A method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore

Abstract

The present invention relates to the field of mineral processing data processing technology, and in particular to a method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore, comprising collecting multi-source process mineralogical data output by an X-ray fluorescence spectrometer, a mineral automatic analyzer, etc., converting the data into a structured format and constructing a process mineralogical database; based on the historical data of the database, a random forest algorithm is used to train an association model, learn the association between mineral characteristics such as calcite content, fluorite particle size distribution, and barite monomer dissociation degree and the inhibitor ratio, and optimize the model parameters through flotation test feedback. The combined inhibitor solves the problem of selective separation caused by similar surface properties through the synergistic effect of modified water glass covering the fluorite calcium active sites, citric acid regulating the barite barium active sites, and modified starch assisting in enhancing inhibition. Structured data processing and model closed-loop optimization solve the problem of inefficient multi-source data integration, thereby improving flotation separation efficiency and concentrate grade.

Description

Technical Field

[0001] The present invention relates to the technical field of mineral processing data processing, and in particular to a method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore. Background Art

[0002] High-calcium fluorite and barite paragenetic minerals often face challenges in flotation separation due to the similarities in their crystal structure and surface properties. The surface active sites of the two are prone to non-selective adsorption with collectors, making it difficult to effectively separate the target minerals. Due to limited targets, a single inhibitor can usually only exert an inhibitory effect on the surface characteristics of one of the minerals, and it is difficult to simultaneously regulate the binding ability of the two minerals with the collector. The design of a combined inhibitor must be based on the differences in the surface ions of the two minerals - the surface of high-calcium fluorite is rich in calcium ions, and the surface of barite is mainly barium ions. Inhibitors of different compositions can produce specific effects with these two ions respectively: one component covers the calcium active sites on the surface of fluorite by chemical adsorption, reducing its affinity with the collector; the other component attaches to the barium active sites on the surface of barite through electrostatic effects or hydrogen bonds, changing its surface hydrophobicity. When the two components work synergistically, they weaken the adsorption capacity of fluorite and barite on the collector respectively, but there are differences in the degree of inhibition, so that one of the minerals preferentially combines with the collector and floats up, while the other remains in a hydrophilic state due to the inhibition effect and remains in the slurry, ultimately achieving selective separation of the two minerals.

[0003] In the process mineralogy of high-calcium fluorite and barite co-occurring ores, analytical data (such as elemental content data from X-ray fluorescence spectrometers and particle size distribution and monomer dissociation degree data from mineral automatic analyzers (MLAs)) suffer from incompatible formats and inconsistent dimensions, leading to inefficient data integration and correlation analysis. This makes it difficult to quickly establish correlations between mineral properties and flotation process parameters. For example, X-ray fluorescence spectrometers produce tabular data for elemental content of calcite, fluorite, and barite through chemical analysis, while the MLA generates a histogram of mineral particle size distribution and a matrix of monomer dissociation degree through image recognition and statistics. These two data types are stored as text tables and graphical statistics, respectively. Manual extraction of key parameters and format conversion are required for cross-comparison and correlation between calcite content, mineral particle size, and flotation depressant dosage. This manual process is not only time-consuming but can also affect the accuracy of subsequent flotation process optimization due to data extraction errors, extending the cycle from data collection to process adjustments. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the present invention provides a method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore, which solves the separation problem of insufficient selectivity of a single inhibitor due to similar surface properties in high-calcium fluorite and barite paragenetic ore, and the problem of low efficiency of integration and correlation analysis caused by incompatible formats and inconsistent dimensions of multi-source analysis data in process mineralogy research.

[0005] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0006] The present invention provides a method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore, comprising:

[0007] Step 1: Collect process mineralogy data of high-calcium fluorite and barite paragenetic ore, the process mineralogy data including element content data output by X-ray fluorescence spectrometer, particle size distribution data and monomer dissociation degree correlation data output by mineral automatic analyzer, and corresponding data of inhibitor component, concentrate grade and recovery rate in historical flotation tests;

[0008] Step 2: Convert the process mineralogy data into a structured format. The tabular data in the process mineralogy data is stored as database columns through field mapping. The key parameters of the graphic data are extracted through image recognition and stored as numerical fields. The sample number and acquisition timestamp metadata are added to form a process mineralogy database.

