Intelligent separation method and device for co-associated minerals

Abstract

The embodiment of the specification discloses an intelligent separation method and device of a co-associated ore, relates to the technical field of ore separation, and is used for solving the problem of poor separation and identification effect. The method comprises the following steps: collecting X-ray energy spectrum data of the co-associated ore to be separated to obtain the equivalent thickness of at least two base substances; a two-dimensional feature space of equivalent atomic number and surface density in the co-associated ore to be separated is constructed based on the equivalent thickness of the at least two base substances; the distribution cluster form of each co-associated ore to be separated in the two-dimensional feature space is obtained to dynamically extract a dynamic reference endpoint; the projection position of each co-associated ore to be separated on the connecting line of the dynamic reference endpoint is determined to determine the continuous grade estimation of the target ore in the co-associated ore to be separated according to the projection position; and the time sequence control signal corresponding to the co-associated ore to be separated is generated according to the continuous grade estimation to control the separation of the target ore in the co-associated ore to be separated to the corresponding collection area.

Description

Technical Field

[0001] This specification relates to the field of ore separation technology, and in particular to an intelligent separation method and equipment for associated minerals. Background Technology

[0002] Associated minerals refer to the coexistence of two or more industrially valuable mineral resources in the same deposit, such as lead-zinc ore, copper-molybdenum ore, and rare earth iron ore. These mineral resources are characterized by low grade of valuable minerals, fine grain size, and complex mineral composition, which poses a significant challenge to mineral processing and separation.

[0003] Currently, the separation of traditional associated and co-existing ores generally employs processes such as gravity separation, magnetic separation, and flotation. These separation processes require crushing the ore to liberate valuable minerals from gangue minerals. However, due to the fine particle size of valuable minerals and the mutual encapsulation of various minerals in associated and co-existing ores, fine grinding or even ultrafine grinding is often required to achieve liberation, resulting in high energy consumption and high costs. With the development of technology, intelligent separation technology based on X-ray transmission has been applied in the field of ore pre-selection and waste disposal. This technology uses X-rays to irradiate the ore and identifies and separates valuable ores from waste rock based on the differences in X-ray absorption capacity of different minerals. Existing X-ray separation technologies mainly include single-energy X-ray separation and dual-energy X-ray separation. However, existing intelligent X-ray separation technologies lack the ability to continuously and quantitatively identify ore grades, and cannot adapt to the characteristics of large fluctuations in composition and continuous changes in grade in associated and co-existing ores. Summary of the Invention

[0004] To address the aforementioned problems, one or more embodiments of this specification provide an intelligent separation method and apparatus for associated minerals.

[0005] One or more embodiments of this specification employ the following technical solutions: This specification provides one or more embodiments of a smart separation method for associated minerals, the method comprising: X-ray energy spectrum data of the associated minerals to be separated are collected, and the X-ray energy spectrum data are subjected to matrix material decomposition processing to obtain the equivalent thickness of at least two matrix materials; wherein, the X-ray energy spectrum data includes attenuation information of the high-energy spectral band and the low-energy spectral band. Based on the equivalent thickness of the at least two base materials, a two-dimensional feature space of the equivalent atomic number and surface density of the co-associated ore to be separated is constructed. The distribution cluster morphology of each of the associated minerals to be separated is obtained in the two-dimensional feature space to dynamically extract the dynamic reference endpoints of each of the associated minerals to be separated; wherein, the dynamic reference endpoints include: a first reference endpoint representing high-grade minerals and a second reference endpoint representing low-grade gangue. The projection position of each of the associated ores to be separated on the line connecting the dynamic reference endpoints is determined, so as to determine the continuous grade estimate of the target mineral in the associated ores to be separated based on the projection position; Based on the continuous grade estimate, a timing control signal for the spray valve corresponding to the associated ore to be separated is generated, and the target ore in the associated ore to be separated is separated into the corresponding collection area by controlling the timing control signal for the spray valve.

[0006] Optionally, in one or more embodiments of this specification, X-ray energy dispersive spectroscopy data of the associated ore to be separated are collected, specifically including: The associated ore to be separated is irradiated by an X-ray emission source, and the X-ray signal transmitted through the associated ore to be separated is collected by a detector array. Based on the X-ray signal, obtain X-ray energy spectrum data containing attenuation information for different energy ranges.

[0007] Optionally, in one or more embodiments of this specification, the X-ray energy spectrum data is subjected to matrix material decomposition processing to obtain the equivalent thickness of at least two matrix materials, specifically including: Based on the X-ray energy spectrum data, the low-energy spectrum transmission intensity value and the high-energy spectrum transmission intensity value of the associated ore to be separated are determined. Based on the transmission intensity values ​​of the low-energy spectral band and the high-energy spectral band, a transmission signal pair of the associated minerals to be separated is constructed. The matching results are obtained by matching the transmission signal pairs with the existing hierarchical fingerprint database; wherein, the existing hierarchical fingerprint database is constructed by hierarchically storing the multidimensional feature fingerprints of the standard samples after calibration tests based on various standard samples with known thickness and known composition; If the confidence level corresponding to the matching result is higher than the preset first threshold, then the equivalent thickness of at least two base materials is obtained. If the confidence level corresponding to the matching result is lower than a preset first threshold, the neighboring multidimensional feature fingerprint corresponding to the transmission signal pair is obtained, and the equivalent thickness of at least two base materials is obtained based on the interpolation fitting of the neighboring multidimensional feature fingerprint and the multidimensional feature fingerprint corresponding to the matching result.

