Rare earth ore in-situ leaching intelligent control system and method based on digital twinning

Abstract

This invention relates to the field of automatic control technology for rare earth mining, specifically disclosing an intelligent control system and method for in-situ leaching of rare earth minerals based on digital twins. The system collects real-time data on groundwater level, leaching agent concentration, mother liquor flow rate, and spatial geological data; calculates the leaching state vector field based on seepage mechanics and ion exchange constitutive relations; corrects the optimal injection template based on this vector field, generating dynamic injection volume distribution maps and dynamic injection concentration distribution maps; performs deviation analysis on actual injection parameters, outputting a binary map of leaching anomaly areas; spatially overlays this binary map with grade distribution maps and permeability zoning maps, analyzing the topological relationship between the injection anomaly areas and high-grade and low-permeability areas; judges the risk of leaching rate decline or leakage based on the topological relationship, generating solenoid valve opening commands and variable frequency pump speed commands for each injection well, achieving precise addition of leaching agent; this invention can improve rare earth recovery rate and reduce environmental pollution caused by leaching agent leakage.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology for rare earth mining, specifically to an intelligent control system and method for in-situ leaching of rare earth minerals based on digital twins. Background Technology

[0002] Ion-adsorption rare earth deposits (also known as ion-adsorption rare earth deposits) are precious mineral resources unique to my country, mainly distributed in the southern regions. In these deposits, rare earth elements are adsorbed onto the surface of clay minerals in the form of hydrated ions or hydroxyl hydrated ions. Currently, in-situ leaching is commonly used for mining. This involves injecting leaching solutions such as ammonium sulfate or magnesium sulfate into the ore body through injection wells. The ammonium or magnesium ions in the leaching agent exchange rare earth ions, forming a rare earth mother liquor, which is then collected in a mother liquor treatment workshop through a collection pipeline network.

[0003] The existing technology has the following shortcomings: In the in-situ leaching process of ion-adsorption rare earth minerals, the spatiotemporal dynamic coupling between the seepage field and the ion concentration field is difficult to characterize in real time, which leads to the inability of the injection parameters to be adaptively adjusted according to the internal state of the ore body, resulting in low rare earth leaching rate and lateral leakage of leaching agent. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent control system and method for in-situ leaching of rare earth minerals based on digital twins, so as to solve the problems mentioned above.

[0005] The objective of this invention can be achieved through the following technical solutions: A digital twin-based intelligent control method for in-situ leaching of rare earth minerals includes the following steps: S1 is used to collect in real time the groundwater level, the concentration of leaching agent in the injection pipeline, the flow rate of rare earth mother liquor in the collection pipeline, and the spatial coordinates and geological stratification data of the mine during the in-situ leaching process of ion-adsorption rare earth ore. S2, based on groundwater level, leaching agent concentration and geological stratification data, calculates the seepage field and ion concentration field inside the mine at the current moment according to seepage mechanics and constitutive relations of ion exchange reaction, and outputs the leaching state vector field corresponding to the physical mine state; S3, based on the leaching state vector field, perform geometric and parameter correction on the pre-stored optimal injection template for the mine, and generate a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state; S4. Using dynamic injection volume distribution map and dynamic injection concentration distribution map, deviation analysis is performed on the actual injection parameters of each injection well collected in real time, and the areas of insufficient injection and excessive injection are extracted, and a binary map of the leaching abnormal area is output. S5. Spatially overlay the binary map of the leaching anomaly area with the grade distribution map and permeability zoning map of the mine to analyze the spatial topological relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body. S6, based on the spatial topology, determines whether abnormal injection areas will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage, and generates solenoid valve opening commands and variable frequency pump speed commands for each injection well based on the judgment results, so as to achieve precise addition of leaching agent.

[0006] As a further aspect of the present invention: S2 specifically includes: Based on the groundwater level data from each observation well, a three-dimensional distribution map of groundwater equipotential surfaces inside the mine was constructed. The pressure gradient is calculated along the normal direction of the equipotential surface, and the seepage velocity components are solved grid by grid according to Darcy's law to form the seepage velocity vector field. The seepage velocity vector field is spatially matched with the measured values ​​of the leaching agent concentration in each injection well. The convection-diffusion equation of the ion exchange reaction is solved iteratively along the streamline direction to update the ion concentration value of each grid. The percolation velocity vector field and the updated ion concentration field are encapsulated together as a leaching state vector field output.

[0007] As a further aspect of the present invention: the formation of the seepage velocity vector field specifically includes: The three-dimensional groundwater equipotential surface is discretized into a set of triangular patches, and the centroid coordinates and unit normal vector determined by the isopleth distribution of pressure values ​​are extracted for each patch. Along the unit normal vector direction of each patch, extend forward and backward by one grid step each, obtain the pressure values ​​at the two extension endpoints, and use the central difference method to calculate the pressure gradient magnitude at the corresponding patch. The pressure gradient magnitude is multiplied by the permeability coefficient of the grid and then divided by the dynamic viscosity coefficient. The resulting scalar value is projected along the unit normal direction to generate the seepage velocity vector component of each patch. All patch components are then combined to form the seepage velocity vector field.

[0008] As a further aspect of the present invention: S3 specifically includes: Extract the seepage velocity amplitude and ion concentration gradient direction at each injection well location from the leaching state vector field, and read the preset injection volume and preset concentration at the corresponding location in the optimal injection template for the mine. The ratio of the seepage velocity amplitude to the target seepage velocity threshold is used as the injection volume correction coefficient. At the same time, the cosine value of the angle between the ion concentration gradient direction and the injection diffusion direction is calculated, and the increase or decrease correction amount of the preset concentration is determined according to the sign of the cosine value. The injection volume correction coefficient and concentration increase / decrease correction of each injection well are spatially distributed to the surrounding un-welled area to generate continuous dynamic injection volume distribution map and dynamic injection concentration distribution map respectively.

[0009] As a further aspect of the present invention: the generation of continuous dynamic injection volume distribution map and dynamic injection concentration distribution map specifically includes: With each injection well as the center, an elliptical influence domain is constructed according to the anisotropic radius set in the mine permeability zoning map, and the correction coefficient and correction amount at the injection well are assigned as the center value of the ellipse. Within the elliptical influence domain, the decay rate is calculated along the major and minor axes, and weight values ​​are assigned based on the normalized Euclidean distance from each grid point to the center point within the elliptical domain. The weight values ​​of the overlapping areas of the elliptical influence domains of multiple injection wells are normalized and accumulated. The accumulated weight values ​​are then used to perform a weighted average of each correction coefficient and correction amount to generate a continuous distribution map of the entire mining area.