[0009] Step 3: Using a machine learning algorithm to train an association model based on historical data on mineral properties and inhibitor composition in a process mineralogy database, the mineral property data includes calcite content, fluorite particle size distribution, and barite monomer dissociation degree;

[0010] Step 4: receiving real-time mineral property data of the ore to be processed, inputting the real-time mineral property data of the ore to be processed into the correlation model, and obtaining a recommended inhibitor composition ratio;

[0011] Step 5: Prepare a combined depressant by mixing raw materials according to the recommended ratio, record the flotation test results, and feed the test results back to the correlation model to update the model parameters. The flotation test results include concentrate grade, recovery rate, and mineral floating order data.

[0012] Furthermore, in the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 1 comprises:

[0013] Calcite, fluorite and barite content distribution data output by X-ray fluorescence spectrometer;

[0014] The mineral automatic analyzer outputs the data on the proportion of each mineral particle size range and the intergrowth ratio of fluorite, barite and calcite;

[0015] The graphic data output by the mineral automatic analyzer is used to extract the particle size interval value and dissociation percentage parameters through image recognition.

[0016] Furthermore, in the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 3 comprises:

[0017] The random forest algorithm was used to train the association model, and the input variables were the calcite content, fluorite particle size distribution, and barite monomer dissociation degree data from the process mineralogy database;

[0018] The historical data in the process mineralogy database were divided into a training set and a validation set in a ratio of 7:3. The training set was used to learn the correlation law between mineral characteristics and inhibitor ratios in the correlation model.

[0019] The validation set is used to evaluate the prediction error of the association model and optimize the model stability by adjusting the tree depth and feature sampling ratio hyperparameters.

[0020] Furthermore, in the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 4 comprises:

[0021] Real-time mineral property data of the ore to be processed include calcite content, fluorite particle size distribution range and barite monomer dissociation degree;

[0022] Input the real-time mineral property data into the association model trained in step 3. If the calcite content is higher than the average threshold of the historical data in the process mineralogy database and the fluorite particle size is smaller than the median particle size of the historical data, the association model outputs a ratio for increasing the proportion of modified water glass;

[0023] If the dissociation degree of the barite monomer is lower than the dissociation degree benchmark value of the historical data, the correlation model outputs a ratio of increasing the proportion of citric acid.

[0024] Furthermore, in the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 5 comprises:

[0025] The raw materials of the combined inhibitor are modified water glass and citric acid, and the raw material ratio is determined based on the ratio of modified water glass to citric acid recommended by the correlation model in step 4;

[0026] The raw materials were mixed by mechanical stirring at a speed of 100-150 rpm to uniformly disperse the modified water glass and citric acid;

[0027] The stirring time is 15-20 minutes, and the mixing environment is at room temperature, so that the raw materials are fully mixed to form a homogeneous inhibitor system.

[0028] Furthermore, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore of the present invention further comprises:

[0029] Apply the combined depressant prepared in step 5 to a flotation test, and record the floating order of fluorite and barite, the calcium fluoride content, the barium sulfate content, and the residual calcite in the concentrate during the test;

[0030] Compare and analyze the flotation result data recorded in the test with the ratio recommended by the correlation model in step 4 to generate a feedback adjustment signal for adjusting the characteristic weight coefficient of the mineral characteristics to the inhibitor ratio in the correlation model in step 3;

[0031] After the correlation model is updated, the inhibitor ratio recommendation logic for subsequent ore to be processed is optimized based on the updated feature weight coefficient to improve the matching degree between the recommended ratio and the actual flotation effect.

[0032] Furthermore, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore of the present invention further comprises:

[0033] Modified starch is used to help enhance the inhibitory effect of modified water glass on fluorite by covering the calcium active sites on the surface of fluorite;

[0034] Modified water glass covers the calcium active sites on the surface of fluorite through chemical adsorption, which is used to reduce the affinity between fluorite and the collector, so that the fluorite surface remains hydrophilic;

[0035] Citric acid is attached to the barium active sites on the barite surface through electrostatic interaction or hydrogen bonding, which is used to change the hydrophobicity of the barite surface and regulate its binding ability with the collector;

[0036] The three act synergistically in the homogeneous inhibitor system formed in step 5, differentially inhibiting the adsorption capacity of fluorite and barite on the collector, so that one of the minerals floats preferentially and the other remains in the slurry for selective separation.

