Alteration rock type gold ore rich ore section prospecting method

Abstract

The invention discloses an altered rock type gold ore rich ore section prospecting method, and belongs to the field of geological data processing, mineral resource evaluation and investment decision support, and the method comprises the steps: obtaining and fusing multi-source geological data to construct a three-dimensional seepage physical model; performing numerical simulation based on the physical model to obtain flow field parameters; calculating and generating a fluid focusing index according to the flow field parameters; performing regression calibration on the focusing index by using the drilling data to establish a grade thickness prediction model; and delineating and generating a three-dimensional rich ore exploration target area by using the prediction model according to a grade threshold value. According to the method, a technical path of combining physical process simulation and machine learning regression is adopted, and a closed-loop feedback correction mechanism connecting theoretical prediction and calibration results is constructed, so that end-to-end quantitative prediction of the hidden ore body from geological causes to commercial values can be realized, the success rate of deep prospecting is remarkably improved, and the exploration investment risk is reduced.

Description

Technical Field

[0001] This invention relates to the fields of geological data processing, mineral resource evaluation and investment decision support, and in particular to a method for prospecting rich sections of altered rock-type gold deposits. Background Technology

[0002] Altered rock gold deposits are an important type of gold resource globally, and the exploration of deep, concealed rich ore sections is both a key focus and a challenge in current mineral exploration work. Traditional prospecting typically relies on direct exploration methods such as geological mapping, drilling and sampling, and chemical analysis. By obtaining point and line geological data within a small area, these methods provide a basic basis for geological experts to delineate prospecting target areas, playing a crucial role in mineral resource discovery.

[0003] With the development of exploration technology, the amount of geological, geophysical, and geochemical data acquired is increasing dramatically. However, existing technologies still have shortcomings in the deep integration and comprehensive interpretation of multi-source heterogeneous data. In addition, the construction of mineral exploration prediction models largely relies on the static experience of geological experts, lacking dynamic simulation of key physicochemical processes such as the migration of ore-forming fluids, and failing to integrate the assessment of geological mineralization potential with the economic benefits of mining, resulting in high risks in mineral exploration decisions. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides a prospecting method for rich sections of altered rock-type gold deposits. It employs a technical approach that combines physical process simulation with machine learning regression and constructs a closed-loop feedback correction mechanism that connects theoretical predictions with calibration results. This enables end-to-end quantitative prediction of concealed ore bodies from their geological origins to their commercial value, significantly improving the success rate of deep prospecting and reducing exploration investment risks.

[0005] The above objectives can be achieved through the following approach:

[0006] A method for prospecting rich sections of altered rock-type gold deposits includes acquiring and integrating geological, geophysical, and core data of the exploration area to construct a three-dimensional seepage physical model; performing numerical simulation based on the three-dimensional seepage physical model to calculate flow field parameters; calculating the fluid pressure and velocity distribution in the flow field parameters to construct a fluid focusing index; performing regression calibration on the relationship between the fluid focusing index and the actual mineralization enrichment degree to establish a grade-thickness prediction model; and delineating the rich ore exploration target area based on the grade-thickness prediction model and preset first, second, and third grade thresholds.

[0007] Optionally, the construction of the three-dimensional seepage physical model includes: spatial interpolation and three-dimensional modeling of the geological, geophysical and core data of the exploration area to reproduce the spatial distribution of alteration zoning and generate a three-dimensional geological structure and alteration zoning; assigning permeability coefficient values ​​to different regions in the three-dimensional geological structure and alteration zoning to construct a three-dimensional seepage physical model.

[0008] Optionally, assigning permeability coefficient values ​​to different regions within the three-dimensional geological structure and alteration zones includes: matching and assigning initial permeability values ​​from a preset rock property database based on the rock type and alteration intensity of the three-dimensional geological structure and alteration zones to generate a baseline permeability spatial distribution; performing geometric analysis on the three-dimensional geological structure and alteration zones to generate a structural complexity correction factor; and applying the structural complexity correction factor to the baseline permeability spatial distribution for weighted adjustment to generate permeability coefficient values.

[0009] Optionally, the calculated flow field parameters include: setting boundary conditions in the three-dimensional seepage physical model, performing dynamic calculations to generate a three-dimensional fluid pressure field distribution; simultaneously calculating the three-dimensional flow velocity of the fluid in the three-dimensional seepage physical model to obtain a three-dimensional fluid velocity field distribution; and combining the three-dimensional fluid pressure field distribution with the three-dimensional fluid velocity field distribution to generate flow field parameters.

[0010] Optionally, combining the three-dimensional fluid pressure field distribution with the three-dimensional fluid velocity field distribution includes: calculating the curl of the three-dimensional fluid velocity field distribution to obtain a three-dimensional fluid vortex field; calculating the gradient of the three-dimensional fluid pressure field distribution and extracting the modulus to generate a pressure gradient scalar field; and aggregating the three-dimensional fluid vortex field and the pressure gradient scalar field to generate flow field parameters.

[0011] Optionally, the process of constructing the fluid focusing index includes: weighting and summing the three-dimensional fluid vorticity field and the pressure gradient scalar field to calculate a fluid dynamics favorability index; assigning geochemical acceptability scores to spatial locations based on the three-dimensional geological structure and alteration zoning to generate a geochemical favorability index; and performing calculations between the fluid dynamics favorability index and the geochemical favorability index to construct the fluid focusing index.

[0012] Optionally, the geochemical favorability index includes: rasterizing the three-dimensional geological structure and alteration zoning, quantifying the rock type and local thickness of each raster unit to generate a lithology-structure baseline favorability distribution map; extracting the geochemical data corresponding to the raster unit and quantifying it into a geochemical correction coefficient; and using the geochemical correction coefficient to adjust the lithology-structure baseline favorability distribution map grid by grid to calculate and generate the geochemical favorability index.

[0013] Optionally, establishing the grade-thickness prediction model includes: acquiring borehole data from known mineral deposits, extracting the fluid focusing index, ore grade, and ore thickness, and generating a training sample set; using the fluid focusing index as input features and the ore grade as output label, training a machine learning regression model using the training sample set to obtain an ore grade prediction sub-model; using the fluid focusing index and the actual ore grade as input features and the ore thickness as output label, training a machine learning regression model using the training sample set to obtain an ore thickness prediction sub-model; and cascading the ore grade prediction sub-model and the ore thickness prediction sub-model to construct a grade-thickness prediction model.

