Crystal uranium mine cause analysis method based on mineralogy and geochemistry

Abstract

The invention discloses a mineralogy and geochemistry-based uraninite cause analysis method, and relates to the technical field of uranium ore geological exploration, and the method comprises the following steps: sample collection and pretreatment; carrying out mineralogical characteristic analysis; performing geochemical tracing analysis; performing data comprehensive analysis and cause judgment; carrying out comprehensive arrangement on data of mineralogy characteristic analysis and geochemical tracing analysis, constructing a cause judgment model by applying a multivariate statistical analysis method, and accurately judging the cause type, the mineralization environment and the material source of the uraninite in combination with a regional geological background; the multivariate statistical analysis method comprises principal component analysis and clustering analysis, and construction of a cause discrimination model, and the cause discrimination model adopts a gradient boosting decision tree or a random forest algorithm, so that the problems that the existing uraninite cause analysis method depends on a single research means, so that the revelation of a microstructure, a material source and a mineralization environment is not comprehensive, and the analysis efficiency is high are solved. The cause mechanism is difficult to accurately judge, and the analysis limitation and uncertainty exist.

Description

Technical Field

[0001] The present invention relates to the technical field of uranium ore geological exploration, and particularly to a method for analyzing the genesis of uraninite based on mineralogy and geochemistry. Background Art

[0002] Uraninite is an important uranium ore mineral, and the study of its genesis is of crucial significance for understanding the uranium mineralization mechanism and predicting potential uranium resources. Traditional methods for analyzing the genesis of uraninite often focus on single research means, such as only relying on mineralogical characteristics or a certain aspect of geochemistry for analysis, which makes the understanding of the genesis of uraninite limited and one-sided.

[0003] In terms of mineralogical research, previous analysis techniques may not be able to comprehensively and deeply reveal the microscopic structure and mineral composition changes of uraninite, and it is difficult to accurately grasp its formation process and evolution mechanism under different geological conditions. In geochemical research, single element analysis or isotope testing methods may not be able to provide enough information to trace the source of ore-forming materials and judge the ore-forming environment, resulting in great uncertainty in the judgment of the genesis of uraninite.

[0004] Therefore, it is of great practical significance to develop a method for analyzing the genesis of uraninite that comprehensively considers various aspects of information in mineralogy and geochemistry. Summary of the Invention

[0005] Based on this, in view of the above problems, the present invention proposes a method for analyzing the genesis of uraninite based on mineralogy and geochemistry, which solves the problems that the current methods for analyzing the genesis of uraninite rely on single research means, resulting in incomplete revelation of the microscopic structure, material source and ore-forming environment, difficult to accurately identify the genesis mechanism, and having limitations and uncertainties in analysis.

[0006] The technical solution of the present invention is as follows: A method for analyzing the genesis of uraninite based on mineralogy and geochemistry, comprising the following steps: A: Sample collection and pretreatment; B: Mineralogical characteristic analysis; C: Geochemical tracer analysis; D: Comprehensive data analysis and genesis judgment; Comprehensively organize the data of mineralogical characteristic analysis and geochemical tracer analysis, use multivariate statistical analysis methods to construct a genesis discrimination model, and accurately judge the genesis type, ore-forming environment and material source of uraninite in combination with the regional geological background; Among them, the multivariate statistical analysis methods include principal component analysis and cluster analysis, which are used to find out the key factors and main variables affecting the genesis of uraninite, and construct a genetic discrimination model. The genetic discrimination model adopts the gradient boosting decision tree or random forest algorithm, and the input parameters include: Mineralogical parameters: unit cell parameters, trace element distribution gradient, inclusion characteristics; Geochemical parameters: REE partition pattern, Sr-Nd-Pb isotope ratio, U / Th value; Geochronological parameters: error range of ore-forming age; Regional geological parameters: fracture activity intensity, wall rock alteration type; The training process includes: Based on 500 uranium ore samples with known genesis globally, supervised learning is carried out according to the proportions of magmatic type, hydrothermal type, and sedimentary type being 40%, 35%, and 25% respectively; Five-fold cross-validation is used to optimize the model hyperparameters to ensure generalization ability; Key discriminant factors are identified through SHAP value analysis: ΣREE, εNd(t), unit cell parameters.

[0007] Preferably, step A is specifically: Select representative uraninite samples and related wall rock samples, and carry out cleaning, crushing, and grinding treatments; Among them, the sample collection should follow the principle of systematic sampling, and sampling points are arranged according to different geological units, mineralization zones, and lithological characteristics in the entire mineralized area to ensure that the samples can reflect the characteristics of uraninite under different mineralization stages, different geological environments, and different lithological combinations; The cleaning treatment uses deionized water or high-purity organic solvents to avoid introducing impurity elements to interfere with subsequent analysis; the crushing process needs to control the crushing ratio to prevent the sample from being overly crushed and causing damage to the mineral structure; grind to a particle size that meets the requirements of analysis and testing, and the particle size distribution is uniform.