[0037] Beneficial effects of the present invention:

[0038] The present invention converts multi-source process mineralogical data into a unified format and constructs a database through structured data processing technology, solving the problem of low integration efficiency caused by incompatible formats and inconsistent dimensions of traditional multi-source data, and providing a standardized data basis for subsequent correlation analysis; based on the correlation model trained by the random forest algorithm, combined with real-time mineral property data, dynamically recommends inhibitor ratios, realizes adaptive adjustment of the ratio scheme to different ore properties, and improves the targeted selection of inhibitors; modified water glass, citric acid and modified starch form surface hydrophobicity differences of fluorite with strong hydrophilicity and barite with weak hydrophobicity through synergistic adsorption of differentiated action sites (calcium active sites and barium active sites) in a homogeneous system, effectively solving the problem of insufficient selectivity of a single inhibitor due to similar surface properties, and ultimately improving the flotation separation efficiency and concentrate grade of high-calcium fluorite and barite. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on the drawings without paying any creative labor.

[0040] Figure 1 A flow chart of a method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. The technical solutions provided by each embodiment of the present invention are described in detail below in conjunction with the drawings. In order to better understand the purpose of the present invention, the present invention is further described in detail below.

[0042] See also Figure 1 The present invention provides a method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore, comprising:

[0043] Step 1: Collect process mineralogy data of high-calcium fluorite and barite paragenetic ore, the process mineralogy data including element content data output by X-ray fluorescence spectrometer, particle size distribution data and monomer dissociation degree correlation data output by mineral automatic analyzer, and corresponding data of inhibitor component, concentrate grade and recovery rate in historical flotation tests;

[0044] Step 2: Convert the process mineralogy data into a structured format. The tabular data in the process mineralogy data is stored as database columns through field mapping. The key parameters of the graphic data are extracted through image recognition and stored as numerical fields. The sample number and acquisition timestamp metadata are added to form a process mineralogy database.

[0045] Step 3: Using a machine learning algorithm to train an association model based on historical data on mineral properties and inhibitor composition in a process mineralogy database, the mineral property data includes calcite content, fluorite particle size distribution, and barite monomer dissociation degree;

[0046] Step 4: receiving real-time mineral property data of the ore to be processed, inputting the real-time mineral property data of the ore to be processed into the correlation model, and obtaining a recommended inhibitor composition ratio;

[0047] Step 5: Prepare a combined depressant by mixing raw materials according to the recommended ratio, record the flotation test results, and feed the test results back to the correlation model to update the model parameters. The flotation test results include concentrate grade, recovery rate, and mineral floating order data.

[0048] The present invention provides a method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore, which achieves dynamic optimization of the inhibitor ratio through a data-driven intelligent process. The core steps and underlying technical solutions are supplemented as follows:

[0049] Process mineralogy data collection is the starting point of the method, and multi-source heterogeneous data need to be systematically acquired to support subsequent analysis. Specifically, the X-ray fluorescence spectrometer (XRF) outputs the content distribution data of calcite, fluorite, and barite through chemical element analysis, usually presented in the form of a text table, recording the mass percentage of each mineral in the ore (such as calcite content 15% and fluorite content 40%). The mineral automatic analyzer (MLA) outputs two types of data through image recognition and statistical analysis: one is the proportion of each mineral particle size in the range of 0.074-0.15mm, 0.15-0.3mm, etc. (particle size distribution data); the other is the intergrowth ratio of fluorite, barite, and calcite (monomer dissociation degree intergrowth relationship data). The latter represents the degree of symbiosis between different minerals in matrix form (such as the intergrowth ratio of fluorite and barite is 25%). Historical flotation test data records the ratios of inhibitor components such as modified water glass and citric acid in previous tests (such as 3:1, 2.5:1.5) and the grades of calcium fluoride and barium sulfate in the corresponding concentrates (such as calcium fluoride grade 92% and barium sulfate grade 88%). All data collection is marked with the sample number (such as S20250515-01) and collection time (such as May 15, 2025 10:00) to ensure data traceability.

[0050] Collected multi-source data must be converted into a unified structured format to construct a process mineralogy database. For tabular data output by XRF, column names such as "calcite content" and "fluorite content" are mapped to the database's "calcite_content" and "fluorite_content" fields through field mapping and stored as numeric data. Graphical data output by MLA (such as particle size distribution histograms and intergrowth relationship heat maps) is processed using image recognition technology. Optical character recognition (OCR) is used to extract the values ​​of each particle size range in the histogram (e.g., the 0.1-0.5 mm range accounts for 60%). Image segmentation algorithms are used to identify percentage values ​​in the intergrowth relationship matrix (e.g., the fluorite-barite intergrowth ratio is 25%) and convert these values ​​into numeric fields for storage. All data are appended with sample number and timestamp metadata and stored in a relational database (e.g., MySQL) using the "sample number - acquisition time - data type - data value" structure, forming a queryable and linkable process mineralogy database.