[0014] Optionally, generating a three-dimensional rich mineral exploration target area includes: applying the grade-thickness prediction model to a three-dimensional spatial grid of the exploration area to calculate and generate a three-dimensional attribute volume; performing preliminary screening on the three-dimensional attribute volume to delineate a three-dimensional point cloud of potential mineralization bodies; performing three-dimensional connected component analysis on the three-dimensional point cloud of potential mineralization bodies to identify and segment candidate ore bodies; calculating the average grade of the candidate ore bodies and classifying them according to the first, second, and third grade thresholds to generate a three-dimensional rich mineral exploration target area.

[0015] Optionally, the method further includes: performing a comparative analysis of the spatial distribution of the fluid focusing index and the spatial distribution of the three-dimensional rich ore exploration target area, calculating and generating a spatial deviation residual map; and performing closed-loop feedback correction on the permeability coefficient value based on the spatial deviation residual map.

[0016] Compared with the prior art, the present invention has the following advantages:

[0017] 1. This invention, by constructing a three-dimensional seepage physical model and simulating the dynamic process of ore-forming fluids, transforms mineral exploration methods from relying on subjective analogies of static characteristics by geological experts to dynamic process deduction based on physical laws. This approach can reveal hidden mineralization zones that are difficult to discover using traditional methods, significantly improving the scientific rigor and depth of mineral exploration prediction.

[0018] 2. This invention creatively constructs a fluid focusing index, coupling fluid dynamics favorability with geochemical favorability. This multi-factor fusion quantitative assessment method overcomes the one-sidedness of a single prospecting indicator, and can more accurately locate enrichment areas with favorable physical and chemical conditions, significantly improving the accuracy of target area delineation;

[0019] 3. This invention generates a spatial deviation residual map by comparing the theoretical predictions from physical simulations with the actual results calibrated from borehole data. This residual map is then used to perform closed-loop feedback correction on the initial seepage physical model. This mechanism enables the mineral exploration model to learn and evolve as exploration data increases, achieving continuous iterative improvement in prediction accuracy.

[0020] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 This is a schematic flowchart of a prospecting method for rich sections of altered rock-type gold deposits according to an embodiment of the present invention.

[0023] Figure 2 This is a coupling diagram of the permeability field and fluid transport path in an embodiment of the present invention.

[0024] Figure 3 This is a coupled analysis diagram of the three-dimensional isosurface and slice of the fluid focusing index according to an embodiment of the present invention.

[0025] Figure 4 This is a smooth surface rendering of a three-dimensional rich ore target area according to an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Reference Figure 1 One embodiment of the present invention proposes a prospecting method for rich sections of altered rock-type gold deposits. It adopts a technical approach that combines physical process simulation with machine learning regression and constructs a closed-loop feedback correction mechanism that connects theoretical predictions and calibration results. This enables end-to-end quantitative prediction of concealed ore bodies from geological genesis to commercial value, significantly improving the success rate of deep prospecting and reducing exploration investment risks.

[0028] The system described in this embodiment specifically includes:

[0029] Acquire and integrate geological, geophysical and core data from the exploration area to construct a three-dimensional seepage physical model;

[0030] Based on the aforementioned three-dimensional seepage physical model, numerical simulations were performed to calculate the flow field parameters.

[0031] The fluid pressure and velocity distribution in the flow field parameters are calculated to construct a fluid focusing index;

[0032] Regression calibration was performed on the relationship between the fluid focusing index and the actual mineralization enrichment degree to establish a grade-thickness prediction model;

[0033] Based on the grade-thickness prediction model, and according to the preset first, second and third grade thresholds, a three-dimensional rich ore exploration target area is generated.

[0034] By adopting a technical approach that combines physical process simulation with machine learning regression, and constructing a closed-loop feedback correction mechanism that connects theoretical predictions with calibration results, we can achieve end-to-end quantitative prediction of concealed ore bodies from their geological genesis to their commercial value, significantly improving the success rate of deep mineral exploration and reducing exploration investment risks.

[0035] Optionally, the construction of the three-dimensional seepage physical model includes:

[0036] Spatial interpolation and three-dimensional modeling are performed on the geological, geophysical and core data of the exploration area to reproduce the spatial distribution of alteration zoning and generate three-dimensional geological structure and alteration zoning.

[0037] Specifically, in the spatial interpolation and 3D modeling of geological, geophysical, and core data to generate 3D geological structures and alteration zoning, a geological modeling unit is responsible for processing and fusing multi-source heterogeneous data. This unit can run on professional 3D geological modeling software platforms, such as GOCAD or Micromine. It receives the 3D coordinates of lithological contact points from the input borehole core data and combines them with the structural morphology revealed by geophysical exploration data. By employing an implicit modeling algorithm, such as an interpolation method based on radial basis functions (RBF), it fits these discrete point, line, and surface data to generate a spatially continuous and smooth 3D geological surface. Its core can be expressed by the following formula:

[0038] ,

[0039] in, Represents the boundary surface of the generated geological body; It is any point in three-dimensional space; It is a spatial scalar potential function, and its zero-value surface is the surface we are looking for; It is the position vector of the i-th known geological data point; These are weighting coefficients calculated based on data point constraints. It is a radial basis function, such as a Gaussian function or a thin-plate spline function; Point With data points The Euclidean distance between them. Through this step, a three-dimensional geological structure and alteration zoning model that can accurately reproduce the spatial location, occurrence, and morphology of the ore-controlling fault and the core alteration zone is generated.

[0040] Permeability coefficient values ​​are assigned to different regions in the three-dimensional geological structure and alteration zone to construct a three-dimensional seepage physical model.

[0041] Specifically, in the step of assigning permeability coefficient values ​​to different regions within the three-dimensional geological structure and alteration zones to construct a three-dimensional seepage physical model, a parameterized unit is responsible for transforming the geometric model generated in the previous step into a seepage model with physical properties. This unit executes a permeability assignment method based on a lithology-structure dual-factor coupling. First, according to the rock type and alteration intensity of different alteration zones, such as pyrite-sericite breccia and pyrite-sericite granitic breccia, a basic permeability value is matched for each geological body from a pre-defined rock property database. Then, by analyzing the geometric morphology of the three-dimensional geological structure model, indices such as fault intersection density and fracture zone width are calculated to generate a structural complexity correction factor. Finally, the permeability coefficient value of each spatial grid cell is determined by a weighted adjustment of its basic permeability and structural complexity correction factor, the calculation process of which can be illustrated by the following formula example:

[0042] ,

[0043] in, It is the final permeability coefficient value assigned to this grid cell; It is a baseline permeability obtained by querying a database based on rock type; It is the construction complexity correction factor after normalization at this position; This involves constructing the influence weighting coefficients. Through this step, a three-dimensional seepage physical model that reflects the differences in the conductivity of different geological bodies is finally constructed.