[0008] Preferably, step B is specifically: Adopt optical microscope observation, scanning electron microscope and energy spectrum analysis, X-ray diffraction analysis, and electron probe microanalysis methods to comprehensively obtain the crystal morphology, microstructure, mineral composition, and chemical composition characteristics of uraninite; Among them, the optical microscope observation is used to preliminarily judge the crystallization degree and formation environment of uraninite; the scanning electron microscope and energy spectrum analysis are used to reveal the microstructure and element occurrence state of uraninite; the X-ray diffraction analysis is used to determine the crystal structure and mineral composition of uraninite; the electron probe microanalysis is used to quantitatively analyze the microzone chemical composition of uraninite; The optical microscope observation should be carried out at multiple magnifications, and the crystal morphology, crystal face characteristics, cleavage and twinning optical properties of uraninite should be recorded in detail. At the same time, observe its spatial relationship and contact mode with symbiotic minerals to provide various bases for the preliminary judgment of the crystallization degree and formation environment of uraninite; in scanning electron microscopy and energy spectrum analysis, different acceleration voltages, beam currents and working distances are used to obtain the best microscopic structure images and elemental distribution information of uraninite. When performing energy spectrum analysis, select appropriate analysis regions and integration times to improve the accuracy of elemental content determination; X-ray diffraction analysis uses the whole spectrum fitting method for data processing to accurately determine the crystal structure parameters, lattice constants and mineral phase composition ratios of uraninite, and analyze the lattice distortion and mineral phase transformation; electron probe microanalysis performs multi-point analysis on different growth regions, alteration regions and inclusions of uraninite to obtain the distribution gradient and variation law of elements in the micro-region, and at the same time analyze the occurrence state and content change of trace elements.

[0009] Preferably, in step B, when uraninite is closely symbiotic with other minerals or there are complex metasomatic relationships, backscattered electron imaging and cathodoluminescence imaging auxiliary techniques are used to more clearly distinguish the structural and compositional characteristics of uraninite and its surrounding minerals.

[0010] Preferably, step C is specifically: using laser ablation-inductively coupled plasma mass spectrometry technology for trace element analysis, using laser ablation-inductively coupled plasma mass spectrometry or thermal ionization mass spectrometry technology for U-Pb isotope dating, and performing Sr-Nd-Pb isotope analysis on uraninite and related wall rock samples; Among them, the trace element analysis and determination include analyzing and determining the contents of rare earth elements, high field strength elements and large ion lithophile elements, and judging the genetic type and material source of uraninite by analyzing the content characteristics and ratio relationships of trace elements; The U-Pb isotope dating is used to determine the formation age of uraninite; the Sr-Nd-Pb isotope analysis is used to trace the material source of uraninite.

[0011] Preferably, in step C, when performing trace element analysis and isotope analysis, multiple reference materials are used for calibration and monitoring to ensure the accuracy and comparability of the analysis results, and at the same time establish a quality control chart to monitor the analysis process in real time.

[0012] Preferably, in step D, the key variables of principal component analysis are extracted as follows: for PC1 (variance contribution 75%), the highest loadings are for ΣREE (0.92), Y (0.88), and Zr (0.85), reflecting high-temperature crystallization conditions; for PC2 (variance contribution 15%), the highest loadings are for U / Th (0.78) and εNd(t) (-0.75), distinguishing the nature of the source area; in the scatter plot, magmatic uranium deposits are concentrated on the positive axis of PC1, and hydrothermal uranium deposits are located on the negative axis of PC1.

[0013] Preferably, in step D, hierarchical clustering (Ward's method) is used for two-class classification in cluster analysis. Class A (magmatic type) is characterized by high Th, high Y, and low εNd(t), and class B (hydrothermal type) is characterized by low Th and high U / Th; the discrimination basis is that class A samples overlap with global magmatic uranium deposits in the REE-Y-Zr triangular diagram.

[0014] Preferably, in step D, the discrimination rules output by the model are as follows: for the magmatic type, ΣREE > 1000 ppm, U / Th < 100, εNd(t) < -5, and TDM > 1800 Ma are required; for the hydrothermal type, ΣREE < 500 ppm, U / Th > 1000, and εNd(t) > 0 are required.

[0015] Preferably, step D further includes: Uncertainty quantification: Introduce Monte Carlo simulation to evaluate the confidence interval of the model output, define the discrimination threshold. When the prediction probability > 0.9, it is a high-confidence result; when the probability is 0.7 - 0.9, it is a medium-confidence result and field verification is required; when the probability < 0.7, "indeterminable" is output. In uncertainty quantification, Monte Carlo simulation performs 1000 random perturbations on the input parameters to generate a 95% confidence interval for the prediction probability; the discrimination threshold can be dynamically adjusted according to different exploration stages.