[0051] Based on historical data from a process mineralogy database, a random forest algorithm was used to train an association model to learn the correlation between mineral properties and inhibitor ratios. The model input variables were the calcite content (e.g., 18%), fluorite particle size distribution (e.g., 60% of the 0.1-0.5mm range), and barite monomer dissociation degree (e.g., 92%) from the database. The output variable was the ratio of modified sodium silicate to citric acid (e.g., 3:1). During training, the 80 groups of historical data in the database were divided into training sets (56 groups) and validation sets (24 groups) in a ratio of 7:3: the training set was used to model the positive correlation between the increase in calcite content and the increase in the proportion of modified water glass, and the negative correlation between the decrease in barite dissociation degree and the increase in the proportion of citric acid; the validation set was used to evaluate the model prediction error (such as the deviation between the predicted ratio of 3:1 and the actual optimal ratio of 2.8:1.2), and to optimize the model stability by adjusting the tree depth of the random forest (such as from 5 layers to 7 layers) and the feature sampling ratio (such as from 0.7 to 0.8), and finally obtain an association model that can accurately map mineral properties to inhibitor ratios.

[0052] Real-time mineral property data for the ore being processed is fed into the correlation model in a structured format from the process mineralogy database to generate dynamically recommended inhibitor ratios. This real-time data includes the calcite content of the current batch of ore (e.g., 20%), fluorite particle size distribution (e.g., 55% in the 0.1-0.5mm range), and barite monomer dissociation (e.g., 88%). This data is formatted in the same way as historical data in the database (e.g., "calcite_content=20%," "fluorite_particle_size=55%," "barite_liberation=88%"). The model outputs a ratio based on the association rules obtained through training: if the calcite content (20%) is higher than the average threshold of the historical data in the database (17%) and the fluorite particle size (0.1-0.5mm) is smaller than the historical median particle size (0.3mm), the model outputs a ratio to increase the proportion of modified water glass (such as 3.2:0.8); if the barite dissociation degree (88%) is lower than the historical benchmark value (90%), the model outputs a ratio to increase the proportion of citric acid (such as 2.5:1.5).

[0053] Once the recommended ratio is determined, a combined depressant is prepared by mechanically mixing the raw materials. This is then fed back into the optimization model through flotation testing. Specifically, modified sodium silicate and citric acid are weighed according to the recommended ratio (e.g., 3.2:0.8) and mechanically stirred at 120 rpm for 18 minutes at room temperature to evenly disperse the two raw materials and form a homogeneous system. The prepared depressant is then applied to flotation tests, and the flotation order of fluorite and barite (e.g., fluorite floats within 10 minutes, barite after 20 minutes), as well as the concentrate's calcium fluoride content (e.g., 93%), barium sulfate content (e.g., 89%), and residual calcite content (e.g., 2%) are recorded. After comparing the test data with the model's recommended ratio (3.2:0.8), feedback adjustment signals are generated (e.g., adjusting the weighting coefficient of calcite content to the modified sodium silicate ratio from 0.4 to 0.5) to update the feature weights of the association model. After the model is updated, subsequent inhibitor ratio recommendations for processed ores will be based on the adjusted weight coefficients, improving the match between the recommended ratio and the actual flotation performance.

[0054] Specifically, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 1 comprises:

[0055] Calcite, fluorite and barite content distribution data output by X-ray fluorescence spectrometer;

[0056] The mineral automatic analyzer outputs the data on the proportion of each mineral particle size range and the intergrowth ratio of fluorite, barite and calcite;

[0057] The graphic data output by the mineral automatic analyzer is used to extract the particle size interval value and dissociation percentage parameters through image recognition.

[0058] In the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic minerals described in the present invention, the process mineralogical data collection in step 1 requires systematic acquisition of multi-source heterogeneous data to provide a basis for subsequent data integration and model training.

[0059] The calcite, fluorite, and barite content distribution data output by an X-ray fluorescence spectrometer (XRF) is obtained by spectrally analyzing the mass percentage of each mineral using X-rays to excite elements in the ore sample, producing characteristic fluorescence. Specifically, the ore sample is ground to 200 mesh (approximately 0.074 mm particle size) and pressed into tablets. The tablets are then placed into the XRF instrument for elemental analysis. The instrument uses a calibration curve to convert the elemental signals into mineral content data. For example, the test results for a sample may show a calcite content of 18%, a fluorite content of 45%, and a barite content of 22%. This data is typically output as a two-dimensional table with columns such as "Mineral Type" and "Content (Mass Percentage)" to directly reflect the relative proportions of the target mineral and impurity minerals in the ore.