[0010] As a further aspect of the present invention: S4 specifically includes: The injection volume deviation value is obtained by subtracting the actual injection volume collected in real time from the value of the dynamic injection volume distribution map of each injection well. At the same time, the injection concentration deviation value is obtained by subtracting the actual injection concentration collected in real time from the value of the dynamic injection concentration distribution map. Set injection volume deviation threshold and injection concentration deviation threshold respectively. Mark the grid where the injection volume deviation value is lower than the negative threshold and the injection concentration deviation value is also lower than the negative threshold as the injection insufficient area, and mark the grid where both deviation values ​​are higher than the positive threshold as the injection excessive area. The grids in the insufficient and excessive injection areas are assigned a value of 1, and the remaining grids are assigned a value of 0, outputting a binarized map of the leaching anomaly area.

[0011] As a further aspect of the present invention: S5 specifically includes: The insufficient injection area in the binary map of the leaching anomaly region is spatially intersected with the high-grade area in the grade distribution map and the low-permeability area in the permeability zoning map to extract the overlapping grid set of insufficient injection and high grade and the overlapping grid set of insufficient injection and low permeability. Calculate the proportion of the overlapping grid area of ​​the insufficiently injected high-grade fluid to the total area of ​​the insufficiently injected region, and at the same time calculate the proportion of the overlapping grid area of ​​the insufficiently injected low-permeability fluid to the total area of ​​the insufficiently injected region. When the proportion of high-grade water content is greater than the first threshold and the proportion of low-penetration water content is less than the second threshold, it is determined to be a high-priority topology; when the proportion of low-penetration water content is greater than the second threshold, it is determined to be a non-replenishable topology.

[0012] As a further aspect of the present invention: S6 specifically includes: When a high-priority injection topology relationship is determined, the target injection volume and target concentration of the corresponding overlapping grid are extracted from the leaching state vector field, and the current actual injection volume and actual concentration of the injection well are subtracted to obtain the injection volume increment and concentration increment, respectively. Input the injection volume increment into the pressure-opening characteristic curve of the pipeline where the injection well is located, and look up the increase in the opening angle of the solenoid valve. At the same time, divide the concentration increment by the current storage concentration of the mother liquor tank to obtain the speed increase ratio of the variable frequency pump. The current solenoid valve opening degree and the increase in opening angle are added together to form the opening degree command output. The current variable frequency pump speed is multiplied by 1 and then added to the speed increase ratio to form the speed command output.

[0013] As a further aspect of the present invention: obtaining the increase in the opening angle of the solenoid valve specifically includes: Real-time pressure values ​​are read from the inlet and outlet ends of the pipeline where the injection well is located, and the difference between the two is calculated as the current working pressure difference. The target curve family in the pressure-opening characteristic curve is located using the current working pressure difference as an index. On the target curve family, starting from the actual opening angle of the current solenoid valve, move the distance along the horizontal axis by a distance equal to the change in angle corresponding to the increase in liquid injection volume divided by the cross-sectional area of ​​the pipe and then divided by the flow rate conversion coefficient, and read whether the pressure difference on the vertical axis corresponding to the change in angle is consistent with the current working pressure difference. If they are consistent, the angle change is directly used as the increase in the opening angle. If they are inconsistent, the angle change is corrected until the calculated pressure difference is equal to the current working pressure difference, and the final corrected increase in the opening angle is output.

[0014] A digital twin-based intelligent control system for in-situ leaching of rare earth minerals, comprising: The data synchronization acquisition module is used to collect in real time the groundwater level, the concentration of leaching agent in the injection pipeline, the flow rate of rare earth mother liquor in the collection pipeline, and the spatial coordinates and geological stratification data of the mine during the in-situ leaching process of ion-adsorption rare earth ore. The physical process calculation module, based on groundwater level, leaching agent concentration and geological stratification data, and according to seepage mechanics and constitutive relations of ion exchange reaction, calculates the seepage field and ion concentration field inside the mine at the current moment, and outputs the leaching state vector field corresponding to the physical mine state. The dynamic injection map generation module performs geometric and parameter correction on the pre-stored optimal injection template of the mine based on the leaching state vector field, and generates a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state. The leaching deviation identification module uses dynamic injection volume distribution map and dynamic injection concentration distribution map to perform deviation analysis on the actual injection parameters of each injection well collected in real time, extract the insufficient injection area and the excessive injection area, and output a binary map of the leaching abnormal area. The spatial topology analysis module spatially overlays the binary map of the leaching anomaly area with the grade distribution map and permeability zoning map of the mine to analyze the spatial topology relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body. The control command generation module determines whether abnormal injection areas will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage based on spatial topology. Based on the determination results, it generates solenoid valve opening commands and variable frequency pump speed commands for each injection well to achieve precise addition of leaching agent.

[0015] The beneficial effects of this invention are: (1) By analyzing the deviation of actual injection parameters through dynamic injection volume distribution map and dynamic injection concentration distribution map, and prioritizing injection in high-grade areas and restricting injection in low-permeability areas based on spatial topology, it is possible to reduce rare earth residue caused by insufficient injection and reagent waste caused by excessive injection, thereby improving the rare earth resource recovery rate and reducing the unit consumption of leaching agent.

[0016] (2) Based on the real-time correction of the injection parameters by the leaching state vector field, and combined with the spatial superposition analysis of the low permeability zone and the insufficient injection zone, the topological relationship that cannot be replenished can be determined, which can avoid the large amount of leaching agent remaining in the low permeability zone and lateral leakage, thereby reducing the environmental risk of leaching agent entering the surrounding groundwater. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a system block diagram of the present invention. Detailed Implementation

[0019] 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, and 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.