[0044] For example, in a target exploration area, a geological modeling engineer first imports core logging data from 100 boreholes and 20 two-dimensional seismic exploration profiles into a geological modeling software. The software invokes an implicit modeling algorithm to automatically generate a three-dimensional surface of the ore-controlling fault F1, and delineates a zonal pyrite-sericite alteration zone centered on this fault and a surrounding potassic alteration zone, thus completing the construction of a three-dimensional geological structure and alteration zoning model. Subsequently, a parametric cell is activated, assigning a higher baseline permeability to the pyrite-sericite alteration zone and a lower baseline permeability to the potassic alteration zone based on a rock property database. Simultaneously, by calculating the curvature and thickness of the fault, a higher structural complexity correction factor is set for the expanded and tortuous parts of the fault zone. Finally, the permeability coefficient value of each three-dimensional grid cell is determined by weighting its baseline permeability with the structural complexity correction factor of its location, thereby generating a three-dimensional seepage physical model that can be used for subsequent fluid simulation.

[0045] Optionally, assigning permeability coefficient values ​​to different regions within the three-dimensional geological structure and alteration zones includes:

[0046] Based on the three-dimensional geological structure and the rock type and alteration intensity of the alteration zone, an initial permeability value is assigned by matching from a preset rock property database to generate a baseline permeability spatial distribution.

[0047] Specifically, in the step of generating a baseline permeability spatial distribution based on rock type and alteration intensity, a parameterized unit is responsible for assigning an initial permeability value to each spatial grid cell in the three-dimensional geological structure and alteration zoning model. This unit first identifies the rock type of each grid cell, for example, whether it is "pyrite-sericite-altered breccia" or "pyrite-sericite-altered granite." Subsequently, the unit performs a query and matching process in a pre-defined rock property database. This database pre-stores empirical permeability values ​​for different altered rock types at typical alteration intensities, obtained through laboratory core testing or literature research. Through this matching process, each grid cell is assigned an initial permeability value determined solely by its own petrological characteristics; all these values ​​together constitute a three-dimensional baseline permeability spatial distribution.

[0048] Geometric analysis was performed on the three-dimensional geological structure and alteration zoning to generate a structural complexity correction factor;

[0049] Specifically, in the step of performing geometric analysis on the three-dimensional geological structure and alteration zoning to generate a structural complexity correction factor, a geometric analysis module is responsible for quantifying the degree of structural development at each spatial location. According to mineral exploration theory, structural intersections and wide fracture zones typically have higher connectivity and permeability. This module extracts two key geometric indicators by calculating the three-dimensional geological structure model: one is the fault intersection density, which characterizes the density of fault structures near a point; the other is the fracture zone thickness. By normalizing and weighting these two indicators, a comprehensive structural complexity correction factor can be obtained, as shown in the formula:

[0050] ,

[0051] in, It is the construction complexity correction factor for the generated structure; and These are the preset weights for fracture density and fracture zone thickness, respectively; Normalized fracture intersection density value; It is the normalized value of the fracture zone thickness.

[0052] The construction complexity correction factor is applied to the baseline permeability spatial distribution for weighted adjustment to generate permeability coefficient values.

[0053] Specifically, in the step of applying a structural complexity correction factor to the baseline permeability spatial distribution for weighted adjustment to generate permeability coefficient values, a fusion computational unit couples the outputs of the first two steps. This step aims to simulate the enhancing effect of structural development on the inherent permeability of rocks. For each grid cell in the 3D model, its final permeability coefficient value is determined by both the baseline permeability determined by its lithology and the structural complexity of its location. An exemplary weighted adjustment algorithm is shown in the formula:

[0054] ,

[0055] in, This is the final permeability coefficient value generated by this grid cell; This is the baseline spatial distribution value of the permeability of this unit; It is the construction complexity correction factor corresponding to this unit; It is a preset construction influence coefficient used to adjust the extent to which construction complexity enhances penetration.

[0056] For example, in a three-dimensional geological model, there are two spatial locations, A and B, both belonging to pyrite-sericite breccia and therefore assigned the same high baseline permeability. However, location A is situated at the intersection of two secondary faults with a wider fracture zone, and a geometric analysis module calculates its structural complexity correction factor to be 0.9. Location B, on the other hand, is located within a single, straight fault zone, with a structural complexity correction factor of only 0.2. In a fusion calculation unit, the final permeability coefficient value for location A is significantly increased, while the increase for location B is much smaller. In this way, the present invention generates a heterogeneous permeability field that can accurately reflect the dual control of lithology and structure, providing a highly realistic physical basis for subsequent fluid dynamics simulations, such as... Figure 2 The diagram shown is a schematic representation of the coupling between the permeability field distribution, the high-permeability fracture zone, and the fluid transport streamlines in an embodiment of the present invention. It illustrates the control effect of the high-permeability fracture zone on the fluid transport path in a heterogeneous permeability field, where the fluid preferentially migrates along the high-permeability channel.

[0057] Optionally, the calculated flow field parameters include:

[0058] Boundary conditions are set in the three-dimensional seepage physical model, dynamic calculations are performed, and a three-dimensional fluid pressure field distribution is generated.

[0059] Specifically, in the step of setting boundary conditions and performing dynamic calculations to generate a three-dimensional fluid pressure field distribution, a simulation setting unit is responsible for defining physical constraints for the fluid simulation. This unit sets a constant high-temperature, high-pressure fluid input boundary deep within the three-dimensional seepage physical model to simulate a ore-forming hydrothermal source from the mantle or deep magma chamber. Simultaneously, a lower-pressure free outflow boundary is set at the top of the model. After setting, a computational fluid dynamics (CFD) solver, such as one based on the finite element method (FEM) or the finite volume method (FVM), solves for the pressure distribution throughout the model domain according to Darcy's law. Darcy's law describes the relationship between the seepage velocity and the pressure gradient of a fluid in a porous medium, as shown in the formula:

[0060] ,

[0061] in, It is the fluid flux or Darcy velocity vector; It is the permeability tensor of the rock, and its value comes from the generated permeability coefficient value; It is the dynamic viscosity coefficient of the fluid, which is related to the fluid temperature and composition; This is the pressure gradient. By solving this equation, a three-dimensional fluid pressure field distribution covering the entire model space is generated.

