions, a three-dimensional mineralization prediction model is established, and the prediction accuracy is evaluated. Based on this three-dimensional mineralization prediction model, three-dimensional quantitative mineralization prediction can accurately determine prospecting target areas. This application, by introducing fluid parameters as prediction variables, more comprehensively considers mineralization prediction elements, making it more geologically significant and more reliable for predicting hydrothermal deposits, with broad application prospects.

[0138] In one exemplary embodiment, this application also provides a computer device, which may be a server or a terminal. Figure 3As shown, the computer device includes a processor, memory, input / output interfaces, and a communication interface. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the described mineralization prediction method based on fluid parameters and fluid field modeling.

[0139] In one exemplary embodiment, this application also provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the described mineralization prediction method based on fluid parameters and fluid field modeling.

[0140] In one exemplary embodiment, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned mineralization prediction method based on fluid parameters and fluid field modeling.

[0141] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by hardware related to computer program instructions. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any reference to memory or other media in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0142] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0143] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A mineralization prediction method based on fluid parameters and fluid field modeling, characterized in that, include: Collect fluid inclusion samples from the target area and conduct petrographic studies to determine the mineralization stage of each fluid inclusion sample; Fluid inclusion samples from different mineralization stages were analyzed and tested to obtain multiple fluid parameters for different mineralization stages. These multiple fluid parameters include homogenization temperature, freezing point temperature, salinity, mineralization pressure, fluid composition, inclusion size, and inclusion distribution density. Exploratory data analysis was conducted on multiple fluid parameters at different mineralization stages to identify key fluid parameters that are closely related to the content of main ore-forming elements at different mineralization stages. Machine learning or deep learning algorithms are used to model the numerical relationships between key fluid parameters and main ore-forming elements in different mineralization stages, and to determine the optimal numerical model for each mineralization stage. Specifically, this involves using key fluid parameters as independent variables and the content of main ore-forming elements as dependent variables, and employing various machine learning or deep learning algorithms to establish multiple quantitative numerical models between the independent and dependent variables. Machine learning algorithms include linear regression, decision trees, support vector machines, and random forests; deep learning algorithms include convolutional neural networks, recurrent neural networks, long short-term memory networks, and Transformers. The model with the best fit among these quantitative numerical models is then selected as the optimal numerical model. Different spatial interpolation algorithms were used to process the key fluid parameters of different mineralization stages to establish three-dimensional fluid field models of different mineralization stages. Based on the three-dimensional fluid field models and optimal numerical models of different mineralization stages, and combined with mineralization geological elements, three-dimensional mineralization prediction is carried out to determine the prospecting target area.

2. The mineralization prediction method based on fluid parameters and fluid field modeling according to claim 1, characterized in that, The collection of fluid inclusion samples from the target area and the conduct of petrographic studies to determine the mineralization stage of each fluid inclusion sample specifically include: Fluid inclusion petrographic studies were conducted under a polarizing microscope. Based on the symbiotic and interpenetrating relationships between minerals in the fluid inclusion samples, fluid inclusion samples from different mineralization stages were identified.

3. The mineralization prediction method based on fluid parameters and fluid field modeling according to claim 2, characterized in that, The analysis and testing of fluid inclusion samples from different mineralization stages yielded multiple fluid parameters for each mineralization stage, specifically including: Homogenization temperature measurement was performed on fluid inclusion samples from different mineralization stages to obtain their homogenization temperature and freezing point temperature, and their salinity and mineralization pressure were calculated. LA-ICP-MS analysis was performed on fluid inclusion samples from different mineralization stages to determine their trace components; ICP-MS analysis was conducted on fluid inclusion samples from different mineralization stages to determine the geochemical element content in the fluids. The fluid ion composition in fluid inclusion samples from different mineralization stages was analyzed using liquid chromatography. Fluid inclusion samples from different mineralization stages were observed using a polarizing microscope. The size and abundance of the inclusions were statistically analyzed, and the distribution density of the inclusions was determined based on the abundance.

4. The mineralization prediction method based on fluid parameters and fluid field modeling according to claim 3, characterized in that, The exploratory data analysis of multiple fluid parameters at different mineralization stages identified key fluid parameters closely related to the content of main ore-forming elements at different mineralization stages, specifically including: Exploratory data analysis of multiple fluid parameters at different mineralization stages was conducted using statistical analysis methods to study their statistical characteristics and distribution patterns, thereby identifying key fluid parameters closely related to the content of main ore-forming elements at different mineralization stages. The statistical analysis methods included correlation analysis, factor analysis, and principal component analysis.

5. The mineralization prediction method based on fluid parameters and fluid field modeling according to claim 4, characterized in that, The key fluid parameters for different mineralization stages are processed using different spatial interpolation algorithms to establish three-dimensional fluid field models for different mineralization stages, specifically including: Using the spatial coordinates of fluid inclusion samples from different mineralization stages as independent variables and the corresponding key fluid parameters as dependent variables, multiple spatial interpolation models between independent and dependent variables were constructed using the three-dimensional distance power ratio method, ordinary kriging method, machine learning or deep learning algorithms, and the spatial interpolation model with the highest accuracy was determined as the optimal spatial interpolation model. The three-dimensional geological model of the target area is divided into multiple block models, and the center coordinates of each block model are input into the optimal spatial interpolation model to obtain the key fluid parameters corresponding to each block model. The key fluid parameters corresponding to each block model are used as their attribute values ​​to jointly construct three-dimensional fluid field models for different mineralization stages.

6. The mineralization prediction method based on fluid parameters and fluid field modeling according to claim 5, characterized in that, The method involves using three-dimensional fluid field models and optimal numerical models for different mineralization stages, combined with mineralization geological elements, to conduct three-dimensional mineralization prediction and determine prospecting target areas. Specifically, this includes: Based on the three-dimensional geological model of the target area, the metallogenic geological elements corresponding to each block model are extracted, including lithology, fault structure and fault curvature; By inputting the key fluid parameters corresponding to each block model into the optimal numerical model, the content of the main ore-forming elements corresponding to each block model is obtained. The classification label is determined based on the content of the main ore-forming elements corresponding to each block model; The key fluid parameters and mineralized geological elements corresponding to each block model are used as inputs, and the corresponding classification labels are used as outputs. Different machine learning or deep learning algorithms are used to construct multiple classification models. The model with the best prediction performance among multiple classification models is selected as the three-dimensional mineralization prediction model. The three-dimensional mineralization prediction model was used to determine the mineral exploration target area.

7. A computer device, comprising: 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 mineralization prediction method based on fluid parameters and fluid field modeling as described in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the mineralization prediction method based on fluid parameters and fluid field modeling as described in any one of claims 1 to 6.

9. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the mineralization prediction method based on fluid parameters and fluid field modeling as described in any one of claims 1 to 6.

Images