[0020] Please see Figure 1 As shown, this invention is an intelligent control method for in-situ leaching of rare earth minerals based on digital twins, comprising the following steps: S1 is used to collect in real time the groundwater level, the concentration of leaching agent in the injection pipeline, the flow rate of rare earth mother liquor in the collection pipeline, and the spatial coordinates and geological stratification data of the mine during the in-situ leaching process of ion-adsorption rare earth ore. S2, based on groundwater level, leaching agent concentration and geological stratification data, calculates the seepage field and ion concentration field inside the mine at the current moment according to seepage mechanics and constitutive relations of ion exchange reaction, and outputs the leaching state vector field corresponding to the physical mine state; S3, based on the leaching state vector field, perform geometric and parameter correction on the pre-stored optimal injection template for the mine, and generate a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state; S4. Using dynamic injection volume distribution map and dynamic injection concentration distribution map, deviation analysis is performed on the actual injection parameters of each injection well collected in real time, and the areas of insufficient injection and excessive injection are extracted, and a binary map of the leaching abnormal area is output. S5. Spatially overlay the binary map of the leaching anomaly area with the grade distribution map and permeability zoning map of the mine to analyze the spatial topological relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body. S6, based on the spatial topology, determines whether abnormal injection areas will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage, and generates solenoid valve opening commands and variable frequency pump speed commands for each injection well based on the judgment results, so as to achieve precise addition of leaching agent.

[0021] In S1, data is collected in real time during the in-situ leaching process of ion-adsorption rare earth ore, including groundwater level, leaching agent concentration in the injection pipeline, rare earth mother liquor flow rate in the collection pipeline, and spatial coordinates and geological stratification data of the mine. Specifically, this includes: At the in-situ leaching site of ion-adsorption rare earth ore, submersible pressure level gauges were installed in each injection well and observation well. The sensor probes of the pressure level gauges were located in the middle of the filter pipe inside the well and were connected to a ground data acquisition terminal via cables. The data acquisition terminal read the pressure values ​​every 5 minutes and performed compensation calculations based on pre-measured atmospheric pressure to obtain the groundwater level elevation data in each well, while also recording the data acquisition timestamp.

[0022] Online conductivity sensors are installed on the main pipe and each branch pipe of the injection pipeline. The sensor probes extend into the inner wall of the pipe and contact the liquid flow. When the leaching agent solution flows through the probe, the sensor measures the conductivity value of the solution in real time. The data acquisition module reads the conductivity value every 2 minutes through the analog input channel and converts the conductivity into the mass percentage concentration of the leaching agent according to the pre-calibrated conductivity-concentration conversion curve, and stores it in the local cache.

[0023] An electromagnetic flowmeter is installed at the main outlet after the liquid collection pipelines converge. The flowmeter's lining material is polytetrafluoroethylene (PTFE) to adapt to the acidic and alkaline environment of the rare earth mother liquor. The excitation coil of the electromagnetic flowmeter is powered by a 24-volt DC power supply from the data acquisition terminal. It collects the instantaneous flow signal every minute, and the cumulative flow value is transmitted to the counting channel of the data acquisition terminal via pulse output to calculate the volumetric flow rate of the rare earth mother liquor and the total liquid collection volume per unit time.

[0024] The spatial coordinates and geological stratification data of the mine were obtained as follows: Coordinate measurements were performed on each injection wellhead, observation wellhead, and collection wellhead using a real-time dynamic differential global positioning system to obtain the geodetic coordinates and elevations of each wellhead. Geological stratification data was compiled based on core logging data from exploration boreholes used in the mine's earlier construction phases. The ore body was divided from top to bottom into topsoil, completely weathered layer, semi-weathered layer, and bedrock layer, and the elevations of the top and bottom plates of each layer were recorded, forming a three-dimensional geological stratification database.

[0025] In S2, based on groundwater level, leaching agent concentration, and geological stratification data, and according to seepage mechanics and constitutive relations of ion exchange reactions, the seepage field and ion concentration field inside the mine at the current moment are calculated, and the leaching state vector field corresponding to the physical mine state is output, specifically including: First, based on the groundwater level data measured from each observation well, a three-dimensional equipotential surface distribution map of groundwater within the mine is constructed. Specifically, the groundwater level elevation values ​​from all observation wells are used as known data points. Radial basis function interpolation is employed to interpolate the groundwater level elevation at each 0.5-meter-side cubic grid node within the three-dimensional space of the mine, obtaining the groundwater level elevation at each grid node. Adjacent grid nodes with the same groundwater level elevation are then connected using a smooth surface, forming a series of equipotential surfaces corresponding to different groundwater levels. The water pressure at each point on each equipotential surface is equal.

[0026] Secondly, the pressure gradient is calculated along the normal direction of the equipotential surface, and the seepage velocity components are solved grid by grid according to Darcy's law to form a seepage velocity vector field. The specific implementation method will be described in detail in subsequent paragraphs.

[0027] Next, the seepage velocity vector field is spatially matched with the measured values ​​of the leaching agent concentration in each injection well. The convection-diffusion equation for the ion exchange reaction is iteratively solved along the streamline direction, updating the ion concentration value of each grid. Specifically, the measured leaching agent concentration of each injection well is used as the boundary condition for the concentration field. Starting from the grid where the injection well is located, the streamline is traced along the seepage velocity vector direction, advancing one grid at each time step. Within each grid, the change in ion concentration is jointly determined by the convection and diffusion terms. The convection term is calculated by multiplying the concentration flowing into the grid by the seepage velocity, while the diffusion term is calculated by dividing the concentration difference between adjacent grids by the grid spacing and then multiplying by the diffusion coefficient. The ion exchange reaction term adopts the instantaneous equilibrium assumption, meaning that when the leaching agent flows through the grid, it replaces ammonium ions in the solution with rare earth ions on the ore body surface according to the ion exchange equilibrium constant. The amount of replacement depends on the remaining rare earth grade within the grid and the current ammonium ion concentration. The iterative calculation continues until the relative change in ion concentration value of each grid within two consecutive time steps is less than one-thousandth. At this point, the concentration field is considered to have converged, and the updated ion concentration field is obtained.

[0028] Finally, the seepage velocity vector field and the updated ion concentration field are stored together as the leaching state vector field. The specific encapsulation method is as follows: each grid node stores two physical quantities, namely the seepage velocity vector (containing direction and magnitude information) and the ion concentration value (in grams per liter). The data set of all grid nodes is the leaching state vector field, which is used for subsequent steps.

[0029] The following describes the three steps for forming the seepage velocity vector field.

[0030] The first step is to discretize the 3D groundwater equipotential surface into a set of triangular patches, and extract the centroid coordinates and unit normal vector of each patch. Specifically, for each equipotential surface, the Delaunay triangulation algorithm is used to divide it into multiple triangular patches, where the three vertices of each triangular patch are grid nodes on the equipotential surface. The centroid coordinates of each triangular patch are calculated, which is the average of the coordinates of the three vertices. Simultaneously, based on the equal distribution of pressure values ​​at the three vertices, the normal direction of the plane containing that patch is determined and normalized to obtain the unit normal vector, which points in the direction of decreasing pressure.