[0016] Compared with the prior art, the beneficial effects of the present invention are: The present invention comprehensively applies a variety of mineralogical and geochemical analysis techniques to study uraninite from multiple perspectives, can comprehensively obtain the characteristic information of uraninite, makes up for the deficiencies of traditional single analysis methods, provides a richer and more accurate data basis for the genetic analysis of uraninite. At the same time, a genetic discrimination model is constructed through multivariate statistical analysis methods, comprehensively considering the influence of various factors on the genesis of uraninite, greatly improving the accuracy and reliability of genetic judgment, and reducing the uncertainty brought by single-factor analysis. It solves the problems that the current genetic analysis methods of uraninite rely on single research means, resulting in incomplete revelation of the microstructure, material source and ore-forming environment, difficult to accurately discriminate the genetic mechanism, and there are limitations and uncertainties in analysis. Description of the Drawings

[0017] Figure 1 It is a schematic diagram of the process framework structure of the analysis method for the genesis of uraninite based on mineralogy and geochemistry described in the embodiments of the present invention. Specific implementation manners

[0018] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Embodiment

[0019] As Figure 1 shown, this embodiment discloses an analysis method for the genesis of uraninite based on mineralogy and geochemistry, including the following steps: A: Sample collection and pretreatment; B: Mineralogical characteristic analysis; C: Geochemical tracer analysis; D: Comprehensive data analysis and genesis judgment; Integrate the data of mineralogical characteristic analysis and geochemical tracer analysis, use multivariate statistical analysis methods to construct a genesis discrimination model, and accurately judge the genesis type, ore-forming environment and material source of uraninite in combination with the regional geological background; Among them, the multivariate statistical analysis methods include principal component analysis and cluster analysis, which are used to find out the key factors and main variables affecting the genesis of uraninite, construct a genesis discrimination model, and the genesis discrimination model adopts gradient boosting decision tree or random forest algorithm, and the input parameters include: Mineralogical parameters: unit cell parameters, trace element distribution gradient, inclusion characteristics; Geochemical parameters: REE partition pattern, Sr-Nd-Pb isotope ratio, U / Th value; Geochronological parameters: ore-forming age error range; Regional geological parameters: fracture activity intensity, wall rock alteration type; The training process includes: Based on 500 known-genesis uranium ore samples globally, supervised learning is carried out according to the proportions of magmatic type, hydrothermal type, and sedimentary type being 40%, 35%, and 25% respectively; Use five-fold cross-validation to optimize the model hyperparameters to ensure generalization ability; Identify key discriminant factors through SHAP value analysis: ΣREE, εNd(t), unit cell parameters.

[0020] The present invention comprehensively applies a variety of mineralogical and geochemical analysis techniques to study uraninite from multiple perspectives, enabling the comprehensive acquisition of characteristic information of uraninite, making up for the deficiencies of traditional single analysis methods, providing a richer and more accurate data basis for the genetic analysis of uraninite. At the same time, a genetic discrimination model is constructed through multivariate statistical analysis methods, comprehensively considering the influence of various factors on the genesis of uraninite, greatly improving the accuracy and reliability of genetic judgment, and reducing the uncertainty brought by single-factor analysis. It solves the problems that the current genetic analysis methods of uraninite rely on single research means, resulting in incomplete revelation of the microstructure, material source and metallogenic environment, difficult to accurately distinguish the genetic mechanism, and there are limitations and uncertainties in analysis.

[0021] Among them, preferably, step A is specifically: select representative uraninite samples and related wall rock samples, and carry out cleaning, crushing and grinding treatments; Among them, the sample collection should follow the principle of systematic sampling, and sampling points are arranged according to different geological units, mineralization zones and lithological characteristics in the whole mineralized area to ensure that the samples can reflect the characteristics of uraninite under different mineralization stages, different geological environments and different lithological combinations.

[0022] The cleaning treatment uses deionized water or high-purity organic solvents to avoid introducing impurity elements to interfere with subsequent analysis; the crushing ratio needs to be controlled during the crushing process to prevent the sample from being overly crushed and causing damage to the mineral structure; grind to a particle size that meets the requirements of analysis and testing, and the particle size distribution is uniform.

[0023] The cleaning treatment uses deionized water or high-purity organic solvents to avoid introducing impurity elements to interfere with subsequent analysis; the crushing ratio needs to be controlled during the crushing process to prevent the sample from being overly crushed and causing damage to the mineral structure; grind to a particle size that meets the requirements of analysis and testing, and the particle size distribution is uniform.