[0008] Optionally, in one or more embodiments of this specification, a two-dimensional feature space of the equivalent atomic number and areal density of the co-existing ore to be separated is constructed based on the equivalent thickness of the at least two base materials, specifically including: Based on the equivalent atomic numbers corresponding to the at least two base substances, each of the at least two base substances is divided into a first type of base substance and a second type of base substance; wherein, the first type of base substance represents a base substance with a low atomic number component content, and the second type of base substance represents a base substance with a high atomic number component content; A two-dimensional rectangular coordinate system is constructed with the equivalent thickness of the first type of base material as the abscissa and the equivalent thickness of the second type of base material as the ordinate. The equivalent thicknesses of at least two base materials of each of the associated minerals to be separated are mapped to the horizontal and vertical coordinates of the two-dimensional rectangular coordinate system to obtain the feature points of the associated minerals to be separated. The feature points of each of the associated minerals to be separated within the preset acquisition time window are aggregated to obtain a two-dimensional feature space distribution map, and the two-dimensional feature space of the equivalent atomic number and surface density of the associated minerals to be separated is obtained based on the two-dimensional feature space distribution map.

[0009] Optionally, in one or more embodiments of this specification, obtaining the distribution cluster morphology of each of the associated ores to be separated in the two-dimensional feature space specifically includes: Based on the two-dimensional feature space distribution map, the distribution density of feature points of each of the associated and co-existing ores to be separated within the preset acquisition time window is obtained; Clustering parameters are determined based on the distribution density, and the feature points are clustered based on the clustering parameters to obtain multiple cluster groups; Extract the distribution cluster morphology corresponding to each of the clusters; wherein, the distribution cluster morphology includes: the centroid position of each cluster, the boundary contour of the cluster, and the density distribution characteristics of feature points within the cluster.

[0010] Optionally, in one or more embodiments of this specification, dynamically extracting the dynamic reference endpoints of each of the associated ores to be separated specifically includes: Based on the pre-defined mineral composition characteristics of the target ore, or the separation records of historical associated ores to be separated, identify the target cluster containing the target ore and the background cluster. The relative positional relationship between the target cluster and the background cluster is obtained. Based on the continuous direction between the target cluster and the background cluster, the distribution density abrupt change points of the target cluster are searched to obtain critical feature points. Based on pre-set confidence constraints, the critical feature points are screened to obtain the dynamic reference endpoints of the associated ores to be separated.

[0011] Optionally, in one or more embodiments of this specification, the projected position of each of the associated ores to be separated on the line connecting the dynamic reference endpoints is determined, so as to determine the continuous grade estimate of the target mineral in the associated ores to be separated based on the projected position, specifically including: Based on the pre-defined mineral composition characteristics of the target ore, or the separation records of historical associated ores to be separated, determine the high-grade endpoint; A grade mapping reference axis is constructed based on the connection between the dynamic reference endpoint and the high-grade endpoint; Obtain the characteristic points of each of the associated ores to be separated, and their vertical projection points on the grade mapping reference axis; Obtain a first distance between the vertical projection point and the high-quality endpoint, and a second distance between the vertical projection point and the dynamic reference point; The continuous grade estimate of the target mineral in each of the associated ores to be separated is determined by the proportional relationship between the first distance and the second distance; wherein the continuous grade estimate is negatively correlated with the first distance and positively correlated with the second distance.

[0012] Optionally, in one or more embodiments of this specification, a timing control signal for the spray valve corresponding to the associated ore to be separated is generated based on the continuous grade estimate, specifically including: Compare the estimated continuous grade with the preset grade threshold; If the estimated continuous grade is greater than or equal to the preset grade threshold, then it is determined that the associated ore to be separated belongs to the target ore, and a corresponding valve opening signal is generated. If the estimated continuous grade is less than the preset grade threshold, it is determined that the separated associated ore does not belong to the target ore, and based on the corresponding valve opening state, it is determined whether to generate a valve closing signal. The timing sequence of the valve opening signal and the valve closing signal is sorted to obtain the valve timing control signal.

[0013] Optionally, in one or more embodiments of this specification, the target ore in the associated ore to be separated is separated to the corresponding collection area by controlling the timing control signal of the spray valve, specifically including: The timing control signal of the injection valve is sent to the injection valve array so that the injection valve array can start the corresponding injection valve according to the timing control signal of the injection valve; The directional airflow generated by the high-pressure gas from the spray valve blows the target ore away from its original trajectory from the original ore flow and separates it into the corresponding collection area.