[0031] The second step involves extending the grid forward and backward by one grid step along the unit normal direction of each patch, obtaining the pressure values ​​at the two extended endpoints. The central difference method is then used to calculate the pressure gradient amplitude at the corresponding patch. Specifically, starting from the centroid of the patch, move one grid step (0.5 meters) along the unit normal direction to the forward endpoint, then move one grid step in the opposite direction to the backward endpoint. Read the pressure values ​​at the grid nodes containing the forward and backward endpoints, which are obtained by multiplying the groundwater level by the unit weight of water. Subtract the pressure value at the backward endpoint from the pressure value at the forward endpoint, then divide by twice the grid step size (1 meter). The quotient is the pressure gradient amplitude at that patch, expressed in Pascals per meter.

[0032] The third step involves multiplying the pressure gradient amplitude by the permeability coefficient of the corresponding grid and then dividing by the dynamic viscosity coefficient. The resulting scalar value is projected along the unit normal direction to generate the seepage velocity vector component for each facet. All facet components are then aggregated to form a seepage velocity vector field. The specific calculation process is as follows: The permeability coefficient of the grid containing the facet's centroid is obtained. This coefficient was determined by previous hydraulic tests and is measured in square meters. The dynamic viscosity coefficient of the leaching agent is obtained by measuring the leaching agent solution at the current temperature using an indoor viscometer and is measured in Pascal-seconds. The product of the pressure gradient amplitude and the permeability coefficient is divided by the dynamic viscosity coefficient to obtain a scalar value, which represents the magnitude of the seepage velocity. This scalar value is multiplied by the unit normal vector of the facet to obtain the seepage velocity vector component. This calculation is repeated for each triangular facet to obtain the seepage velocity vector components for all facets. These vector components are then weighted and averaged according to their respective grid node positions and distributed to each grid node, ultimately forming a seepage velocity vector field covering the entire three-dimensional grid of the mine.

[0033] In S3, based on the leaching state vector field, the pre-stored optimal injection template for the mine is geometrically and parametrically corrected to generate a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state. Specifically, this includes: The overall steps for generating dynamic injection volume distribution maps and dynamic injection concentration distribution maps are as follows: First, extract the seepage velocity amplitude and ion concentration gradient direction at each injection well location from the leaching state vector field, and read the preset injection volume and preset concentration at the corresponding location in the mine's optimal injection template. Specifically, for each injection well, locate the corresponding grid node in the leaching state vector field based on its spatial coordinates, read the stored seepage velocity vector at that node, and calculate the magnitude of this vector as the seepage velocity amplitude, in meters per day. Simultaneously, read the ion concentration value at that node and calculate the concentration difference between that node and its eight adjacent grid nodes, taking the direction with the largest difference as the ion concentration gradient direction. Then, read two preset values ​​at the injection well coordinates from the pre-stored mine's optimal injection template: the preset injection volume (cubic meters per hour) and the preset concentration (mass percentage).

[0034] Secondly, the ratio of the seepage velocity amplitude to the target seepage velocity threshold is calculated as the injection volume correction coefficient. The target seepage velocity threshold is set at 0.5 meters per day. This threshold was determined by previous indoor column leaching experiments. When the seepage velocity is below 0.5 meters per day, the prolonged contact time between the leaching agent and the ore body can easily lead to exacerbated side reactions; when it is above 0.5 meters per day, insufficient contact time results in incomplete ion exchange. The formula for calculating the injection volume correction coefficient is: ; in, This is a dimensionless correction factor for the injection volume. This represents the seepage velocity amplitude, expressed in meters per day. The target seepage velocity threshold is set at 0.5 meters per day. A value greater than 1 indicates that the current seepage rate is too fast, and the injection volume needs to be reduced; when A value less than 1 indicates that the seepage rate is too slow and the injection volume needs to be increased.

[0035] Simultaneously, the cosine of the angle between the ion concentration gradient direction and the injection diffusion direction is calculated, and the adjustment amount for increasing or decreasing the preset concentration is determined based on the sign of this cosine value. The injection diffusion direction is taken as the direction of the seepage velocity vector at the injection well location, i.e., the main direction of the leaching agent diffusion from the injection well to the surrounding area. The angle between the ion concentration gradient direction and the injection diffusion direction is denoted as... The remaining chord values The calculation method is as follows: divide the dot product of the two direction vectors by the product of the magnitudes of the two vectors. Concentration increase / decrease correction amount. The determination rule is: if A value greater than 0 indicates that the direction of the ion concentration gradient is basically consistent with the direction of diffusion of the injected solution. This means that the current concentration is increasing along the diffusion direction, and the preset concentration needs to be reduced. The reduction amount is the preset concentration multiplied by 0. Multiply by a coefficient of 0.3; if A value less than 0 indicates that the two directions are opposite, meaning the current concentration is decreasing along the diffusion direction, and the preset concentration needs to be increased by multiplying the preset concentration by 0. The absolute value is then multiplied by a coefficient of 0.3. This coefficient of 0.3 is obtained from multiple field tests and regressions, and is used to control the correction amplitude to avoid oscillations.

[0036] Finally, the injection volume correction coefficient and concentration increase / decrease correction of each injection well are spatially distributed to the surrounding un-welled area, generating continuous dynamic injection volume distribution maps and dynamic injection concentration distribution maps, respectively. The specific implementation method will be described in detail in the next section.

[0037] The specific steps for generating a continuous distribution map of spatial layout are as follows: First, using each injection well as the center, construct an elliptical influence domain according to the anisotropic radius set in the mine permeability zoning map. Assign the correction coefficient and adjustment amount at the injection well to the center value of the ellipse. Specifically, read the mine permeability zoning map, which divides the mining area into multiple regions. For each region, mark the permeability ratio along the east-west and north-south directions on the horizontal plane. For each injection well, using its coordinates as the center, take the radius in the east-west direction as the base radius of 10 meters multiplied by the east-west permeability ratio, and the radius in the north-south direction as the base radius of 10 meters multiplied by the north-south permeability ratio, forming an elliptical region. If the permeability ratio in a certain direction is greater than 1, the corresponding radius is extended; if it is less than 1, it is shortened. Assign the injection volume correction coefficient value at the injection well to the value at the center point of the ellipse; the concentration increase / decrease correction amount is handled similarly.