[0060] The Mineral Analyzer (MLA) outputs data on the percentage of each mineral size range. Using a scanning electron microscope (SEM) and an energy dispersive spectrometer (EDS), the system then uses statistical algorithms to analyze the distribution of minerals within different size ranges. For example, the system divides ore particles into three ranges: 0.074-0.15mm, 0.15-0.3mm, and 0.3-0.6mm. The system then calculates the percentage of fluorite, barite, and calcite particles within each range (e.g., 60% of fluorite falls within the 0.15-0.3mm range). This data is typically presented as a bar chart or line graph, visually reflecting the characteristics of the mineral size distribution.

[0061] The intergrowth ratio data for fluorite, barite, and calcite output by the mineral analyzer is obtained by identifying the intergrowth relationships between different minerals within ore particles. The intergrowth ratio is defined as the percentage of a particular mineral particle intergrowthed with other mineral particles relative to the total number of particles of that mineral (e.g., 25% of fluorite particles are intergrowthed with barite, and 15% with calcite). MLA uses image segmentation technology to distinguish the boundaries of different minerals and combines compositional analysis to determine intergrowth relationships. The data is ultimately output as a matrix or heat map, such as "Fluorite-barite intergrowth ratio 25%" or "Fluorite-calcite intergrowth ratio 15%." This data is used to assess the degree of dissociation between individual minerals and is a key indicator of the difficulty of flotation separation.

[0062] Graphical data output by automated mineral analyzers (such as particle size distribution histograms and intergrowth ratio thermograms) requires image recognition technology to extract key parameters. Specifically, optical character recognition (OCR) is used to extract particle size range values ​​(e.g., "0.15-0.3mm") and corresponding percentages (e.g., "60%") from the histogram axes and annotations. Image segmentation algorithms are then used to identify the intergrowth ratio percentage (e.g., "25%") from the color gradient of the thermogram. These extracted parameters are formatted and stored as numeric fields (e.g., "particle_size_range: 0.15-0.3mm," "proportion: 60%"), which, along with the tabular data output by XRF, form the raw data layer of the process mineralogy database.

[0063] Specifically, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 3 comprises:

[0064] The random forest algorithm was used to train the association model, and the input variables were the calcite content, fluorite particle size distribution, and barite monomer dissociation degree data from the process mineralogy database;

[0065] The historical data in the process mineralogy database were divided into a training set and a validation set in a ratio of 7:3. The training set was used to learn the correlation law between mineral characteristics and inhibitor ratios in the correlation model.

[0066] The validation set is used to evaluate the prediction error of the association model and optimize the model stability by adjusting the tree depth and feature sampling ratio hyperparameters.

[0067] In the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic minerals described in the present invention, the association model training in step 3 realizes intelligent association learning of mineral properties and inhibitor ratios through the random forest algorithm. Its specific implementation requires the combination of data preprocessing, model construction, and parameter optimization.

[0068] The input variables for the correlation model are structured data from the process mineralogy database, including calcite content (e.g., mass percentage values ​​of 15%, 18%, and 20%), fluorite particle size distribution (e.g., statistical values ​​such as 50% and 60% for the 0.1-0.5 mm range), and barite monomer dissociation (e.g., the inverse of the intergrowth ratio of 85%, 90%, and 92%). These variables are stored in the database as numeric fields (e.g., "calcite_content," "fluorite_size_distribution," and "barite_liberation"). These variables are read through programming interfaces (e.g., Python's pandas library) and converted into a feature matrix recognizable by the model, with each row representing a historical test sample and each column corresponding to a mineral property variable.

[0069] Before model training, the historical data in the process mineralogy database must be divided into a training set and a validation set in a 7:3 ratio. For example, if the database contains 100 sets of historical experimental data (each set includes calcite content, fluorite particle size distribution, barite dissociation degree, and corresponding inhibitor ratio), 70 sets are randomly selected as the training set (for model learning) and the remaining 30 sets as the validation set (for model evaluation). This data partitioning ensures consistent sample distribution (for example, the ratio of high, medium, and low calcite content samples in the training and validation sets is the same) to avoid model overfitting or underfitting due to data bias.