[0014] This specification provides one or more embodiments of an intelligent separation device for associated minerals, the device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform any of the methods described above.

[0015] The above-described at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects: By constructing a two-dimensional feature space based on the equivalent thickness of at least two base materials, equivalent atomic number and areal density are used to decouple the composition and thickness information of the associated ores to be separated, eliminating the interference of particle size variation on grade judgment. Dynamic reference endpoints are dynamically extracted through cluster analysis, enabling the sorting strategy to adapt to changes in ore properties without manual recalibration. Furthermore, by determining the projection position of each associated ore to be separated on the line connecting the dynamic reference endpoints, continuous grade estimates of the target minerals in the associated ores are determined based on these projection positions, achieving continuous quantification of ore grade and providing an accurate data foundation for subsequent separation. Based on the comparison between the grade estimate and a preset threshold, a timing control signal for the spray valve is generated to separate the target ore, achieving precise separation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a schematic flowchart of an intelligent separation method for associated minerals provided in the embodiments of this specification; Figure 2 This is a schematic diagram of the structure of an intelligent separation device for associated minerals provided in the embodiments of this specification. Detailed Implementation

[0017] This specification provides an intelligent separation method and device for associated minerals.

[0018] like Figure 1 As shown, this specification provides a flowchart illustrating an intelligent separation method for associated minerals. Based on... Figure 1 As can be seen, in one or more embodiments of this specification, a method for intelligent separation of associated minerals includes the following steps: S101: Collect X-ray energy spectrum data of the associated ore to be separated, and perform matrix material decomposition processing on the X-ray energy spectrum data to obtain the equivalent thickness of at least two matrix materials; wherein, the X-ray energy spectrum data includes attenuation information of the high-energy spectral band and the low-energy spectral band.

[0019] Associated ores refer to ores in the same deposit that, in addition to the main target mineral being mined, also contain one or more other mineral resources with industrial utilization value.

[0020] X-ray energy spectrum data is the number of transmitted photons in different energy ranges when a beam of X-rays containing multiple energies irradiates a mineral particle. Different energy X-rays interact with different elements in the mineral, some energy is absorbed or scattered, and the remaining energy penetrates the mineral and reaches the detector. The detector records the number of transmitted photons in different energy ranges.

[0021] After crushing, the associated ores are separated into single-layer, single-column, non-overlapping associated ores. These ores are then fed to an X-ray detection area at a preset speed via a vibrating feeder. X-ray energy spectrum data of the associated ores is collected in this area, simultaneously acquiring transmission intensity values ​​in both low and high energy spectrum bands, forming attenuation information for both high and low energy spectrum bands. Since the associated ores, in addition to the primary target mineral, also contain one or more other minerals, they are considered as a mixture of two or more base materials with different atomic numbers in a certain proportion. Based on two or more pre-selected standard materials, the equivalent thickness of at least two base materials can be obtained by analyzing the X-ray energy spectrum data. By decomposing the base material, the mixed attenuation signal of the ore particles in the associated ore to be separated is decomposed into a linear superposition of several base material contribution components, thus realizing the conversion from mixed signal to component characteristic quantity.

[0022] Specifically, in one or more embodiments of this specification, X-ray energy dispersive spectroscopy data of the associated ore to be separated are collected, including: After the associated ore to be separated is transported to the X-ray detection area, it is irradiated by an X-ray emission source. An array of detectors, positioned opposite the X-ray emission source, collects the X-ray signals transmitted through the ore. The X-ray emission source can be a tungsten target X-ray tube or other X-ray generator suitable for the mining environment; no specific limitations are imposed here. The detector array can be a linear cadmium telluride photon counting detector or a scintillator detector. Taking a cadmium telluride photon counting detector as an example, the detector divides the received X-ray photons into several energy channels according to their energy levels. Low-energy channels record the transmission intensity values ​​of the low-energy spectral band, and high-energy channels record the transmission intensity values ​​of the high-energy spectral band. Therefore, by normalizing the transmittance of each spectral band and taking the negative logarithm of the X-ray signal, X-ray energy spectrum data containing attenuation information for different energy ranges can be obtained.

[0023] Specifically, in one or more embodiments of this specification, the above-described process of decomposing X-ray energy spectral data into base material to obtain the equivalent thickness of at least two base materials includes the following steps: Based on X-ray energy dispersive spectroscopy (EDS) data, the low-energy and high-energy transmission intensity values ​​of the associated ore to be separated were determined. Then, the low-energy and high-energy transmission intensity values ​​were recorded to construct a transmission signal pair for the associated ore to be separated.