[0038] The second step involves calculating the attenuation rate along the major and minor axes within the elliptical influence domain, and assigning weights based on the normalized Euclidean distance from each grid point to the center point within the elliptical domain. The attenuation rate is calculated as follows: the length of the semi-major axis of the ellipse is denoted as... The length of the minor semi-axis is denoted as For any grid point within the ellipse, first calculate the distance from that point to the center of the ellipse, then calculate the projection component of that distance along the major axis. ratio And the projection components along the minor axis and ratio .Pick and The larger value in the range is used as the normalized distance. , The value ranges from 0 to 1. Weight value Calculated according to the Gaussian decay function: That is, negative 4 times the natural constant. The square power. Thus, the weight of points near the center of the ellipse is close to 1, and the weight of points near the boundary of the ellipse is close to 0.018.

[0039] The third step involves normalizing and accumulating the weight values ​​of the overlapping elliptical influence domains of multiple injection wells. Then, the accumulated weight values ​​are used to calculate a weighted average of each correction coefficient and adjustment amount, generating a continuous distribution map of the entire mining area. Specifically, for each grid point within the mining area, the elliptical influence domains of all injection wells covering that grid point are first identified, and the weight value calculated for each influence domain at that grid point is recorded. and the correction factor value at the injection well. Or corrected value Then calculate the normalized total weight of the grid point. Then calculate the correction coefficient after weighted average. Correction amount after weighted average After calculating for each grid point individually, the values ​​of each grid point are... The values ​​are filled into the corresponding planar coordinate positions to form a dynamic injection volume distribution map; the values ​​of each grid point are then entered. The values ​​are entered and then superimposed with the preset concentration values ​​in the optimal injection template for the mine to form a dynamic injection concentration distribution map. Both distribution maps are stored in a square grid with a side length of 0.5 meters for use in subsequent steps.

[0040] In S4, dynamic injection volume distribution maps and dynamic injection concentration distribution maps are used to perform deviation analysis on the actual injection parameters of each injection well collected in real time, extracting areas of insufficient and excessive injection, and outputting a binary map of leaching anomaly areas, specifically including: First, for each injection well location and its surrounding grid, the injection volume deviation is obtained by subtracting the actual injection volume collected in real time from the value of that grid in the dynamic injection volume distribution map. The specific calculation process is as follows: Read the injection volume value corresponding to that grid in the dynamic injection volume distribution map (in cubic meters per hour); read the actual injection volume of that well at the same moment from the real-time data collected by the on-site electromagnetic flowmeter (in cubic meters per hour); subtract the latter from the former; the difference is the injection volume deviation, which can be positive, negative, or zero. Simultaneously, the injection concentration deviation is obtained by subtracting the value of that grid in the dynamic injection concentration distribution map from the actual injection concentration collected in real time. The values ​​in the dynamic injection concentration distribution map are in mass percentage, while the actual injection concentration is converted from the concentration by the online conductivity sensor, also in mass percentage. Subtracting the two values ​​yields the injection concentration deviation.

[0041] Secondly, injection volume deviation thresholds and injection concentration deviation thresholds are set separately. The injection volume deviation threshold is set to ±0.05 cubic meters per hour, which is determined based on the measurement error range (±3%) of the electromagnetic flowmeter in the injection pipeline and the allowable fluctuation range of the injection well. The injection concentration deviation threshold is set to ±0.2 mass percentages, which is obtained by taking twice the safety margin after superimposing the measurement error (±0.1%) of the online conductivity sensor and the preparation accuracy (±0.1%) of the leaching agent preparation station. Grids with injection volume deviation values ​​below -0.05 cubic meters per hour and injection concentration deviation values ​​below -0.2 mass percentages are marked as insufficient injection areas, indicating that the actual injected quantity and concentration of leaching agent at this location are lower than the recommended values ​​on the dynamic map. Grids with injection volume deviation values ​​above +0.05 cubic meters per hour and injection concentration deviation values ​​above +0.2 mass percentages are marked as excessive injection areas, indicating that the actual injected value is higher than the recommended value.

[0042] Finally, the grids marked as insufficient and excessive leaching areas are assigned the value 1 in the binary image, while all other unmarked grids are assigned the value 0, resulting in a binary map of leaching anomalies. In this binary map, a value of 1 represents the location of a grid with leaching anomalies, and a value of 0 represents a location with normal leaching. The spatial resolution of the binary map is consistent with the dynamic leaching volume distribution map, i.e., each grid has a side length of 0.5 meters, covering the entire mining area, for use in subsequent spatial overlay analysis.

[0043] In S5, the binary map of the leaching anomaly area is spatially overlaid with the grade distribution map and permeability zoning map of the mine to analyze the spatial topological relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body, specifically including: First, spatial intersection operations are performed between the insufficient injection areas in the binary map of the leaching anomaly region and the high-grade areas in the mine's grade distribution map and the low-permeability areas in the permeability zoning map, respectively. This extracts overlapping grid sets for both insufficient injection and high-grade areas, and also for both insufficient injection and low permeability areas. Specifically, the binary map of the leaching anomaly region is read, and all grids marked as insufficient injection areas with a value of 1 are extracted, each grid having a side length of 0.5 meters. Simultaneously, the mine's grade distribution map is read. Based on previous borehole sampling and testing results, areas with rare earth oxide grades greater than 0.5 parts per thousand are defined as high-grade areas, and this is stored as a grid layer. Each grid in the insufficient injection area is compared with the corresponding grid in the grade distribution map. If the grid falls within a high-grade area, it is added to the overlapping grid set for both insufficient injection and high-grade areas. Similarly, the permeability zoning map is read. Based on on-site hydraulic tests, areas with permeability less than 0.3 meters per day are defined as low-permeability areas, and this is stored as a grid layer. Each grid in the insufficient injection area is compared with the low-permeability area layer. If the grid is in the low-permeability area, it is assigned to the overlapping grid set of insufficient injection and low permeability.

[0044] Secondly, the proportion of overlapping grid areas with insufficient and high-grade fluid injection to the total area of ​​the insufficient-injection region is calculated, as well as the proportion of overlapping grid areas with insufficient and low-permeability fluid injection to the total area of ​​the insufficient-injection region. The specific calculation process is as follows: Count the total number of grids in the insufficient-injection region. Multiply the area of ​​each grid (0.25 square meters, i.e., 0.5 meters multiplied by 0.5 meters) by the total number of grids to obtain the total area of ​​the insufficient-injection region. Count the number of grids in the overlapping grid set with insufficient and high-grade fluid injection, multiply by 0.25 square meters to obtain the area of ​​that overlapping region, divide this area by the total area of ​​the insufficient-injection region, and then multiply by 100% to obtain the proportion of high-grade fluid injection. Similarly, count the number of grids in the overlapping grid set with insufficient and low-permeability fluid injection, multiply by 0.25 square meters to obtain the low-permeability overlapping area, divide this area by the total area of ​​the insufficient-injection region, and then multiply by 100% to obtain the proportion of low-permeability fluid injection. For example, if there are a total of 400 grids in the insufficient injection area, with an area of ​​100 square meters, and 300 high-grade overlapping grids, with an area of ​​75 square meters, then the high-grade ratio is 75%.