[0062] Simultaneously calculate the three-dimensional flow velocity of the fluid in the three-dimensional seepage physical model to obtain the three-dimensional fluid velocity field distribution;

[0063] Specifically, in the step of simultaneously calculating the three-dimensional flow velocity to obtain the three-dimensional fluid velocity field distribution, this step is coupled with and completed synchronously with the previous step of generating the pressure field. After solving for the pressure field, the CFD solver immediately calculates the specific fluid flux vector for each grid cell according to the formula. The magnitude and direction of this vector represent the three-dimensional flow velocity at that point. The set of velocity vectors from all grid cells constitutes the three-dimensional fluid velocity field distribution. This velocity field visually demonstrates the migration path and speed of ore-forming fluids in complex underground fracture channels.

[0064] The flow field parameters are generated by combining the three-dimensional fluid pressure field distribution with the three-dimensional fluid velocity field distribution.

[0065] Specifically, in the step of combining the three-dimensional fluid pressure field distribution and the three-dimensional fluid velocity field distribution to generate flow field parameters, a data integration module is responsible for packaging the results calculated in the first two steps. This module integrates the three-dimensional scalar field data representing pressure magnitude with the three-dimensional vector field data representing velocity magnitude and direction into a unified, structured dataset. This dataset is the final generated flow field parameters, and its data structure can be a composite data object containing two main keys, "Pressure" and "Velocity". This complete flow field parameter provides a direct data foundation for further calculations of derived parameters such as vorticity and gradient.

[0066] For example, after establishing a three-dimensional seepage physical model, a simulation setting unit set a constant pressure boundary of 50 MPa at its bottom -3000 meters and a free outflow boundary equivalent to hydrostatic pressure at its top 0 meters. A CFD solver started the calculation and converged after several hours of iterative solving. The result is two three-dimensional data volumes: the first data volume stores the pressure value of each grid cell in the model, forming a three-dimensional fluid pressure field distribution; the second data volume stores the velocity vector (vx, vy, vz) of each grid cell, forming a three-dimensional fluid velocity field distribution. The results show that the fluid mainly moves upward at high speed along the fractured zone with higher permeability. These two data volumes are combined into a single file by a data integration module and output as flow field parameters.

[0067] Optionally, combining the three-dimensional fluid pressure field distribution with the three-dimensional fluid velocity field distribution includes:

[0068] The curl of the three-dimensional fluid velocity field distribution is calculated to obtain the three-dimensional fluid vortex field;

[0069] Specifically, in the step of calculating the curl of the three-dimensional fluid velocity field distribution to obtain the three-dimensional fluid vorticity field, this step aims to quantify the local rotational or shear intensity of the fluid during its migration in a porous medium. By applying the curl operator to the obtained three-dimensional fluid velocity field distribution, the vorticity vector of each spatial grid cell can be calculated. Regions with higher vorticity values ​​in the vorticity field typically correspond to areas with complex geological formations that cause abrupt changes in the fluid path, such as bends or intersections of fractures. These regions are often favorable spaces where fluid physicochemical conditions change abruptly, promoting mineral precipitation. The calculation of a vorticity vector can be expressed by the formula:

[0070] ,

[0071] in, It is the vorticity vector at any point in the calculated three-dimensional fluid vorticity field; It is the curl operator; It is the fluid velocity vector at that point.

[0072] The gradient of the three-dimensional fluid pressure field distribution is calculated, and the modulus is extracted to generate a pressure gradient scalar field.

[0073] Specifically, in the step of calculating the gradient of the three-dimensional fluid pressure field distribution and extracting the magnitude to generate a pressure gradient scalar field, the purpose of this step is to quantify the dynamic gradient driving fluid migration. First, spatial gradient calculation is performed on the obtained three-dimensional fluid pressure field distribution to obtain a pressure gradient vector field, where each vector points in the direction of the fastest pressure increase. Then, the magnitude of each vector in this vector field is extracted, thus generating a pressure gradient scalar field. Regions with higher values ​​in this scalar field represent areas with the largest fluid pressure difference and the strongest driving force, which are also key areas where mineralization may occur. The calculation process for a pressure gradient scalar value is shown in the formula:

[0074] ,

[0075] in, It is the value at any point in the pressure gradient scalar field; It is the pressure gradient vector at that point; , , They are pressure Partial derivatives in the directions of the three coordinate axes.

[0076] The three-dimensional fluid vortex field and the pressure gradient scalar field are aggregated to generate flow field parameters.

[0077] Specifically, in the step of aggregating the three-dimensional fluid vorticity field and the pressure gradient scalar field to generate flow field parameters, this step is responsible for integrating the derived data, which better reflects the mineralization potential, obtained from the first two steps, with the original flow field data. In this step, the calculated vorticity field and pressure gradient scalar field, along with the already obtained basic pressure field and velocity field, are packaged into a structured, multi-attribute composite dataset. This composite dataset is the final, refined flow field parameter, providing richer and more insightful input information for subsequent calculations of the fluid focusing index.

[0078] For example, after a numerical simulation, a set of basic pressure and velocity field data is obtained. A subsequent calculation process is initiated. First, the curl of the velocity field is calculated according to the formula. The result shows that the vorticity value is abnormally high at a fracture corner, indicating that the fluid is rotating violently at this point. Next, the gradient of the pressure field is calculated, and it is found that the pressure gradient modulus also reaches a peak in the same region, indicating a huge pressure difference. Finally, these two newly generated vorticity field and pressure gradient field data are appended to the original pressure and velocity field data to form a more comprehensive flow field parameter set containing four data layers, which can be used for the next step of mineralization favorability analysis.

[0079] The constructed fluid focusing index includes:

[0080] The three-dimensional fluid vorticity field and the pressure gradient scalar field are weighted and summed to generate a fluid dynamics favorability index.