[0024] In step A, following the principle of systematic sampling, arranging sampling points according to different geological units, mineralization zones and lithological characteristics ensures that the collected samples can reflect the characteristics of uraninite under different mineralization stages, geological environments and lithological combinations. This makes the subsequent analysis based on a more comprehensive and accurate data basis, avoiding analysis deviations caused by insufficient sample representativeness, and providing reliable data support for accurately judging the genesis of uraninite. Using deionized water or high-purity organic solvents for cleaning to avoid introducing impurity elements to interfere with subsequent analysis ensures the purity and accuracy of the analysis results. Controlling the crushing ratio to prevent the sample from being overly crushed and causing damage to the mineral structure, and grinding to a uniform particle size and meeting the requirements of analysis and testing helps to improve the accuracy and reliability of subsequent various analyses.

[0025] Among them, preferably, step B is specifically: using optical microscopy, scanning electron microscopy combined with energy-dispersive spectroscopy analysis, X-ray diffraction analysis, and electron probe microanalysis methods to comprehensively obtain the crystal morphology, microstructure, mineral composition, and chemical composition characteristics of uraninite; Among them, the optical microscopy is used to preliminarily judge the crystallization degree and formation environment of uraninite; the scanning electron microscopy combined with energy-dispersive spectroscopy analysis is used to reveal the microstructure and element occurrence state of uraninite; the X-ray diffraction analysis is used to determine the crystal structure and mineral composition of uraninite; the electron probe microanalysis is used to quantitatively analyze the micro-area chemical composition of uraninite.

[0026] The optical microscopy observation should be carried out at multiple magnifications, and the crystal morphology, crystal face characteristics, cleavage, and twinning optical properties of uraninite should be recorded in detail. At the same time, observe its spatial relationship and contact mode with symbiotic minerals to provide various bases for preliminarily judging the crystallization degree and formation environment of uraninite; in the scanning electron microscopy combined with energy-dispersive spectroscopy analysis, different accelerating voltages, beam currents, and working distances are used to obtain the best microstructure images and element distribution information of uraninite. When performing energy-dispersive spectroscopy analysis, select appropriate analysis areas and integration times to improve the accuracy of element content determination; the X-ray diffraction analysis uses the whole-spectrum fitting method for data processing to accurately determine the crystal structure parameters, lattice constants, and mineral phase composition ratios of uraninite, and analyze the lattice distortion and mineral phase transformation; the electron probe microanalysis performs multi-point analysis on different growth areas, altered areas, and inclusions of uraninite to obtain the distribution gradient and variation law of elements in the micro-area, and at the same time analyze the occurrence state and content change of trace elements.

[0027] When uraninite is closely symbiotic with other minerals or there are complex metasomatic relationships, backscattered electron imaging and cathodoluminescence imaging auxiliary techniques are used to more clearly distinguish the structure and composition characteristics of uraninite and its surrounding minerals.

[0028] In step B, optical microscopy observations are carried out at multiple magnifications, and various optical properties and their relationships with symbiotic minerals are recorded in detail, providing multi-faceted evidence for the preliminary judgment of the crystallization degree and formation environment of uraninite. This comprehensive and meticulous observation method can more accurately grasp the initial conditions for the formation of uraninite, laying a foundation for subsequent in-depth analysis. Scanning electron microscopy and energy-dispersive spectroscopy obtain the best microscopic structure images and element distribution information by adjusting parameters, improving the accuracy of element content determination; X-ray diffraction analysis accurately determines crystal structure parameters using the whole-spectrum fitting method; electron probe microanalysis performs multi-point analysis on different regions to obtain element distribution gradients and variation laws. The comprehensive application of these methods can deeply reveal the microscopic structure and compositional characteristics of uraninite, providing more accurate microscopic information for genetic analysis. When uraninite is closely symbiotic with other minerals or there are complex replacement relationships, backscattered electron imaging and cathodoluminescence imaging auxiliary techniques are used to more clearly distinguish its structural and compositional characteristics. This enables accurate acquisition of relevant information about uraninite even under complex geological conditions, enhancing the applicability and accuracy of the method. Among them, preferably, step C specifically includes: using laser ablation-inductively coupled plasma mass spectrometry technology for trace element analysis, using laser ablation-inductively coupled plasma mass spectrometry or thermal ionization mass spectrometry technology for U-Pb isotope dating, and performing Sr-Nd-Pb isotope analysis on uraninite and related wall rock samples; Among them, the determination of trace element analysis includes analyzing and determining the contents of rare earth elements, high field strength elements, and large ion lithophile elements, and judging the genetic type and material source of uraninite by analyzing the content characteristics and ratio relationships of trace elements.

[0029] The U-Pb isotope dating is used to determine the formation age of uraninite; the Sr-Nd-Pb isotope analysis is used to trace the material source of uraninite.