[0045] Finally, a first threshold and a second threshold are set. The first threshold is set at 60%, based on the mine's economic assessment results. When the proportion of high-grade areas in the insufficient injection area exceeds 60%, priority injection has significant economic benefits. The second threshold is set at 40%, based on field injection tests. When the proportion of low-permeability areas in the insufficient injection area exceeds 40%, continued injection will cause a large amount of leaching agent to remain in the low-permeability area, increasing the risk of leakage and preventing effective rare earth extraction. When the proportion of high-grade areas is greater than 60% and the proportion of low-permeability areas is less than 40%, it is determined to be a high-priority injection topology, indicating that the insufficient injection area should be prioritized for injection. When the proportion of low-permeability areas is greater than 40%, it is determined to be a non-injectable topology, indicating that the area is not suitable for injection due to its low permeability, otherwise it will exacerbate the risk of lateral leakage of the leaching agent. If neither of the above two conditions is met, it is determined to be a routine observation topology, indicating that injection is not needed at this time or further analysis is required.

[0046] In S6, based on spatial topology, it determines whether abnormal injection areas will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage. Based on the determination results, it generates solenoid valve opening commands and variable frequency pump speed commands for each injection well to achieve precise addition of leaching agent. Specifically, this includes: The overall steps for generating solenoid valve opening commands and variable frequency pump speed commands are as follows: First, when a high-priority injection topology is determined based on spatial topology, the target injection volume and target concentration of the corresponding overlapping grid are extracted from the leaching state vector field. These are then subtracted from the current actual injection volume and actual concentration of the injection well to obtain the injection volume increment and concentration increment, respectively. Specifically, the grid node corresponding to the high-grade overlapping grid with insufficient injection volume is located in the leaching state vector field. The target injection volume value (in cubic meters per hour) and target concentration value (in mass percentage) stored at this node are read. Simultaneously, the current actual injection volume (measured by an electromagnetic flowmeter) and actual injection concentration (converted from an online conductivity sensor) of the injection well are read from real-time field data. The difference between the target injection volume and the actual injection volume is the injection volume increment, which can be positive or negative. The difference between the target concentration and the actual concentration is the concentration increment.

[0047] Secondly, the incremental injection volume is input into the pressure-opening characteristic curve of the pipeline where the injection well is located, and the increase in the opening angle of the solenoid valve is obtained by reverse lookup. Simultaneously, the concentration increment is divided by the current storage concentration in the mother liquor tank to obtain the percentage increase in the variable frequency pump's speed. The specific method for reverse lookup of the pressure-opening characteristic curve will be described in detail in the next section. The current storage concentration in the mother liquor tank is monitored in real time by an online concentration sensor at the tank outlet, and the unit is mass percentage. The calculation result of dividing the concentration increment by the current storage concentration indicates that if the concentration increment is positive, the speed increase percentage is positive, indicating that the variable frequency pump speed needs to be increased to increase the concentrate supply; if the concentration increment is negative, the speed increase percentage is negative, indicating that the speed needs to be decreased.

[0048] Finally, the current solenoid valve opening degree is superimposed with the increase in the opening angle to obtain the opening degree command output. The current variable frequency pump speed is multiplied by 1 and then added to the speed increase ratio to obtain the speed command output. The current solenoid valve opening degree is read in real time by the valve position feedback sensor, in degrees (0 degrees represents fully closed, 90 degrees represents fully open). The current opening degree is added to the increase in the opening angle. If the result exceeds 90 degrees, it is limited to 90 degrees; if it is less than 0 degrees, it is limited to 0 degrees, thus obtaining the final opening degree command, which is sent to the solenoid valve actuator via a 4 to 20 mA current signal. The current variable frequency pump speed is read in real time by the frequency converter, in revolutions per minute (RPM). The speed command is calculated by multiplying the current speed by (1 plus the speed increase ratio). For example, if the speed increase ratio is 0.15, then multiply by 1.15 to obtain the target speed command, which is sent to the frequency converter via the analog output channel.

[0049] The specific steps to obtain the increase in the solenoid valve opening angle are as follows: First, read the real-time pressure values ​​from the inlet and outlet ends of the pipeline where the injection well is located, calculate the difference between the two as the current working differential pressure, and use this working differential pressure as an index to locate the target curve family in the pressure-opening characteristic curve family. The inlet pressure sensor is installed on the pipeline before the branch of the injection main pipe, and the outlet pressure sensor is installed on the pipeline after the solenoid valve and before the injection wellhead. Subtract the outlet pressure value from the inlet pressure value to obtain the current working differential pressure, in megapascals (MPa). The pressure-opening characteristic curves are pre-calibrated experimentally. This curve family uses the differential pressure as a parameter, with the solenoid valve opening angle on the horizontal axis and the injection volume through the valve on the vertical axis. Based on the current working differential pressure value, find the corresponding characteristic curve in the curve family as the target curve.

[0050] The second step involves moving a distance along the horizontal axis on the target curve family, starting from the actual opening angle of the current solenoid valve. This distance is equal to the change in angle corresponding to the increase in injection volume divided by the pipe cross-sectional area and then divided by the velocity conversion factor. The pressure difference on the vertical axis corresponding to this change in angle is then read to see if it matches the current working pressure difference. The pipe cross-sectional area is calculated based on the inner diameter of the injection pipe; when the inner diameter is 50 mm, the cross-sectional area is approximately 0.0019625 square meters. The velocity conversion factor is set to 3600, used to convert cubic meters per hour to meters per second. Specifically, the increase in injection volume (cubic meters per hour) is divided by 3600 to obtain cubic meters per second, and then divided by the pipe cross-sectional area (square meters) to obtain the change in velocity (meters per second). This change in velocity is then correlated with the pre-calibrated valve flow coefficient to find the corresponding change in angle. On the target curve, move the angle change to the right or left from the current actual opening angle position, read the corresponding vertical coordinate (i.e., the calculated pressure difference), and compare whether the calculated pressure difference is equal to the current working pressure difference read in the first step.