[0081] Specifically, in the step of weighted summation of the three-dimensional fluid vorticity field and pressure gradient scalar field to calculate the hydrodynamic favorability index, this step aims to comprehensively evaluate the favorable influence of hydrodynamic conditions on mineralization. First, the obtained three-dimensional data volumes of the vorticity field and pressure gradient scalar field are normalized, mapping their values ​​to a unified range of 0 to 1. Then, the comprehensive hydrodynamic favorability index is calculated by weighted summation of these two normalized favorability indices. The calculation process is shown in the formula:

[0082] ,

[0083] in, It is a calculated fluid dynamics favorability index, which is a three-dimensional scalar field; and These are the preset weighting coefficients for the vorticity field and the pressure gradient field, reflecting the relative importance of their contribution to mineralization as believed by geological experts. It is the normalized vorticity field; It is the normalized pressure gradient scalar field.

[0084] Based on the aforementioned three-dimensional geological structure and alteration zoning, geochemical acceptability scores are assigned to spatial locations to generate a geochemical favorability index.

[0085] Specifically, in the step of assigning geochemical acceptability scores to spatial locations based on three-dimensional geological structure and alteration zoning to generate a geochemical favorability index, this step aims to quantify the chemical capacity of different rocks as "host rocks" to accept and precipitate minerals. This process first quantifies a lithological-structural baseline favorability based on the rock type of each spatial grid cell, such as "pyrite-sericite-quartzized breccia" or "potassic granite," and the local thickness of its alteration zone. Next, indicator element geochemical data for that location are extracted as correction coefficients to dynamically adjust the baseline favorability, ultimately calculating and generating the geochemical favorability index.

[0086] The fluid focusing index is constructed by calculating the fluid dynamics favorability index and the geochemical favorability index.

[0087] Specifically, in the step of constructing the fluid focusing index by calculating the hydrodynamic favorability index and the geochemical favorability index, this step couples physical and chemical conditions to obtain a final comprehensive evaluation of mineralization potential that reflects both favorable timing and location. The final fluid focusing index is constructed by multiplying the hydrodynamic favorability index and the geochemical favorability index generated in the first two steps on each spatial grid cell. The calculation process is shown in the formula:

[0088] ,

[0089] in, It is the final fluid focusing index; It is a fluid dynamics advantage index; This is a geochemical favorability index. The significance of using a product operation is that a high focusing index only occurs when both hydrodynamic and geochemical conditions are favorable. Failure to meet either condition will significantly reduce the final mineralization potential assessment. This aligns with the geological principle that mineral precipitation requires both physical and chemical conditions to occur. Figure 3 The diagram shown is a schematic diagram of the coupling analysis of the three-dimensional isosurface of the fluid focusing index and the horizontal slice cloud map in an embodiment of the present invention. The three-dimensional nested isosurface displays mineralization potential areas at different levels, while the superimposed horizontal slice cloud map reveals the spatial distribution pattern of the deep fluid focusing index, providing accurate target area positioning for three-dimensional mineral exploration.

[0090] For example, at point A in the 3D model, due to its location at a sharp turning point in the fault structure, the calculated hydrodynamic favorability index is as high as 0.9. However, the rock at point A is the surrounding potassic granite, and its geochemical favorability index, calculated based on factors such as lithology and thickness, is only 0.2. Therefore, the final fluid focusing index for point A, calculated according to the formula, is 0.9 * 0.2 = 0.18. Meanwhile, at point B, the hydrodynamic conditions are generally average, with a favorability index of only 0.5. However, this point is located within a thick pyrite-sericite breccia zone, which serves as the main host ore-bearing lithology, and its geochemical favorability index is as high as 0.95. Therefore, the final fluid focusing index for point B is 0.5 * 0.95 = 0.475. Although the hydrophysical conditions at point B are not as good as at point A, its superior chemical and physical host environment ultimately leads the model to evaluate it as a location with greater mineralization potential.

[0091] Optionally, the geochemical favorability index includes:

[0092] The three-dimensional geological structure and alteration zoning are rasterized, and the rock type and local thickness of each raster unit are quantified to generate a lithology-structure baseline favorability distribution map.

[0093] Specifically, in the step of rasterizing and quantifying the three-dimensional geological structure and alteration zoning to generate a lithology-structure baseline favorability distribution map, this step aims to quantify the static favorability of the rock body as a mineralization space. First, the generated three-dimensional geological structure and alteration zoning model is rasterized. For each raster cell, its lithology-structure baseline favorability is calculated using a preset scoring function based on its rock type and the local thickness of its alteration zone. This function assigns different base scores to different altered rock types. For example, according to references, the "pyrite-sericite-quartz altered breccia" as the core host rock receives the highest score, while the surrounding "pyrite-sericite-quartz altered granite" receives a lower score. Simultaneously, the function also uses the alteration zone thickness at that location as a gain term. An exemplary calculation function is shown in the formula:

[0094] ,

[0095] in, It is the lithological-structural benchmark favorability of this grid unit; It is a basic score obtained by querying the database based on the rock type; Thickness affects the weight; This represents the normalized local thickness of the alteration zone at that location. Performing this calculation on all grid cells generates a lithological-structural baseline favorability map.

[0096] Extract the geochemical data corresponding to the raster unit and quantify it into geochemical correction coefficients;

[0097] Specifically, in the step of extracting geochemical data corresponding to raster cells and quantifying it into geochemical correction coefficients, this step aims to quantify the chemical "footprint" left by ore-forming fluid activity. Geochemical data of indicator elements corresponding to the spatial location of each raster cell are extracted from multi-source geological data. These indicator elements typically include gold (Au), silver (Ag), arsenic (As), and antimony (Sb). By normalizing the concentration values ​​of these elements and assigning different weights based on their indicative significance in the primary halo zoning theory, a comprehensive geochemical correction coefficient can be calculated. An exemplary calculation function is shown in the formula:

[0098] ,

[0099] in, It is the generated geochemical correction factor; It is the number of selected indicator elements; It is the preset weight of the i-th indicator element. For example, arsenic (As) and antimony (Sb), which are near-ore indicator elements, may have a higher weight. It is the normalized concentration value of the i-th indicator element in this grid cell.

[0100] The geochemical correction coefficient is used to adjust the lithology-structure baseline favorability distribution map grid by grid to calculate and generate the geochemical favorability index.

[0101] Specifically, in the step of adjusting the lithology-structure baseline favorability distribution map using geochemical correction coefficients to calculate the geochemical favorability index, this step aims to couple the static favorability of rocks with the dynamic chemical imprint of fluids. The geochemical correction coefficients generated in the previous step are used as gain adjustment terms and applied to the lithology-structure baseline favorability distribution map generated in the first step for grid-by-grid dynamic adjustment. The calculation process can be illustrated by the following formula:

[0102] ,

[0103] in, It is the final geochemical favorability index; It is the lithological-structural benchmark favorability of this grid unit; This is the geochemical correction factor corresponding to the grid cell. The geochemical favorability index generated in this step comprehensively reflects the ore-hosting "hardware conditions" and ore-forming "software environment" of a spatial location.