[0024] Then, the obtained transmission signal pairs are matched with the existing graded fingerprint database to obtain matching results. The existing graded fingerprint database is a database constructed by hierarchically storing the multidimensional feature fingerprints of standard samples after calibration experiments on standard samples with known thicknesses and compositions. That is, these standard samples have known compositions and known thicknesses, where the composition includes at least two base materials, and the thickness of each base material varies in a gradient within a preset range. By combining standard samples with different thicknesses and compositions, a calibration sample set covering the expected range of ore compositions is formed. Then, each standard sample in the calibration sample set is subjected to X-ray detection to obtain its transmission intensity values ​​in the low-energy and high-energy spectral bands, forming a transmission signal pair for that standard sample. Simultaneously, the corresponding base material thickness combination for that standard sample is recorded. The transmission signal pair of each standard sample is associated and stored with its corresponding base material thickness combination to form a fingerprint record. The above operation is repeated for all standard samples to obtain a fingerprint record set. The fingerprint record set is stored hierarchically to construct a hierarchical fingerprint database. For example, the fingerprint database can be divided into a fast matching layer and a precise matching layer. The fast matching layer stores representative fingerprint records sampled at relatively coarse thickness intervals for fast coarse matching, while the precise matching layer stores the entire set of fingerprint records for fine matching.

[0025] For each pair of transmission signals from associated minerals to be separated, the fingerprint record with the highest similarity is selected as the candidate matching result, and the confidence level of the matching result is calculated. If the confidence level of the matching result is higher than a preset first threshold, the matrix material thickness combination stored in the corresponding candidate matching result is directly read to obtain the equivalent thickness of at least two matrix materials. The confidence level can be determined by weighted summation based on the distance between the real-time transmission signal pair and the candidate fingerprint record, as well as the density distribution of fingerprint records surrounding the candidate fingerprint record.

[0026] If the confidence level corresponding to the matching result is lower than a preset first threshold, then the neighboring multidimensional feature fingerprints corresponding to the transmission signal pair are obtained. Based on the interpolation fitting of the neighboring multidimensional feature fingerprints and the multidimensional feature fingerprints corresponding to the matching result, the equivalent thicknesses of at least two matrix materials are obtained. That is, the transmission signal pairs of these neighboring fingerprint records and their corresponding matrix material thickness combinations are used as known sample points, and the matrix material thickness combinations corresponding to the real-time transmission signal pairs are calculated using interpolation fitting. Various interpolation fitting methods can be used. For example, the inverse distance weighted interpolation method can be used, where the contribution of neighboring fingerprint records to the interpolation result is inversely proportional to the square of their distance to the real-time transmission signal pair. Alternatively, the radial basis function interpolation method can be used, constructing an interpolation surface with neighboring fingerprint records as nodes, etc.

[0027] S102: Based on the equivalent thickness of the at least two base materials, construct a two-dimensional feature space of the equivalent atomic number and surface density of the co-existing ore to be separated.

[0028] After obtaining the equivalent thickness of at least two base materials of the associated ore to be separated based on the above step S101, in order to solve the problem of inaccurate grade judgment caused by the coupling of composition and thickness information in traditional X-ray sorting technology, the embodiments of this specification construct a two-dimensional feature space of equivalent atomic number and areal density of the associated ore to be separated based on the equivalent thickness of at least two base materials.

[0029] Specifically, in one or more embodiments of this specification, a two-dimensional feature space of the equivalent atomic number and areal density of the co-existing ore to be separated is constructed based on the equivalent thickness of the at least two base materials, specifically including: After obtaining the equivalent thicknesses of at least two base materials, each of the at least two base materials is divided into a first type of base material and a second type of base material based on their corresponding equivalent atomic numbers. The first type of base material represents the base material with a low atomic number content, used to characterize the content of low atomic number components in ore particles. The second type of base material represents the base material with a high atomic number content, used to characterize the content of high atomic number components in ore particles. It should be noted that when there are exactly two base materials, the first type and the second type of base material correspond to these two base materials respectively. When there are more than two base materials, multiple low atomic number base materials can be grouped into a first type of base material group, and their equivalent thicknesses can be weighted and summed or vectored to obtain the comprehensive equivalent thickness of the first type of base material. Similarly, multiple high atomic number base materials can be grouped into a second type of base material group, and their equivalent thicknesses can be aggregated to obtain the comprehensive equivalent thickness of the second type of base material.

[0030] After completing the above classification, a two-dimensional rectangular coordinate system is constructed with the equivalent thickness of the first type of matrix material as the abscissa and the equivalent thickness of the second type of matrix material as the ordinate. It can be understood that the abscissa of this coordinate system reflects the relative content of low atomic number components in the ore particles, and the ordinate reflects the relative content of high atomic number components in the ore particles. The equivalent thicknesses of at least two matrix materials of each associated ore to be separated are mapped to the abscissa and ordinate of the two-dimensional rectangular coordinate system according to the above matrix material type classification results, respectively, to obtain points in the constructed two-dimensional rectangular coordinate system, which serve as characteristic points of the associated ore to be separated.