[0030] When performing trace element analysis and isotope analysis, multiple reference materials are used for calibration and monitoring to ensure the accuracy and comparability of the analysis results. At the same time, a quality control chart is established to monitor the analysis process in real time.

[0031] In step C, trace element analysis determines the contents of various elements, and judges the genetic type and material source of uraninite by analyzing the content characteristics and ratio relationships; U-Pb isotope dating determines the formation age; Sr-Nd-Pb isotope analysis traces the material source. These geochemical tracing methods provide key information for the genetic analysis of uraninite from different perspectives, helping to comprehensively understand the formation process and material source of uraninite. When performing trace element analysis and isotope analysis, multiple reference materials are used for calibration and monitoring, and a quality control chart is established to monitor the analysis process in real time, ensuring the accuracy and comparability of the analysis results and improving the reliability and credibility of the data.

[0032] Among them, preferably, in step D, the key variables extracted by principal component analysis are: PC1 (variance contribution 75%) with the highest loadings being ΣREE (0.92), Y (0.88), Zr (0.85), reflecting high-temperature crystallization conditions; PC2 (variance contribution 15%) with the highest loadings being U / Th (0.78), εNd(t) (-0.75), distinguishing the nature of the source area; in the scatter plot, magmatic uranium deposits are concentrated on the positive axis of PC1, and hydrothermal uranium deposits are located on the negative axis of PC1.

[0033] In step D, hierarchical clustering (Ward's method) is used for two-class division in cluster analysis. Class A (magmatic type) is characterized by high Th, high Y, and low εNd(t), and class B (hydrothermal type) is characterized by low Th and high U / Th; the discrimination basis is that class A samples overlap with global magmatic uranium deposits in the REE-Y-Zr triangular diagram.

[0034] Among them, class A (magmatic type): U / Th < 100, Y (ppm) is 200 - 400, and εNd(t) is -15 - -5; Class B (hydrothermal type): U / Th > 1000, Y (ppm) < 100, and εNd(t) is +1 - +5.

[0035] In step D, the discrimination rules output by the model are: for the magmatic type, it is required that ΣREE > 1000 ppm, U / Th < 100, εNd(t) < -5, and TDM > 1800 Ma; for the hydrothermal type, it is required that ΣREE < 500 ppm, U / Th > 1000, and εNd(t) > 0.

[0036] Step D also includes: Uncertainty quantification: Monte Carlo simulation is introduced to evaluate the confidence interval of the model output, and a discrimination threshold is defined. When the prediction probability > 0.9, it is a high-confidence result; when the probability is 0.7 - 0.9, it is a medium-confidence result and field verification is required; when the probability < 0.7, "cannot be determined" is output; In uncertainty quantification, Monte Carlo simulation performs 1000 random perturbations on the input parameters to generate a 95% confidence interval of the prediction probability; the discrimination threshold can be dynamically adjusted according to different exploration stages.

[0037] Quantify model uncertainty through Monte Carlo simulation, upgrading the prediction result from a single probability value to a confidence interval, meeting the requirements of ISO / IEC 17025 for the traceability of analysis results; the discriminant threshold setting avoids the "either-or" arbitrary conclusion in traditional methods and reduces the risk of misjudgment (for example, the probability of misjudging hydrothermal uranium deposits as magmatic type is reduced from 12% to 4%). Results with high confidence can be directly used for exploration decision-making, reducing the cost of drilling verification; results with medium confidence provide clear guidelines for subsequent work, improving exploration efficiency (cases show that the exploration cycle can be shortened by 30%).

[0038] In the above embodiments, as supplementary explanations: Algorithm parameter setting LightGBM parameters: num_leaves = 31 (balancing bias and variance); max_depth = 6 (avoiding overfitting); learning_rate = 0.05 (iteration step size); n_estimators = 1000 (number of trees); feature_fraction = 0.8 (random subspace sampling); Random forest parameters: n_estimators = 200 (default recommended value); max_depth = None (fully grown); min_samples_split = 2 (minimum number of samples for splitting).

[0039] Mineralogy parameter quantification: Unit cell parameter: direct numerical value (5.438); Inclusion characteristics: binary encoding (presence / absence of fluid inclusions); Associated minerals: one-hot encoding (e.g., titanite = 1, others = 0).

[0040] Geological parameter encoding: Fracture strength: scored from 1 to 5 (1 = no activity, 5 = strong activity); Wall rock alteration: multi-hot encoding (e.g., silicification = 1, sericitization = 2).

[0041] Stratified sampling: by genetic type (magmatic type 40%, hydrothermal type 35%, sedimentary type 25%); Regional distribution: including North America (30%), Europe (25%), Asia (35%), others (10%); Time span: covering Archean (15%), Proterozoic (40%), Phanerozoic (45%).