[0104] For example, in a 3D model, a raster cell was identified as a 30-meter-thick pyrite-sericite breccia, exhibiting excellent rock type and thickness conditions, resulting in a calculated lithology-structure baseline favorability of 0.9. Simultaneously, analysis of the cell's geochemical data revealed extremely high concentrations of the indicator elements arsenic (As) and antimony (Sb), leading to a calculated geochemical correction factor of 0.8. Ultimately, the geochemical favorability index for this raster cell was calculated using the formula 0.9*(1+0.8)=1.62, a very high favorability value. In contrast, another cell with the same lithology and thickness conditions, but with normal geochemical data and a correction factor of only 0.1, would have a final favorability index of only 0.9*(1+0.1)=0.99.

[0105] Optionally, establishing the grade-thickness prediction model includes:

[0106] Obtain borehole data from known mineral deposits, extract the fluid focusing index, ore grade, and ore thickness, and generate a training sample set;

[0107] Specifically, in the step of acquiring borehole data and generating a training sample set, this step aims to prepare "standard answers" for the subsequent supervised learning process. First, core analysis data from all historical boreholes within an exploration area are collected. This data precisely records the actual ore grade and thickness at specific three-dimensional spatial coordinates. Then, these coordinates are spatially matched with a three-dimensional fluid focusing index distribution map to extract the fluid focusing index value corresponding to each known ore point location. By combining this fluid focusing index value, the actual ore grade, and the actual ore thickness into a data tuple, the collection of all these tuples constitutes the training sample set used for model training.

[0108] Using the fluid focusing index as the input feature and the ore body grade as the output label, a machine learning regression model is trained using the training sample set to obtain an ore body grade prediction sub-model.

[0109] Specifically, in the step of training the ore body grade prediction sub-model, the purpose of this step is to establish a nonlinear mapping relationship between the physical simulation results and the degree of mineralization enrichment. A machine learning regression model, such as a random forest or gradient boosting decision tree (GBDT), is used for model training. During this process, the fluid focusing index in the training sample set is used as the model's input feature, and the corresponding actual ore body grade is used as the desired output label. Through optimization algorithms, the internal parameters of the model are continuously adjusted until the error between the model's prediction results for the training samples and the true grade label is minimized. After training, the solidified model is the ore body grade prediction sub-model, and its function can be expressed by the formula:

[0110] ,

[0111] in, It is the grade of the ore body predicted by the model; It is a trained sub-model for predicting ore body grade; It is the input fluid focusing index.

[0112] Using the fluid focusing index and the actual ore body grade as input features, and the ore body thickness as the output label, a machine learning regression model is trained using the training sample set to obtain an ore body thickness prediction sub-model.

[0113] Specifically, in the training step of the orebody thickness prediction sub-model, a feature enhancement training strategy was adopted to reflect the potential geological correlation between grade and thickness. The input feature for this step is a two-dimensional vector that includes not only the fluid focusing index but also the actual orebody grade at that location. The training process also employs a machine learning regression model, using this two-dimensional feature vector as input and the actual orebody thickness corresponding to the training sample set as the desired output label. The trained orebody thickness prediction sub-model, because it considers grade information, can more accurately capture the patterns of thickness variation; its function can be expressed by the formula:

[0114] ,

[0115] in, It is the thickness of the ore body predicted by the model; It is a trained sub-model for predicting ore body thickness; and These are two features that are used as input together.

[0116] The ore body grade prediction sub-model and the ore body thickness prediction sub-model are cascaded and combined to construct a grade-thickness prediction model.

[0117] Specifically, in the step of cascading and combining two sub-models to construct a grade-thickness prediction model, this step defines the collaborative workflow of the two sub-models in the actual prediction stage. When it is necessary to predict an unknown point, the ore body grade prediction sub-model is first invoked to predict the ore body grade based on the fluid focusing index of that point. Then, the fluid focusing index of that point and the grade value predicted in the first step are used as inputs to invoke the ore body thickness prediction sub-model to predict the ore body thickness at that point. This two-step, sequential combination method of first predicting the grade and then using the grade to predict the thickness is called cascading combination. The final grade-thickness prediction model is a complete prediction workflow that includes two sub-models and the above-described invocation logic.

[0118] For example, a geological engineer compiles thousands of data points from 100 existing boreholes to form a training sample set. First, this data is used to train a sub-model for predicting ore body grade. Then, using the "fluid focusing index" and "actual grade" as features and "actual thickness" as a label, a second sub-model for predicting ore body thickness is trained. Now, a new unknown point with a fluid focusing index of 0.85 needs to be predicted. First, 0.85 is input into the grade prediction sub-model, resulting in a predicted grade of 8.2 g / t. Next, the data pair (0.85, 8.2) is input into the thickness prediction sub-model, resulting in a predicted thickness of 15.6 meters. Finally, the prediction result for this unknown point is determined as: grade 8.2 g / t, thickness 15.6 meters.

[0119] Optionally, the generation of the three-dimensional rich mineral exploration target area includes:

[0120] The grade-thickness prediction model is applied to the three-dimensional spatial grid of the exploration area to calculate and generate a three-dimensional attribute volume.

[0121] Specifically, in the step of applying the grade-thickness prediction model to a 3D spatial grid to calculate and generate a 3D attribute volume, this step aims to extend the trained prediction model to the entire target exploration area. First, the entire exploration area is discretized into a 3D spatial grid containing millions or even hundreds of millions of cells. For each grid cell, the corresponding fluid focusing index value is extracted using its spatial coordinates. Then, the grade-thickness prediction model is invoked, and based on its internal cascade logic, a predicted grade value and a predicted thickness value are calculated for that cell. After traversing all grid cells, a 3D attribute volume is generated. This attribute volume is a large 3D array, where each element stores the predicted grade and predicted thickness information for that spatial location.