[0031] To obtain statistically significant ore distribution characteristics, it is necessary to aggregate feature points from multiple ore particles collected within a certain time window. Specifically, in this embodiment, feature points from each associated ore to be separated within a preset collection time window are aggregated to obtain a two-dimensional feature space distribution map. Based on this two-dimensional feature space distribution map, the two-dimensional feature space of the equivalent atomic number and areal density of the associated ore to be separated is obtained. The preset collection time window can be adjusted according to the actual processing volume and real-time requirements. All feature points are superimposed or stored in this coordinate system within the preset collection time window. When the time window ends, the coordinate system contains all feature points from the associated ore to be separated within that time window, forming a feature point set, i.e., a two-dimensional feature space distribution map. Based on this distribution map, the two-dimensional feature space of the equivalent atomic number and areal density of the associated ore to be separated can be reflected.

[0032] S103: Obtain the distribution cluster morphology of each of the associated minerals to be separated in the two-dimensional feature space, so as to dynamically extract the dynamic reference endpoints of each of the associated minerals to be separated; wherein, the dynamic reference endpoints include: a first reference endpoint representing high-grade minerals and a second reference endpoint representing low-grade gangue.

[0033] To address the problem in existing technologies where fixed reference endpoints cannot adapt to dynamic changes in the properties of associated ores, this embodiment, after obtaining the associated ores to be separated in the two-dimensional feature space, dynamically extracts dynamic reference endpoints for each ore based on the distribution cluster morphology of the clusters within that space. These dynamic reference endpoints include a first reference endpoint representing high-grade minerals and a second reference endpoint representing low-grade gangue. In other words, by performing cluster analysis in the two-dimensional feature space of the current batch, high-grade mineral clusters and low-grade gangue clusters are automatically identified, and their respective cluster centers are extracted as dynamic reference endpoints. This process achieves adaptive updating of the reference endpoints, allowing subsequent grade calculations to automatically adjust to ore composition fluctuations without manual intervention, thereby significantly improving the intelligence of associated ore separation.

[0034] Specifically, in one or more embodiments of this specification, obtaining the distribution cluster morphology of each associated ore to be separated in a two-dimensional feature space includes: Based on the two-dimensional feature space distribution map, the distribution density of feature points of each associated ore to be separated within a preset acquisition time window is obtained. Specifically, for each feature point in the two-dimensional feature space distribution map, a circular neighborhood is defined with that point as the center and a preset radius. The number of other feature points contained within this neighborhood is counted, and this number is used as the local density value of that feature point. After traversing all feature points, the distribution density of feature points of each associated ore to be separated within the preset acquisition time window can be obtained. The preset radius can be determined based on a fixed empirical value. Alternatively, a grid-based method can be used, dividing the two-dimensional feature space into several grid cells of equal size, and counting the number of feature points in each grid cell as the density value of that grid.

[0035] Since the choice of clustering parameters directly affects the quality of clustering results, clustering parameters are determined based on the obtained distribution density. This allows for clustering of feature points based on these parameters, resulting in multiple clusters. In one implementation, the process of determining clustering parameters based on distribution density involves: statistically analyzing the local density values ​​of all feature points, calculating the distribution characteristics of these density values, using the neighborhood radius corresponding to a certain quantile of the density value as the cluster neighborhood radius, and using the number of feature points within the neighborhood corresponding to the 10th percentile of the density value distribution as the minimum number of points. This adaptive parameter determination method ensures that the clustering parameters match the actual distribution characteristics of the current batch of feature points, avoiding clustering failure due to fixed parameters. In another implementation, the process of determining clustering parameters based on distribution density involves: traversing preset parameter combinations, evaluating each parameter combination for clustering, and selecting the parameter combination with the optimal evaluation index as the final clustering parameters.

[0036] After determining the clustering parameters, the system performs clustering analysis on the feature point set based on these parameters, obtaining multiple clusters. Then, the distribution morphology corresponding to each cluster is extracted. The distribution morphology is a quantitative description of the geometric and statistical characteristics of the clusters, including: the centroid position of each cluster, the boundary contour of the clusters, and the density distribution characteristics of the feature points within the clusters. The centroid position of each cluster is the geometric center of all feature points in that cluster, reflecting the average position of the cluster in the two-dimensional feature space. The abscissa of the centroid is the average of the abscissas of all feature points within the cluster, and the ordinate of the centroid is the average of the ordinates of all feature points within the cluster. The boundary contour of the cluster describes the distribution range of the cluster in the two-dimensional feature space. In a certain scenario, the boundary of the convex hull of all feature points within the cluster is determined by calculating the convex hull, which is the boundary contour of the cluster. The density distribution characteristics of the feature points within the cluster correspond to the position of the density peak points, the direction of the density gradient, and the distribution of density contour lines within the cluster.