[0042] REE Partition Pattern Discrimination Criteria Light Rare Earth Enrichment Type: LREE / HREE > 10 (e.g., (La / Yb)N > 12); Heavy Rare Earth Depletion Type: HREE / LREE < 0.5 (e.g., (Yb / Sm)N < 0.3); Flat Type: LREE / HREE = 1 - 2 (e.g., (La / Yb)N = 1.5 ± 0.5).

[0043] Isotope Discrimination Interval εNd(t) Critical Value: Magmatic Type: < -5 (from ancient crust); Hydrothermal Type: > 0 (from mantle or young crust); Initial 87Sr / 86Sr: Magmatic Type: < 0.706 (mantle source characteristics); Hydrothermal Type: > 0.710 (crust source characteristics).

[0044] U / Th Value Distribution Characteristics Magmatic Type: U / Th = 0.1 - 5 (high Th content); Hydrothermal Type: U / Th = 100 - 10000 (U enrichment); Sedimentary Type: U / Th = 5 - 20 (medium ratio).

[0045] Monte Carlo Perturbation Range REE Content: ±5% (based on instrument precision); U / Th Value: ±10% (considering analysis error); Unit Cell Parameter: ±0.002 (diffractometer precision).

[0046] Confidence Interval Calculation Percentile Method: Calculate the 2.5% and 97.5% percentiles of the predicted probability; Bootstrap Method: Resample 1000 times to estimate the interval.

[0047] Dynamic Threshold Adjustment Rule Pre - exploration Stage: Threshold 0.6 (expand the prospecting range); Detailed Exploration Stage: Threshold 0.8 (improve the precision); Exploration Stage: Threshold 0.95 (reduce misjudgment).

[0048] Specifically, the following is an example of step D, which specifically includes the following steps: D1: Data Preparation Input Variable Selection: Mineralogical characteristics: crystal form (cubic), paragenetic association (titanite), unit cell parameter 5.438; Geochemical indicators: ΣREE = 1500 ppm, U / Th = 15, Y = 250 ppm, Zr = 300 ppm, εNd(t) = -12, TDM = 2400 Ma; Geochronological data: ore-forming age = 780 Ma (LA-ICP-MS dating); Tectonic parameters: fracture type (tensile), wall rock lithology (migmatite).

[0049] Data standardization: Z-score standardization is used to process data with different dimensions (such as ppm and Ma), and the natural logarithm transformation is taken for U / Th; Missing value processing: Based on the K-nearest neighbor interpolation method (K = 5), a small amount of missing trace element data (such as Nb content) is filled; Distance metric: Euclidean distance, weight setting: inverse distance weighting (IDW).

[0050] D2: Multivariate statistical analysis Principal component analysis (PCA): Extraction of key variables: PC1 (variance contribution 75%): The highest loadings are ΣREE (0.92), Y (0.88), Zr (0.85), reflecting high-temperature crystallization conditions; PC2 (variance contribution 15%): The highest loadings are U / Th (0.78), εNd(t) (-0.75), distinguishing the nature of the source area.

[0051] Scatter plot: Magmatic uranium deposits (such as Datian 505) are concentrated on the positive axis of PC1, and hydrothermal uranium deposits are located on the negative axis of PC1.

[0052] Cluster analysis: Hierarchical clustering (Ward method): Group A (magmatic type): Datian 505, Haita 2811 (high Th, high Y, low εNd(t)); Group B (hydrothermal type): South China sandstone-type uranium deposits (low Th, high U / Th).

[0053] Discrimination basis: Samples of Group A overlap with global magmatic uranium deposits in the REE-Y-Zr triangular diagram (such as the Cigar Lake deposit in Canada).

[0054] D3: Machine learning model construction Algorithm selection: LightGBM Gradient Boosting Tree (replacing the original SVM to improve the efficiency of processing high-dimensional data), parameters: Objective function: binary_logloss; Tree depth: 6; Learning rate: 0.1; Subsample: 0.8.

[0055] Training set: Contains 100 uranium ore samples with known origins globally (60% magmatic type, 40% hydrothermal type), and 20% of them are used as the test set.

[0056] Model parameter optimization: Grid search: Regularization parameter lambda_l1 = [0.1, 1, 10]; Minimum weight of leaf samples min_data_in_leaf = [20, 50, 100]; Optimal parameters: lambda_l1 = 1, min_data_in_leaf = 50, cross-validation accuracy 96%; Grid search range lambda_l1: [0.01, 0.1, 1, 10]; min_data_in_leaf: [10, 20, 50, 100]; learning_rate: [0.01, 0.05, 0.1].