[0070] The training set is used to build a correlation model that learns the relationship between mineral properties and inhibitor ratios. Specifically, a random forest algorithm is used to construct the model: The algorithm extracts multiple subsamples from the training set through bootstrap sampling. Each subsample generates a decision tree. Each node in the tree selects the optimal splitting condition (e.g., "calcite content > 17%)" based on the information gain of the mineral property variable (e.g., calcite content). Ultimately, the voting results from multiple trees output the inhibitor ratio (e.g., modified sodium silicate: citric acid = 3:1). For example, samples with high calcite content (>17%) in the training set often correspond to a high modified sodium silicate ratio (e.g., 3:1). By learning from these patterns, the model establishes a positive correlation between calcite content and modified sodium silicate ratio.

[0071] The validation set is used to evaluate the prediction error of the association model and optimize model stability. After model training, the model is fed with the mineral property data from the validation set (e.g., calcite content 19%, fluorite particle size distribution 55%, and barite dissociation degree 88%). The model then generates a predicted inhibitor ratio (e.g., a predicted ratio of 3.2:0.8). This ratio is then compared with the actual optimal ratio in the validation set (e.g., a ratio of 3:1) to calculate the error (e.g., an absolute deviation of 0.2:0.2). If the error exceeds a preset threshold (e.g., 0.3:0.3), the random forest hyperparameters are adjusted. For example, the tree depth can be increased (from the default 5 to 7 layers) to enhance the model's ability to capture complex associations, or the feature sampling ratio can be reduced (from 0.8 to 0.7) to reduce the impact of noise. Through repeated adjustments and validation, an association model with stable prediction errors within an acceptable range (e.g., a deviation ≤ 0.2:0.2) is ultimately achieved.

[0072] Specifically, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 4 comprises:

[0073] Real-time mineral property data of the ore to be processed include calcite content, fluorite particle size distribution range and barite monomer dissociation degree;

[0074] Input the real-time mineral property data into the association model trained in step 3. If the calcite content is higher than the average threshold of the historical data in the process mineralogy database and the fluorite particle size is smaller than the median particle size of the historical data, the association model outputs a ratio for increasing the proportion of modified water glass;

[0075] If the dissociation degree of the barite monomer is lower than the dissociation degree benchmark value of the historical data, the correlation model outputs a ratio of increasing the proportion of citric acid.

[0076] In the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore described in the present invention, step 4 realizes intelligent recommendation of the inhibitor ratio of the ore to be processed through an association model. The core of this method lies in the structured input of real-time mineral property data and the dynamic decision-making logic of the model.

[0077] The real-time mineral property data for the ore being processed must be formatted consistently with the historical data in the process mineralogy database. This data includes three dimensions: calcite content (expressed as a percentage by mass, e.g., 19%), fluorite particle size distribution (expressed as the percentage of each particle size range, e.g., 55% for the 0.1-0.5 mm range), and barite monomer dissociation (expressed as the inverse of the intergrowth ratio, e.g., 88%). This data is collected using the same testing equipment as in Step 1: calcite content is measured by X-ray fluorescence (XRF), while fluorite particle size distribution and barite monomer dissociation are analyzed and output by a mineral automatic analyzer (MLA). This data is then converted to a structured format (e.g., "calcite_content=19%," "fluorite_size=55%," "barite_liberation=88%") through image recognition and field mapping, ensuring that the field names and data types are identical to those of the historical data in the database.

[0078] After real-time mineral property data is fed into the association model trained in step 3, the model makes dynamic decisions based on the statistical characteristics of historical data. The average threshold and median particle size for the historical data are calculated by statistically calculating the calcite content and fluorite particle size distribution of all historical samples in the process mineralogy database. For example, the average calcite content of 100 historical samples in the database is 17% (the average threshold), and the median particle size of the fluorite particle size distribution is 0.3 mm (meaning that 50% of the samples have a fluorite particle size less than 0.3 mm). If the calcite content of the ore being processed (19%) is higher than the average threshold (17%) and the fluorite particle size (55% of the 0.1-0.5 mm range corresponds to a median particle size of 0.25 mm) is smaller than the historical median particle size (0.3 mm), the model identifies that the fluorite surface in this ore has more exposed calcium active sites and requires stronger calcium site coverage. Therefore, the model outputs a ratio that increases the modified sodium silicate ratio (for example, from the historical average of 3:1 to 3.2:0.8).