[0037] Specifically, reference points representing high-grade and low-grade endmembers in the current batch of ore are determined to provide a benchmark for subsequent grade mapping reference axis construction and projection calculation. In one or more embodiments of this specification, dynamic reference endpoints of each associated ore to be separated are dynamically extracted, specifically including: Based on the above cluster analysis, multiple clusters of associated and co-existing ores to be separated within the current acquisition time window have been obtained in the two-dimensional feature space. These clusters represent ore groups of different grades. Among them, clusters with higher equivalent atomic numbers correspond to high-grade ores, clusters with lower equivalent atomic numbers correspond to low-grade gangue, and clusters in between correspond to intergrowths of different grades. In order to extract dynamic reference endpoints, based on the pre-set mineral composition characteristics of the target ore or the historical separation records of associated and co-existing ores to be separated, target clusters containing the target ore and background clusters were identified. It should be noted that the identification method of the target clusters is as follows: the pre-stored mineral composition characteristics of the target ore are compared with the equivalent atomic number of the centroid of each cluster, and the clusters whose centroids fall within this range are identified as target clusters. If multiple clusters fall within this range, the cluster with the highest equivalent atomic number is selected as the target cluster. Alternatively, target clusters can be identified based on historical sorting records. This involves recording the distribution characteristics of ore particles successfully separated into the concentrate collection area during historical sorting processes in a two-dimensional feature space. These distribution characteristics are then used as feature templates. The distribution characteristics of each current cluster are compared with these templates, and the cluster with the highest matching degree is identified as the target cluster. Background clusters are identified by selecting the cluster with the lowest equivalent atomic number from among multiple clusters. This background cluster represents the lowest-grade ore group in the current batch, primarily composed of gangue.

[0038] The relative positional relationship between the target cluster and the background cluster is obtained. Based on the continuous direction of the target and background clusters, the distribution density abrupt change points of the target cluster are searched to obtain critical feature points. Specifically, if the relative positional relationship between the target and background clusters changes with ore grade from low to high, the main trend direction of the feature point's movement in the two-dimensional feature space is identified. In a certain example, if the centroids of the background and target clusters are (2, 1) and (6, 8) respectively, the continuous direction vector is (4, 7), indicating that from low-grade gangue to high-grade minerals, the equivalent thickness of both the first and second type matrix materials increases, with a greater increase in the ordinate. This is consistent with the physical expectation that high-grade minerals are rich in high atomic number components. After determining the continuous direction, the boundary region of the target cluster is searched along this direction to obtain the distribution density abrupt change points. It can be understood that distribution density mutation points refer to feature points at the edge of the target cluster where the distribution density of feature points changes significantly. These points represent the transition boundary from the interior of the high-density cluster to the sparse outer region.

[0039] After obtaining the critical feature points, they are screened based on pre-set confidence constraints to obtain dynamic reference endpoints for the associated ores to be separated. The pre-set confidence constraints are used to evaluate the reliability of each critical feature point as a dynamic reference endpoint, ensuring the representativeness and stability of the selected reference endpoints. These pre-set confidence constraints include, but are not limited to, boundary clarity constraints, local density constraints, directional consistency, and historical stability constraints comparing the positions of the current batch's critical feature points with the positions of the dynamic reference endpoints in historical batches. In this embodiment, the dynamic reference endpoints include a first reference endpoint and a second reference endpoint. The first reference endpoint represents a high-grade mineral reference point, which can be the critical feature point with the highest confidence on the boundary of the target cluster. The second reference endpoint represents a low-grade gangue reference point, which can be extracted from the boundary of the background cluster or directly using the centroid of the background cluster.

[0040] S104: Determine the projection position of each of the associated ores to be separated on the line connecting the dynamic reference endpoints, so as to determine the continuous grade estimate of the target mineral in the associated ores to be separated based on the projection position.

[0041] Specifically, in one or more embodiments of this specification, the projected position of each associated ore to be separated on the line connecting the dynamic reference endpoints is determined, so as to determine the continuous grade estimate of the target mineral in the associated ore to be separated based on the projected position, specifically including: Based on the pre-defined mineral composition characteristics of the target ore, or the separation records of associated ores to be separated in history, the high-grade endpoint is determined. This involves pre-storing the theoretical or measured characteristic values ​​of the target minerals in the target ore, using these theoretical coordinates as the high-grade endpoint, or statistically analyzing the coordinate distribution in a two-dimensional feature space of the multiple ores with the highest grades successfully separated to the concentrate collection area during historical sorting processes, using the centroid coordinates or density peak coordinates of these ores as the high-grade endpoint.

[0042] A grade mapping baseline axis is constructed by connecting the dynamic reference endpoint and the high-grade endpoint. This baseline axis starts at the low-grade endpoint and ends at the high-grade endpoint, representing the direction of grade increase from low to high. For characteristic points of each associated ore to be separated, its vertical projection point on the grade mapping baseline axis is calculated. Here, the vertical projection point represents the equivalent position of the ore particle in the grade dimension. After obtaining the vertical projection point, the first distance between the vertical projection point and the high-grade endpoint, and the second distance between the vertical projection point and the dynamic reference point can be obtained. The continuous grade estimate of the target mineral in each associated ore to be separated is determined by the proportional relationship between the first distance and the second distance. The continuous grade estimate is negatively correlated with the first distance and positively correlated with the second distance; the closer the projection point is to the high-grade endpoint, the higher the grade estimate, and the closer it is to the low-grade endpoint, the lower the grade estimate. The method for determining the continuous grade estimate of the target mineral in each associated ore to be separated includes, but is not limited to, the linear proportional method.