[0051] The third step involves adjusting the angle change. If the calculated differential pressure matches the current working differential pressure, the angle change is directly used as the increase in the opening angle. If they don't match, the angle change is corrected until the calculated differential pressure equals the current working differential pressure, and the final corrected increase in the opening angle is output. The correction method uses a binary iterative approach: upper and lower boundaries are set for the angle change. Each time, the intermediate value is taken to recalculate the corresponding differential pressure. If the calculated differential pressure is greater than the current working differential pressure, the angle change is decreased; if it is less, the angle change is increased. This iteration is repeated until the absolute value of the difference between the calculated differential pressure and the current working differential pressure is less than 0.001 MPa. The angle change at this point is the final increase in the opening angle. This increase is then added to the current solenoid valve opening.

[0052] Please see Figure 2 As shown, a digital twin-based intelligent control system for in-situ leaching of rare earth minerals includes: The data synchronization acquisition module is used to collect in real time the groundwater level, the concentration of leaching agent in the injection pipeline, the flow rate of rare earth mother liquor in the collection pipeline, and the spatial coordinates and geological stratification data of the mine during the in-situ leaching process of ion-adsorption rare earth ore. The physical process calculation module, based on groundwater level, leaching agent concentration and geological stratification data, and according to seepage mechanics and constitutive relations of ion exchange reaction, calculates the seepage field and ion concentration field inside the mine at the current moment, and outputs the leaching state vector field corresponding to the physical mine state. The dynamic injection map generation module performs geometric and parameter correction on the pre-stored optimal injection template of the mine based on the leaching state vector field, and generates a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state. The leaching deviation identification module uses dynamic injection volume distribution map and dynamic injection concentration distribution map to perform deviation analysis on the actual injection parameters of each injection well collected in real time, extract the insufficient injection area and the excessive injection area, and output a binary map of the leaching abnormal area. The spatial topology analysis module spatially overlays the binary map of the leaching anomaly area with the grade distribution map and permeability zoning map of the mine to analyze the spatial topology relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body. The control command generation module determines whether abnormal injection areas will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage based on spatial topology. Based on the determination results, it generates solenoid valve opening commands and variable frequency pump speed commands for each injection well to achieve precise addition of leaching agent.

[0053] The working principle of this invention is as follows: First, real-time data is collected on the groundwater level, leaching agent concentration in the injection pipeline, rare earth mother liquor flow rate in the collection pipeline, and spatial coordinates and geological stratification data of the mine during the in-situ leaching process of ion-adsorption rare earth ore. Then, based on the collected groundwater level, leaching agent concentration, and geological stratification data, and according to seepage mechanics and constitutive relations of ion exchange reactions, the seepage field and ion concentration field inside the mine at the current moment are calculated, and a leaching state vector field corresponding to the physical mine state is output. Finally, based on this leaching state vector field, the pre-stored optimal injection template for the mine is geometrically and parametrically corrected to generate a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state. The figure shows the following steps: Next, using dynamic injection volume distribution and dynamic injection concentration distribution maps, deviation analysis is performed on the actual injection parameters of each injection well collected in real time. Areas with insufficient or excessive injection are extracted, and a binary map of the leaching anomaly area is output. Then, this binary map of the leaching anomaly area is spatially overlaid with the mine's grade distribution map and permeability zoning map to analyze the spatial topological relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body. Finally, based on the spatial topological relationship, it is determined whether the leaching anomaly area will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage. Based on the judgment results, solenoid valve opening commands and variable frequency pump speed commands are generated for each injection well to achieve precise addition of the leaching agent.

[0054] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A digital twin-based intelligent control method for in-situ leaching of rare earth minerals, characterized in that, Includes the following steps: S1 is used to collect in real time the groundwater level, the concentration of leaching agent in the injection pipeline, the flow rate of rare earth mother liquor in the collection pipeline, and the spatial coordinates and geological stratification data of the mine during the in-situ leaching process of ion-adsorption rare earth ore. S2, based on groundwater level, leaching agent concentration and geological stratification data, calculates the seepage field and ion concentration field inside the mine at the current moment according to seepage mechanics and constitutive relations of ion exchange reaction, and outputs the leaching state vector field corresponding to the physical mine state; S3, based on the leaching state vector field, perform geometric and parameter correction on the pre-stored optimal injection template for the mine, and generate a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state; S4. Using dynamic injection volume distribution map and dynamic injection concentration distribution map, deviation analysis is performed on the actual injection parameters of each injection well collected in real time, and the areas of insufficient injection and excessive injection are extracted, and a binary map of the leaching abnormal area is output. S5. Spatially overlay the binary map of the leaching anomaly area with the grade distribution map and permeability zoning map of the mine to analyze the spatial topological relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body. S6, based on the spatial topology, determines whether abnormal injection areas will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage, and generates solenoid valve opening commands and variable frequency pump speed commands for each injection well based on the judgment results, so as to achieve precise addition of leaching agent.

2. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 1, characterized in that, S2 specifically includes: Based on the groundwater level data from each observation well, a three-dimensional distribution map of groundwater equipotential surfaces inside the mine was constructed. The pressure gradient is calculated along the normal direction of the equipotential surface, and the seepage velocity components are solved grid by grid according to Darcy's law to form the seepage velocity vector field. The seepage velocity vector field is spatially matched with the measured values ​​of the leaching agent concentration in each injection well. The convection-diffusion equation of the ion exchange reaction is solved iteratively along the streamline direction to update the ion concentration value of each grid. The percolation velocity vector field and the updated ion concentration field are encapsulated together as a leaching state vector field output.

3. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 2, characterized in that, The formation of the seepage velocity vector field specifically includes: The three-dimensional groundwater equipotential surface is discretized into a set of triangular patches, and the centroid coordinates and unit normal vector determined by the isopleth distribution of pressure values ​​are extracted for each patch. Along the unit normal vector direction of each patch, extend forward and backward by one grid step each, obtain the pressure values ​​at the two extension endpoints, and use the central difference method to calculate the pressure gradient magnitude at the corresponding patch. The pressure gradient magnitude is multiplied by the permeability coefficient of the grid and then divided by the dynamic viscosity coefficient. The resulting scalar value is projected along the unit normal direction to generate the seepage velocity vector component of each patch. All patch components are then combined to form the seepage velocity vector field.

4. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 1, characterized in that, S3 specifically includes: Extract the seepage velocity amplitude and ion concentration gradient direction at each injection well location from the leaching state vector field, and read the preset injection volume and preset concentration at the corresponding location in the optimal injection template for the mine. The ratio of the seepage velocity amplitude to the target seepage velocity threshold is used as the injection volume correction coefficient. At the same time, the cosine of the angle between the ion concentration gradient direction and the injection diffusion direction is calculated, and the increase or decrease correction amount of the preset concentration is determined according to the sign of the cosine of the angle. The injection volume correction coefficient and concentration increase / decrease correction of each injection well are spatially distributed to the surrounding un-welled area to generate continuous dynamic injection volume distribution map and dynamic injection concentration distribution map respectively.

5. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 4, characterized in that, The generation of continuous dynamic injection volume distribution map and dynamic injection concentration distribution map specifically includes: With each injection well as the center, an elliptical influence domain is constructed according to the anisotropic radius set in the mine permeability zoning map, and the correction coefficient and correction amount at the injection well are assigned as the center value of the ellipse. Within the elliptical influence domain, the decay rate is calculated along the major and minor axes, and weight values ​​are assigned based on the normalized Euclidean distance from each grid point to the center point within the elliptical domain. The weight values ​​of the overlapping elliptical influence domains of multiple injection wells are normalized and accumulated. The accumulated weight values ​​are then used to perform a weighted average of each correction coefficient and correction amount to generate a continuous distribution map of the entire mining area.

6. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 1, characterized in that, S4 specifically includes: The injection volume deviation value is obtained by subtracting the actual injection volume collected in real time from the value of the dynamic injection volume distribution map of each injection well. At the same time, the injection concentration deviation value is obtained by subtracting the actual injection concentration collected in real time from the value of the dynamic injection concentration distribution map. Set injection volume deviation threshold and injection concentration deviation threshold respectively. Mark the grid where the injection volume deviation value is lower than the negative threshold and the injection concentration deviation value is also lower than the negative threshold as the injection insufficient area, and mark the grid where both deviation values ​​are higher than the positive threshold as the injection excessive area. The grids in the insufficient and excessive injection areas are assigned a value of 1, and the remaining grids are assigned a value of 0, outputting a binarized map of the leaching anomaly area.

7. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 1, characterized in that, S5 specifically includes: The insufficient injection area in the binary map of the leaching anomaly area is spatially intersected with the high-grade area in the grade distribution map and the low-permeability area in the permeability zoning map to extract the overlapping grid set of insufficient injection and high grade and the overlapping grid set of insufficient injection and low permeability. Calculate the proportion of the overlapping grid area of ​​the insufficiently injected high-grade fluid to the total area of ​​the insufficiently injected region, and at the same time calculate the proportion of the overlapping grid area of ​​the insufficiently injected low-permeability fluid to the total area of ​​the insufficiently injected region. When the proportion of high-grade water is greater than the first threshold and the proportion of low-penetration water is less than the second threshold, it is determined to be a high-priority topology; when the proportion of low-penetration water is greater than the second threshold, it is determined to be a non-replenishable topology.

8. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 1, characterized in that, S6 specifically includes: When a high-priority injection topology relationship is determined, the target injection volume and target concentration of the corresponding overlapping grid are extracted from the leaching state vector field, and the current actual injection volume and actual concentration of the injection well are subtracted to obtain the injection volume increment and concentration increment, respectively. Input the injection volume increment into the pressure-opening characteristic curve of the pipeline where the injection well is located, and look up the increase in the opening angle of the solenoid valve. At the same time, divide the concentration increment by the current storage concentration of the mother liquor tank to obtain the speed increase ratio of the variable frequency pump. The current solenoid valve opening degree and the increase in opening angle are added together to form the opening degree command output. The current variable frequency pump speed is multiplied by 1 and then added to the speed increase ratio to form the speed command output.

9. The intelligent control method for in-situ leaching of rare earth minerals based on digital twins according to claim 8, characterized in that, The method of obtaining the increase in the opening angle of the solenoid valve specifically includes: Real-time pressure values ​​are read from the inlet and outlet ends of the pipeline where the injection well is located, and the difference between the two is calculated as the current working pressure difference. The target curve family in the pressure-opening characteristic curve is located using the current working pressure difference as an index. On the target curve family, starting from the actual opening angle of the current solenoid valve, move the distance along the horizontal axis by a distance equal to the change in angle corresponding to the increase in liquid injection volume divided by the cross-sectional area of ​​the pipe and then divided by the flow rate conversion coefficient, and read whether the pressure difference on the vertical axis corresponding to the change in angle is consistent with the current working pressure difference. If they are consistent, the angle change is directly used as the increase in the opening angle. If they are inconsistent, the angle change is corrected until the calculated pressure difference is equal to the current working pressure difference, and the final corrected increase in the opening angle is output.

10. A digital twin-based intelligent control system for in-situ leaching of rare earth minerals, characterized in that, A method for implementing the intelligent control method for in-situ leaching of rare earth minerals based on digital twins as described in any one of claims 1-9, comprising: The data synchronization acquisition module is used to collect in real time the groundwater level, the concentration of leaching agent in the injection pipeline, the flow rate of rare earth mother liquor in the collection pipeline, and the spatial coordinates and geological stratification data of the mine during the in-situ leaching process of ion-adsorption rare earth ore. The physical process calculation module, based on groundwater level, leaching agent concentration and geological stratification data, and according to seepage mechanics and constitutive relations of ion exchange reaction, calculates the seepage field and ion concentration field inside the mine at the current moment, and outputs the leaching state vector field corresponding to the physical mine state. The dynamic injection map generation module performs geometric and parameter correction on the pre-stored optimal injection template of the mine based on the leaching state vector field, and generates a dynamic injection volume distribution map and a dynamic injection concentration distribution map that match the current leaching state. The leaching deviation identification module uses dynamic injection volume distribution map and dynamic injection concentration distribution map to perform deviation analysis on the actual injection parameters of each injection well collected in real time, extract the insufficient injection area and the excessive injection area, and output a binary map of the leaching abnormal area. The spatial topology analysis module spatially overlays the binary map of the leaching anomaly area with the grade distribution map and permeability zoning map of the mine to analyze the spatial topology relationship between the leaching anomaly area and the high-grade and low-permeability areas of the ore body. The control command generation module determines whether abnormal injection areas will lead to a decrease in rare earth leaching rate or an increase in the risk of leaching agent leakage based on spatial topology. Based on the determination results, it generates solenoid valve opening commands and variable frequency pump speed commands for each injection well to achieve precise addition of leaching agent.

Images