[0122] The three-dimensional attribute volume is initially screened to delineate the three-dimensional point cloud of potential mineralization bodies;

[0123] Specifically, in the initial screening of 3D attribute volumes to delineate the 3D point cloud of potential mineralizations, the purpose of this step is to filter out areas that are clearly not economically valuable, thereby reducing the complexity of subsequent calculations. Screening is performed by applying two basic economic and technical constraints: a minimum industrial grade threshold and a minimum industrial thickness threshold. Only when the predicted grade and predicted thickness of a grid cell simultaneously meet both of these lower limits is the cell retained. The set of all retained grid cells collectively constitutes a 3D point cloud of potential mineralizations distributed in a scattered pattern in 3D space.

[0124] Perform three-dimensional connected component analysis on the three-dimensional point cloud of the potential mineralization body to identify and segment candidate ore bodies;

[0125] Specifically, in the step of performing 3D connected component analysis on the 3D point cloud of potential mineralization bodies to identify and segment candidate ore bodies, this step is crucial for distinguishing between "mineralization points" and "ore bodies," aiming to identify mineralized areas with continuity and scale. In this step, a 3D connected component analysis algorithm is employed. This algorithm traverses every grid cell in the 3D point cloud of potential mineralization bodies and checks whether its surrounding neighboring cells also belong to the point cloud. In this way, all spatially adjacent or contacting cells are "glued" together, forming an independent connected cluster. The algorithm ultimately segments the entire point cloud into several separate but internally continuous clusters, each of which is identified as an independent candidate ore body.

[0126] The average grade of the candidate ore bodies is calculated, and they are classified according to the first, second, and third grade thresholds to generate a three-dimensional rich ore exploration target area.

[0127] Specifically, in the step of calculating and classifying the average grade of candidate ore bodies to generate three-dimensional rich ore exploration target areas, this step is responsible for the final value assessment and grading of each candidate ore body identified in the previous step. For each individual candidate ore body, the arithmetic mean of the predicted grades of all its internal grid cells is first calculated to obtain the average grade of the candidate ore body, as shown in the formula:

[0128] ,

[0129] in, It is the average grade of a candidate ore body; It is the total number of grid cells that make up the candidate ore body; This is the predicted grade of the i-th unit. Then, the calculated average grade is compared with preset first, second, and third grade thresholds to assign each candidate ore body a grade of "high grade," "medium grade," or "low grade." Finally, a continuous mineralized body in three-dimensional space containing multiple graded mineralized bodies with clearly defined boundaries and scales constitutes the final generated three-dimensional rich ore exploration target area, such as... Figure 4 As shown, a 3D view overlays the final delineated high, medium, and low-grade ore bodies with the actual borehole paths used to verify the model, intuitively demonstrating the spatial correspondence between the prediction results and the actual data.

[0130] For example, a grade-thickness prediction model was applied to a 3D exploration area containing ten million grid cells, generating a 3D attribute volume. After initial screening, the remaining one hundred thousand grid cells that met the criteria constituted a 3D point cloud of potential mineralization bodies. A 3D connected component analysis algorithm processed this point cloud, identifying five large, independent candidate ore bodies and discarding a large number of scattered, discontinuous mineralization points. Subsequently, the average grade of these five candidate ore bodies was calculated: A = 7.5 g / t, B = 4.2 g / t, C = 1.8 g / t, D = 6.1 g / t, and E = 0.9 g / t. Assuming the first, second, and third grade thresholds are 5 g / t, 2 g / t, and 1 g / t, respectively, A and D are classified as "high-grade target areas," B as "medium-grade target areas," C as "low-grade target areas," and E is eliminated because it is below the minimum grade threshold. The final prospecting map will clearly show four exploration target areas, A, B, C, and D, with different value levels.

[0131] Optionally, the method further includes:

[0132] A comparative analysis of the spatial distribution of the fluid focusing index and the spatial distribution of the three-dimensional rich ore exploration target area was conducted, and a spatial deviation residual map was calculated and generated.

[0133] Specifically, in the step of comparing and analyzing the spatial distribution of the fluid focusing index with the spatial distribution of the 3D rich mineral exploration target area to calculate and generate a spatial deviation residual map, this step aims to quantify the inconsistency between theoretical model predictions and data-driven results. First, the 3D rich mineral exploration target area is transformed into a continuous calibration favorability map. For example, high values ​​can be assigned to grid cells within the target area, and low values ​​to cells outside the target area. Then, a grid-by-grid subtraction operation is performed between the fluid focusing index distribution map and this calibration favorability map, the calculation process of which is shown in the formula:

[0134] ,

[0135] in, This is the final generated spatial deviation residual map; It is a three-dimensional distribution map of the fluid focusing index; This is a calibration favorability map derived from the final exploration target area. Positive values ​​in this residual map represent areas where the physical model overestimates the mineralization potential, while negative values ​​represent areas where the physical model underestimates the mineralization potential.

[0136] Based on the spatial deviation residual plot, the permeability coefficient value is corrected using closed-loop feedback.

[0137] Specifically, in the step of closed-loop feedback correction of permeability coefficient values ​​based on the spatial bias residual map, this step is the core of achieving model self-optimization. An inverse optimization algorithm takes the generated spatial bias residual map as input. The core logic of this algorithm is: if the residual of a certain region is significantly positive, it indicates that the assigned permeability coefficient value for that region may be too high and needs to be lowered; conversely, the opposite is also true. The algorithm iteratively calculates and obtains a set of optimal permeability adjustment amounts, which are then applied to the original permeability coefficient values ​​to generate a new version of permeability coefficient values ​​after closed-loop feedback correction. The update process can be expressed by the formula:

[0138] ,

[0139] in, This is a new version of the penetration rate coefficient value that has been corrected based on feedback; This is the original permeability coefficient value; It is a residual plot based on spatial bias This is a function used to calculate the permeability adjustment. This new set of permeability coefficient values ​​will be used in all subsequent new mineral exploration predictions, thus continuously improving the predictive power of the entire physical model as exploration data increases.

[0140] For example, after completing a round of mineral exploration prediction, this method compares the fluid focusing index distribution map generated by physical simulation with the three-dimensional rich mineral exploration target area delineated based on borehole data. It was found that the fluid focusing index was generally higher in the northern part of the exploration area, but the final exploration target area did not delineate a ore body there. This resulted in a large area of ​​significantly positive values ​​on the spatial deviation residual map of this northern region. An inverse optimization algorithm identified this deviation and inferred that its cause was likely an overestimation of the permeability of the rock in this area. Therefore, the algorithm automatically lowered the permeability coefficient value of the rock in this northern region and generated a new version of the physical model. In the next prediction using new borehole data, the updated model will no longer incorrectly overestimate this area, thus achieving model self-correction and evolution.