[0043] S105: Based on the continuous grade estimate, generate a valve timing control signal corresponding to the associated ore to be separated, and separate the target ore in the associated ore to be separated to the corresponding collection area through the control of the valve timing control signal.

[0044] After completing the continuous grade estimation, the target ore grade of each piece of associated ore to be separated has been obtained. At this point, in order to separate the high-grade ore from the low-grade waste rock, the embodiments of this specification will generate a valve timing control signal corresponding to the associated ore to be separated based on the continuous grade estimation, and the target ore in the associated ore to be separated will be separated to the corresponding collection area by controlling the valve timing control signal.

[0045] Specifically, in one or more embodiments of this specification, generating a timing control signal for the spray valve corresponding to the associated ore to be separated, based on the continuous grade estimation, specifically includes: Compare continuous grade estimates with predetermined grade thresholds. In one approach, the predetermined grade threshold is determined based on the current market price of the target ore and beneficiation costs, balancing sales revenue and processing costs. In another approach, the optimal grade threshold is statistically determined by analyzing historical separation data or representative ore sample data, and this optimal threshold is used as the predetermined grade threshold.

[0046] If the estimated continuous grade is greater than or equal to the preset grade threshold, the associated ore to be separated is determined to belong to the target ore, and a corresponding valve opening signal is generated. If the estimated continuous grade is less than the preset grade threshold, the associated ore to be separated is determined not to belong to the target ore, and based on the corresponding valve opening state, it is determined whether to generate a valve closing signal. The valve opening and closing signals are then time-sequentially ordered to obtain the valve timing control signal.

[0047] Specifically, in one or more embodiments of this specification, the target ore in the associated ore to be separated is separated into the corresponding collection area by controlling the timing control signal of the spray valve, specifically including: The timing control signal for the spray valves is sent to the spray valve array, so that the spray valve array can activate the corresponding spray valves according to the timing control signal. Based on the high-pressure gas emitted by the spray valves in response to the timing control signal, a directional airflow is generated, thereby blowing the target ore away from its original trajectory in the original ore flow and separating it to the corresponding collection area.

[0048] like Figure 2 The diagram shows a schematic of a smart separation device for associated minerals. In one or more embodiments of this specification, the device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform any of the methods described above.

[0049] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0050] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0051] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0052] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A method for intelligent separation of associated minerals, characterized in that, The method includes: X-ray energy spectrum data of the associated minerals to be separated are collected, and the X-ray energy spectrum data are subjected to matrix material decomposition processing to obtain the equivalent thickness of at least two matrix materials; wherein, the X-ray energy spectrum data includes attenuation information of the high-energy spectral band and the low-energy spectral band. Based on the equivalent thickness of the at least two base materials, a two-dimensional feature space of the equivalent atomic number and surface density of the co-associated ore to be separated is constructed. The distribution cluster morphology of each of the associated minerals to be separated is obtained in the two-dimensional feature space to dynamically extract the dynamic reference endpoints of each of the associated minerals to be separated; wherein, the dynamic reference endpoints include: a first reference endpoint representing high-grade minerals and a second reference endpoint representing low-grade gangue. The projection position of each of the associated ores to be separated on the line connecting the dynamic reference endpoints is determined, so as to determine the continuous grade estimate of the target mineral in the associated ores to be separated based on the projection position; Based on the continuous grade estimate, a timing control signal for the spray valve corresponding to the associated ore to be separated is generated, and the target ore in the associated ore to be separated is separated into the corresponding collection area by controlling the timing control signal for the spray valve.

2. The intelligent separation method for associated minerals according to claim 1, characterized in that, Collect X-ray energy dispersive spectroscopy data of the associated minerals to be separated, specifically including: The associated ore to be separated is irradiated by an X-ray emission source, and the X-ray signal transmitted through the associated ore to be separated is collected by a detector array. Based on the X-ray signal, obtain X-ray energy spectrum data containing attenuation information for different energy ranges.

3. The intelligent separation method for associated minerals according to claim 1, characterized in that, The X-ray energy spectrum data is subjected to matrix material decomposition processing to obtain the equivalent thickness of at least two matrix materials, specifically including: Based on the X-ray energy spectrum data, the low-energy spectrum transmission intensity value and the high-energy spectrum transmission intensity value of the associated ore to be separated are determined. Based on the transmission intensity values ​​of the low-energy spectral band and the high-energy spectral band, a transmission signal pair of the associated minerals to be separated is constructed. The matching results are obtained by matching the transmission signal pairs with the existing hierarchical fingerprint database; wherein, the existing hierarchical fingerprint database is constructed by hierarchically storing the multidimensional feature fingerprints of the standard samples after calibration tests based on various standard samples with known thickness and known composition; If the confidence level corresponding to the matching result is higher than the preset first threshold, then the equivalent thickness of at least two base materials is obtained. If the confidence level corresponding to the matching result is lower than a preset first threshold, the neighboring multidimensional feature fingerprint corresponding to the transmission signal pair is obtained, and the equivalent thickness of at least two base materials is obtained based on the interpolation fitting of the neighboring multidimensional feature fingerprint and the multidimensional feature fingerprint corresponding to the matching result.