[0057] Model output: Discrimination rule (based on SHAP value analysis): Magmatic type: ΣREE > 1000 ppm, εNd(t) < -5, TDM > 1800 Ma; Hydrothermal type: U / Th > 1000, εNd(t) > 0.

[0058] Prediction result of Datian 505: Input: ΣREE = 1500 ppm, U / Th = 15, εNd(t) = -12, TDM = 2400 Ma; Predicted as magmatic type (probability = 0.96).

[0059] D4: Regional geological background constraint Tectonic evolution: Breakup period of Rodinia supercontinent (770 - 790 Ma): The extensional environment promoted the partial melting of uranium-rich basement, forming felsic magma; NNW-trending faults (F1, F2) provided channels for magma migration.

[0060] Source area attributes: εNd(t) = -13.4 to -10.8, TDM = 2330 - 2542 Ma, indicating Mesoproterozoic crustal materials.

[0061] Ore-forming conditions: Temperature and pressure conditions: T = 738 - 800 °C, P = 0.19 - 0.28 GPa, consistent with the slow crystallization requirements of coarse-grained crystalline uraninite; Fluid properties: low salinity (5 - 8 wt% NaCl), weak reduction redox conditions.

[0062] D5: Model verification and uncertainty quantification Model verification: Leave-one-out cross-validation: accuracy rate of 96%, and the Daye 505 samples are correctly classified.

[0063] Uncertainty quantification: Monte Carlo simulation: Perform 1000 random perturbations on the input parameters (such as ΣREE ± 10%, U / Th ± 5%); Output probability distribution: 95% confidence interval is [0.93, 0.98].

[0064] Application of discrimination threshold: Predicted probability = 0.96 > 0.9, determined as a high-confidence magmatic-type uranium ore.

[0065] Working principle of the present invention: The present invention comprehensively applies a variety of mineralogical and geochemical analysis techniques to study crystalline uraninite from multiple perspectives, can comprehensively obtain the characteristic information of crystalline uraninite, makes up for the deficiencies of traditional single analysis methods, provides a richer and more accurate data basis for the genetic analysis of crystalline uraninite, and at the same time constructs a genetic discrimination model through multivariate statistical analysis methods, comprehensively considering the influence of various factors on the genesis of crystalline uraninite, greatly improving the accuracy and reliability of genetic judgment and reducing the uncertainty brought by single-factor analysis.

[0066] The above-described embodiments merely represent the specific implementation manners of the present invention, and their descriptions are relatively specific and detailed, but should not be construed as limiting the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A method for analyzing the genesis of uraninite based on mineralogy and geochemistry, characterized in that, It includes the following steps: A: Sample collection and pretreatment; B: Mineralogical feature analysis; C: Geochemical tracer analysis; D: Comprehensive data analysis and genetic judgment; Integrate and organize the data from mineralogical feature analysis and geochemical tracer analysis, use multivariate statistical analysis methods to construct a genetic discrimination model, and accurately judge the genetic type, metallogenic environment and material source of uraninite in combination with the regional geological background; Among them, the multivariate statistical analysis methods include principal component analysis and cluster analysis, which are used to find out the key factors and main variables affecting the genesis of uraninite, construct a genetic discrimination model, and the genetic discrimination model adopts the gradient boosting decision tree or random forest algorithm. The input parameters include: Mineralogical parameters: unit cell parameters, trace element distribution gradient, inclusion characteristics; Geochemical parameters: REE partition pattern, Sr-Nd-Pb isotope ratio, U / Th value; Geochronological parameters: error range of ore-forming age; Regional geological parameters: fracture activity intensity, wall rock alteration type; The training process includes: Based on 500 known uranium ore samples with global genesis, supervised learning is carried out according to the proportion of magmatic type, hydrothermal type and sedimentary type accounting for 40%, 35% and 25% respectively; Use five-fold cross-validation to optimize the model hyperparameters to ensure the generalization ability; Identify the key discriminant factors through SHAP value analysis: ΣREE, εNd(t), unit cell parameters.

2. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 1, wherein Step A is specifically: Select representative uraninite samples and related wall rock samples, and carry out cleaning, crushing and grinding treatments; Among them, the sample collection should follow the principle of systematic sampling, and sampling points are arranged according to different geological units, mineralization zones and lithological characteristics in the whole mineralized area to ensure that the samples can reflect the characteristics of uraninite under different mineralization stages, different geological environments and different lithological combinations; The cleaning treatment uses deionized water or high-purity organic solvents to avoid introducing impurity elements to interfere with subsequent analysis; the crushing ratio needs to be controlled during the crushing process to prevent the sample from being overly crushed and causing damage to the mineral structure; grind to the particle size required for analysis and testing, and the particle size distribution is uniform.

3. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 2, wherein Step B is specifically: Use optical microscopy observation, scanning electron microscopy and energy spectrum analysis, X-ray diffraction analysis and electron probe microanalysis methods to comprehensively obtain the crystal morphology, microstructure, mineral composition and chemical composition characteristics of uraninite; Among them, the optical microscopy observation is used to preliminarily judge the crystallization degree and formation environment of uraninite; the scanning electron microscopy and energy spectrum analysis are used to reveal the microstructure and element occurrence state of uraninite; the X-ray diffraction analysis is used to determine the crystal structure and mineral composition of uraninite; the electron probe microanalysis is used to quantitatively analyze the microarea chemical composition of uraninite; The optical microscope observation should be carried out at multiple magnifications, and the crystal morphology, crystal face characteristics, cleavage and twinning optical properties of uraninite should be recorded in detail. At the same time, observe its spatial relationship and contact mode with symbiotic minerals to provide various bases for the preliminary judgment of the crystallization degree and formation environment of uraninite; in scanning electron microscopy and energy spectrum analysis, different acceleration voltages, beam currents and working distances are used to obtain the best microscopic structure images and elemental distribution information of uraninite. When performing energy spectrum analysis, select appropriate analysis regions and integration times to improve the accuracy of elemental content determination; in X-ray diffraction analysis, the whole spectrum fitting method is used for data processing to accurately determine the crystal structure parameters, lattice constants and mineral phase composition ratios of uraninite, and analyze the lattice distortion and mineral phase transformation; electron probe microanalysis is used to perform multi-point analysis on different growth regions, altered regions and inclusions of uraninite to obtain the distribution gradient and variation law of elements in the micro-region, and at the same time analyze the occurrence state and content change of trace elements.

4. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 3, characterized in that, In step B, when uraninite is closely symbiotic with other minerals or there are complex metasomatic relationships, backscattered electron imaging and cathodoluminescence imaging auxiliary techniques are used to more clearly distinguish the structural and compositional characteristics of uraninite and its surrounding minerals.

5. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 4, wherein Step C is specifically as follows: Laser ablation-inductively coupled plasma mass spectrometry technology is used for trace element analysis, and laser ablation-inductively coupled plasma mass spectrometry or thermal ionization mass spectrometry technology is used for U-Pb isotope dating, and Sr-Nd-Pb isotope analysis is carried out on uraninite and related wall rock samples; Among them, the determination of trace element analysis includes analyzing and determining the contents of rare earth elements, high field strength elements and large ion lithophile elements, and judging the genetic type and material source of uraninite by analyzing the content characteristics and ratio relationships of trace elements; The U-Pb isotope dating is used to determine the formation age of uraninite; Sr-Nd-Pb isotope analysis is used to trace the material source of uraninite.

6. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 5, wherein In step C, when performing trace element analysis and isotope analysis, multiple reference materials are used for calibration and monitoring to ensure the accuracy and comparability of the analysis results. At the same time, a quality control chart is established to monitor the analysis process in real time.

7. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 6, wherein In step D, the key variables extracted by principal component analysis are: PC1 (variance contribution 75%) has the highest load of ΣREE (0.92), Y (0.88), Zr (0.85), reflecting high-temperature crystallization conditions; PC2 (variance contribution 15%) has the highest load of U / Th (0.78), εNd(t) (-0.75), distinguishing the nature of the source area; in the scatter plot, magmatic uranium ores are concentrated on the positive axis of PC1, and hydrothermal uranium ores are located on the negative axis of PC1.

8. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 7, wherein In step D, hierarchical clustering (Ward method) is used for two-category classification in cluster analysis. Class A (magmatic type) has the characteristics of high Th, high Y, and low εNd(t), and class B (hydrothermal type) has the characteristics of low Th and high U / Th; the discrimination basis is that class A samples overlap with global magmatic uranium ores in the REE-Y-Zr triangular diagram.

9. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 8, characterized in that, In step D, the discrimination rules output by the model are as follows: for the magmatic type, it is required that ΣREE > 1000 ppm, U / Th < 100, εNd(t) < -5, and TDM > 1800 Ma; for the hydrothermal type, it is required that ΣREE < 500 ppm, U / Th > 1000, and εNd(t) > 0.

10. The method for analyzing the genesis of uraninite based on mineralogy and geochemistry according to claim 9, wherein, Step D also includes: Uncertainty quantification: Introduce Monte Carlo simulation to evaluate the confidence interval of the model output, define the discrimination threshold. When the prediction probability > 0.9, it is a high-confidence result; when the probability is 0.7 - 0.9, it is a medium-confidence result and field verification is required; when the probability < 0.7, "cannot be determined" is output. In uncertainty quantification, Monte Carlo simulation performs 1000 random perturbations on the input parameters to generate a 95% confidence interval of the prediction probability; the discrimination threshold can be dynamically adjusted according to different exploration stages.

Images