[0079] If the barite monomer dissociation degree (88%) of the ore being processed is lower than the historical dissociation degree baseline (90%), the model determines that the barite is highly intertwined with other minerals and that the surface barium active sites are encapsulated, necessitating increased action on the barium sites to change their hydrophobicity. Therefore, the model outputs a ratio that increases the citric acid ratio (for example, from the historical average of 2.8:1.2 to 2.5:1.5). This ratio adjustment logic is based on the association rules learned during the model training phase (for example, for every 1% increase in calcite content, the modified water glass ratio increases by 0.1; for every 1% decrease in barite dissociation degree, the citric acid ratio increases by 0.05). The final recommended ratio is calculated by linearly combining the weight coefficients of each mineral property variable (for example, a weight of 0.4 for calcite content and 0.3 for barite dissociation degree).

[0080] Specifically, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to the present invention, step 5 comprises:

[0081] The raw materials of the combined inhibitor are modified water glass and citric acid, and the raw material ratio is determined based on the ratio of modified water glass to citric acid recommended by the correlation model in step 4;

[0082] The raw materials were mixed by mechanical stirring at a speed of 100-150 rpm to uniformly disperse the modified water glass and citric acid;

[0083] The stirring time is 15-20 minutes, and the mixing environment is at room temperature, so that the raw materials are fully mixed to form a homogeneous inhibitor system.

[0084] In the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore described in the present invention, steps 4 and 5 together constitute a closed-loop process of "model recommendation-raw material preparation", and the precise preparation of the combined inhibitor is achieved through real-time data-driven ratio recommendation and standardized mixing process.

[0085] The real-time mineral property data input and model decisions in step 4 are based on the historical statistical characteristics of the process mineralogy database. Real-time data collection and processing are consistent with step 1: calcite content is determined by X-ray fluorescence (XRF) analysis (e.g., the calcite content of the current batch of ore is 19%). Fluorite particle size distribution is calculated using a mineral automatic analyzer (MLA) to calculate the percentage of each size range (e.g., 55% of the 0.1-0.5 mm range corresponds to a median particle size of 0.25 mm). The degree of dissociation of barite monomers is calculated by MLA analysis of intergrowth relationships (e.g., if the intergrowth ratio is 12%, the degree of dissociation is 88%). These data are converted to a structured format through field mapping (e.g., "calcite_content = 19%," "fluorite_size = 55%," "barite_liberation = 88%"), which is identical to the field names (e.g., "calcite_content") and data types (numeric) of the historical data in the database.

[0086] When making model decisions, the historical statistical features of the process mineralogy database are determined through a preprocessing step: the average threshold for calcite content is the average of all historical samples in the database (e.g., 17%), the median fluorite particle size is the particle size corresponding to the 50th percentile in the historical samples (e.g., 0.3mm), and the baseline for barite dissociation is the lowest dissociation degree that achieves optimal flotation performance in the historical samples (e.g., 90%). If the real-time calcite content (19%) is higher than the average threshold (17%) and the fluorite particle size (0.25mm) is lower than the median particle size (0.3mm), the model identifies that the calcium active sites on the fluorite surface are more fully exposed and require enhanced calcium site coverage. Therefore, the model outputs a modified sodium silicate ratio (e.g., from the historical average of 3:1 to 3.2:0.8). If the real-time barite dissociation degree (88%) is lower than the baseline value (90%), the model determines that the barium active sites on the barite surface are encapsulated and the effect of the barium sites needs to be strengthened, so it outputs an increase in the citric acid ratio (for example, from the historical average of 2.8:1.2 to 2.5:1.5).

[0087] The combined inhibitor preparation in Step 5 is based on the recommended ratio in Step 4. For raw material weighing, modified sodium silicate (320 g) and citric acid (80 g) are weighed separately using an electronic balance (0.1 g accuracy) according to the recommended ratio (e.g., 3.2:0.8), ensuring a mass error of no more than ±1%. A laboratory mechanical stirrer (e.g., IKA RW 20) with an anchor-type impeller (5 cm diameter) is used for mixing to improve mixing of high-viscosity raw materials. The stirring speed is set to 120 rpm (within the range of 100-150 rpm). This speed avoids uneven mixing caused by too low a speed and prevents the introduction of bubbles caused by too high a speed. The stirring time is controlled to 18 minutes (within the range of 15-20 minutes) using a timer (e.g., a digital timer) to ensure thorough contact between the raw materials. The mixing environment is kept at room temperature (25 ± 5°C), eliminating the need for additional heating or cooling, reducing energy consumption and preventing temperature fluctuations from affecting raw material properties.