4. The intelligent separation method for associated minerals according to claim 1, characterized in that, Based on the equivalent thickness of the at least two base materials, a two-dimensional feature space of equivalent atomic number and areal density of the associated ore to be separated is constructed, specifically including: Based on the equivalent atomic numbers corresponding to the at least two base substances, each of the at least two base substances is divided into a first type of base substance and a second type of base substance; wherein, the first type of base substance represents a base substance with a low atomic number component content, and the second type of base substance represents a base substance with a high atomic number component content; A two-dimensional rectangular coordinate system is constructed with the equivalent thickness of the first type of base material as the abscissa and the equivalent thickness of the second type of base material as the ordinate. The equivalent thicknesses of at least two base materials of each of the associated minerals to be separated are mapped to the horizontal and vertical coordinates of the two-dimensional rectangular coordinate system to obtain the feature points of the associated minerals to be separated. The feature points of each of the associated minerals to be separated within the preset acquisition time window are aggregated to obtain a two-dimensional feature space distribution map, and the two-dimensional feature space of the equivalent atomic number and surface density of the associated minerals to be separated is obtained based on the two-dimensional feature space distribution map.

5. The intelligent separation method for associated minerals according to claim 4, characterized in that, Obtaining the distribution cluster morphology of each of the associated minerals to be separated in the two-dimensional feature space specifically includes: Based on the two-dimensional feature space distribution map, the distribution density of feature points of each of the associated and co-existing ores to be separated within the preset acquisition time window is obtained; Clustering parameters are determined based on the distribution density, and the feature points are clustered based on the clustering parameters to obtain multiple cluster groups; Extract the distribution cluster morphology corresponding to each of the clusters; wherein, the distribution cluster morphology includes: the centroid position of each cluster, the boundary contour of the cluster, and the density distribution characteristics of feature points within the cluster.

6. The intelligent separation method for associated minerals according to claim 5, characterized in that, Dynamically extracting the dynamic reference endpoints of each of the associated minerals to be separated, specifically including: Based on the pre-defined mineral composition characteristics of the target ore, or the separation records of historical associated ores to be separated, identify the target cluster containing the target ore and the background cluster. The relative positional relationship between the target cluster and the background cluster is obtained. Based on the continuous direction between the target cluster and the background cluster, the distribution density abrupt change points of the target cluster are searched to obtain critical feature points. Based on pre-set confidence constraints, the critical feature points are screened to obtain the dynamic reference endpoints of the associated ores to be separated.

7. The intelligent separation method for associated minerals according to claim 1, characterized in that, Determine the projection position of each of the associated ores to be separated on the line connecting the dynamic reference endpoints, and determine the continuous grade estimate of the target mineral in the associated ores to be separated based on the projection position, specifically including: Based on the pre-defined mineral composition characteristics of the target ore, or the separation records of historical associated ores to be separated, determine the high-grade endpoint; A grade mapping reference axis is constructed based on the connection between the dynamic reference endpoint and the high-grade endpoint; Obtain the characteristic points of each of the associated ores to be separated, and their vertical projection points on the grade mapping reference axis; Obtain a first distance between the vertical projection point and the high-quality endpoint, and a second distance between the vertical projection point and the dynamic reference point; The continuous grade estimate of the target mineral in each of the associated ores to be separated is determined by the proportional relationship between the first distance and the second distance; wherein the continuous grade estimate is negatively correlated with the first distance and positively correlated with the second distance.

8. The intelligent separation method for associated minerals according to claim 1, characterized in that, Based on the continuous grade estimate, a timing control signal for the spray valve corresponding to the associated ore to be separated is generated, specifically including: Compare the estimated continuous grade with the preset grade threshold; If the estimated continuous grade is greater than or equal to the preset grade threshold, then it is determined that the associated ore to be separated belongs to the target ore, and a corresponding valve opening signal is generated. If the estimated continuous grade is less than the preset grade threshold, it is determined that the separated associated ore does not belong to the target ore, and based on the corresponding valve opening state, it is determined whether to generate a valve closing signal. The timing sequence of the valve opening signal and the valve closing signal is sorted to obtain the valve timing control signal.

9. The intelligent separation method for associated minerals according to claim 1, characterized in that, The target ore in the associated ore to be separated is separated into the corresponding collection area by controlling the timing signal of the spray valve, specifically including: The timing control signal of the injection valve is sent to the injection valve array so that the injection valve array can start the corresponding injection valve according to the timing control signal of the injection valve; The directional airflow generated by the high-pressure gas from the spray valve blows the target ore away from its original trajectory from the original ore flow and separates it into the corresponding collection area.

10. An intelligent separation device for associated minerals, characterized in that, The device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in any one of claims 1-9.

Images