[0141] It should be noted that the formulas described above, through the principle of dimensional consistency and mathematical standardization methods (such as normalization, dimensionless parameter conversion, or unit system unification), can translate physical quantities with different properties into unitless standard values ​​or parameters that can be superimposed in the same dimension. This eliminates the interference of different dimensions on the computational logic, allowing the formulas to retain the original data distribution characteristics while possessing mathematical rationality and adaptability to objective laws. These are conventional technical methods and will not be elaborated further. The electrical connections between the various units described above do not necessarily represent direct or indirect connections; any indirect connection method is applicable to the embodiments of this invention as long as it achieves the purpose of this invention. The above descriptions are merely exemplary embodiments of this invention and should not be construed as limiting the scope of this invention.

[0142] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A method for prospecting rich sections of altered rock-type gold deposits, characterized in that, The method includes: Acquire and integrate geological, geophysical and core data from the exploration area to construct a three-dimensional seepage physical model; Based on the aforementioned three-dimensional seepage physical model, numerical simulations were performed to calculate the flow field parameters. The fluid pressure and velocity distribution in the flow field parameters are calculated to construct a fluid focusing index; Regression calibration was performed on the relationship between the fluid focusing index and the actual mineralization enrichment degree to establish a grade-thickness prediction model; Based on the grade-thickness prediction model, and according to the preset first, second and third grade thresholds, a three-dimensional rich ore exploration target area is generated.

2. The method for prospecting rich sections of altered rock-type gold deposits according to claim 1, characterized in that, The construction of the three-dimensional seepage physical model includes: Spatial interpolation and three-dimensional modeling are performed on the geological, geophysical and core data of the exploration area to reproduce the spatial distribution of alteration zoning and generate three-dimensional geological structure and alteration zoning. Permeability coefficient values ​​are assigned to different regions in the three-dimensional geological structure and alteration zone to construct a three-dimensional seepage physical model.

3. The method for prospecting rich sections of altered rock-type gold deposits according to claim 2, characterized in that, Assigning permeability coefficient values ​​to different regions within the three-dimensional geological structure and alteration zones includes: Based on the three-dimensional geological structure and the rock type and alteration intensity of the alteration zone, an initial permeability value is assigned by matching from a preset rock property database to generate a baseline permeability spatial distribution. Geometric analysis was performed on the three-dimensional geological structure and alteration zoning to generate a structural complexity correction factor; The construction complexity correction factor is applied to the baseline permeability spatial distribution for weighted adjustment to generate permeability coefficient values.

4. The method for prospecting rich sections of altered rock-type gold deposits according to claim 1, characterized in that, The calculated flow field parameters include: Boundary conditions are set in the three-dimensional seepage physical model, dynamic calculations are performed, and a three-dimensional fluid pressure field distribution is generated. Simultaneously calculate the three-dimensional flow velocity of the fluid in the three-dimensional seepage physical model to obtain the three-dimensional fluid velocity field distribution; The flow field parameters are generated by combining the three-dimensional fluid pressure field distribution with the three-dimensional fluid velocity field distribution.

5. The method for prospecting rich sections of altered rock-type gold deposits according to claim 4, characterized in that, The combination of the three-dimensional fluid pressure field distribution and the three-dimensional fluid velocity field distribution includes: The curl of the three-dimensional fluid velocity field distribution is calculated to obtain the three-dimensional fluid vortex field; The gradient of the three-dimensional fluid pressure field distribution is calculated, and the modulus is extracted to generate a pressure gradient scalar field. The three-dimensional fluid vortex field and the pressure gradient scalar field are aggregated to generate flow field parameters.

6. The method for prospecting rich sections of altered rock-type gold deposits according to claim 4, characterized in that, The constructed fluid focusing index includes: The three-dimensional fluid vorticity field and the pressure gradient scalar field are weighted and summed to generate a fluid dynamics favorability index. Based on the aforementioned three-dimensional geological structure and alteration zoning, geochemical acceptability scores are assigned to spatial locations to generate a geochemical favorability index. The fluid focusing index is constructed by calculating the fluid dynamics favorability index and the geochemical favorability index.

7. The method for prospecting rich sections of altered rock-type gold deposits according to claim 6, characterized in that, The geochemical favorability index includes: The three-dimensional geological structure and alteration zoning are rasterized, and the rock type and local thickness of each raster unit are quantified to generate a lithology-structure baseline favorability distribution map. Extract the geochemical data corresponding to the raster unit and quantify it into geochemical correction coefficients; The geochemical correction coefficient is used to adjust the lithology-structure baseline favorability distribution map grid by grid to calculate and generate the geochemical favorability index.

8. The method for prospecting rich sections of altered rock-type gold deposits according to claim 1, characterized in that, The establishment of the grade-thickness prediction model includes: Obtain borehole data from known mineral deposits, extract the fluid focusing index, ore grade, and ore thickness, and generate a training sample set. Using the fluid focusing index as the input feature and the ore body grade as the output label, a machine learning regression model is trained using the training sample set to obtain an ore body grade prediction sub-model. Using the fluid focusing index and the actual ore body grade as input features, and the ore body thickness as the output label, a machine learning regression model is trained using the training sample set to obtain an ore body thickness prediction sub-model. The ore body grade prediction sub-model and the ore body thickness prediction sub-model are cascaded and combined to construct a grade-thickness prediction model.

9. The method for prospecting rich sections of altered rock-type gold deposits according to claim 1, characterized in that, The generated three-dimensional rich mineral exploration target area includes: The grade-thickness prediction model is applied to the three-dimensional spatial grid of the exploration area to calculate and generate a three-dimensional attribute volume. The three-dimensional attribute volume is initially screened to delineate the three-dimensional point cloud of potential mineralization bodies; Perform three-dimensional connected component analysis on the three-dimensional point cloud of the potential mineralization body to identify and segment candidate ore bodies; The average grade of the candidate ore bodies is calculated, and they are classified according to the first, second, and third grade thresholds to generate a three-dimensional rich ore exploration target area.

10. A method for prospecting rich sections of altered rock-type gold deposits according to claim 2, characterized in that, The description also includes: A comparative analysis of the spatial distribution of the fluid focusing index and the spatial distribution of the three-dimensional rich ore exploration target area was conducted, and a spatial deviation residual map was calculated and generated. Based on the spatial deviation residual plot, the permeability coefficient value is corrected using closed-loop feedback.

Images