[0088] During the mixing process, observe the raw materials to determine dispersion uniformity: Initially (0-5 minutes), the modified sodium silicate (viscous liquid) and citric acid (white powder) appear separated into layers. Between 5 and 15 minutes, the paddles create a vortex, gradually dissolving and dispersing the citric acid powder. After 15 minutes, the color and viscosity of the system converge (e.g., a translucent, uniform liquid), indicating a homogeneous inhibitor system. The final inhibitor is transferred via pipette to a sealed container and labeled with the sample number (e.g., S20250515-02) and preparation date (e.g., May 15, 2025, 14:30) for subsequent flotation testing.

[0089] Specifically, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore of the present invention further includes:

[0090] Apply the combined depressant prepared in step 5 to a flotation test, and record the floating order of fluorite and barite, the calcium fluoride content, the barium sulfate content, and the residual calcite in the concentrate during the test;

[0091] Compare and analyze the flotation result data recorded in the test with the ratio recommended by the correlation model in step 4 to generate a feedback adjustment signal for adjusting the characteristic weight coefficient of the mineral characteristics to the inhibitor ratio in the correlation model in step 3;

[0092] After the correlation model is updated, the inhibitor ratio recommendation logic for subsequent ore to be processed is optimized based on the updated feature weight coefficient, thereby improving the matching degree between the recommended ratio and the actual flotation effect.

[0093] In the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore described in the present invention, flotation test verification and model feedback optimization are the key to achieving dynamic iteration of inhibitor ratios. By comparing and analyzing the actual flotation effect with the model recommended results, the model's adaptability to complex ore characteristics is continuously improved.

[0094] Flotation tests using combined depressants should be conducted under standardized conditions. Specifically, a laboratory XFD single-cell flotation cell was used. The slurry volume was set at 1 L, the slurry concentration was controlled at 30% (the ratio of solids mass to total slurry mass), the aeration rate was 0.15 m³ / h, and the agitation speed was 1800 rpm. The combined depressant prepared in Step 5 (e.g., a 3.2:0.8 ratio of modified sodium silicate to citric acid) was added to the slurry at a dosage of 100 g / t (100 g per ton of ore) and stirred for 3 minutes to allow the depressant to fully interact with the mineral. Aeration was then initiated, and the order in which fluorite and barite floated was recorded: fluorite, due to its highly hydrophobic surface, typically floated within 1-3 minutes, while barite, due to its highly hydrophilic surface due to the inhibitory effect, typically floated after 5-8 minutes. If barite floats too early (e.g., within 2 minutes), it indicates an insufficient citric acid ratio and requires increased interaction with the barium sites.

[0095] Flotation test data includes concentrate grade and residual impurities. The concentrate is collected by scraping, dehydrated, and dried. X-ray fluorescence (XRF) is used to analyze the calcium fluoride (fluorite grade) and barium sulfate (barite grade) contents. For example, one test revealed 93% calcium fluoride and 89% barium sulfate. Residual calcite is determined by chemical titration (e.g., EDTA complexometric titration for calcium content) and converted to a mass percentage of calcite (e.g., 2%). All data is annotated with the test number (e.g., T20250515-01) and time (e.g., May 15, 2025, 15:00) and stored in the "Test Verification" table of the process mineralogy database, along with the mix ratio recommended by the correlation model in step 4 (e.g., 3.2:0.8).

[0096] Comparison and analysis of experimental data with the model's recommended mix ratios are achieved through deviation calculation. For example, the model recommends a mix ratio of modified sodium silicate to citric acid of 3.2:0.8. The actual flotation results showed a calcium fluoride grade of 93% (target grade: 92%) and a calcite residual of 2% (target residual: 3%), indicating that this mix effectively suppressed calcite. However, the barite grade was slightly lower at 89% (target grade: 90%), likely due to an insufficient citric acid ratio. By calculating the deviation of each indicator (e.g., barite grade deviation = -1%) and combining it with mineral characteristics (e.g., barite dissociation degree of 88%), a feedback adjustment signal is generated. If the barite dissociation degree and grade deviation are negatively correlated (lower dissociation degree indicates greater grade deviation), the characteristic weight coefficient of the barite dissociation degree in the model is increased (e.g., from 0.3 to 0.4), thereby strengthening the influence of this variable on the citric acid ratio.

[0097] The feature weights of the correlation model are adjusted using a random forest algorithm to assess feature importance. Each mineral property variable in the model (calcite content, fluorite particle size distribution, barite dissociation degree) is assigned a weight coefficient, reflecting its influence on the inhibitor ratio (for example, a calcite weight of 0.4 indicates a 40% impact on the modified sodium silicate ratio). Feedback adjustment signals adjust these weight coefficients to update the model. For example, if the correlation between barite dissociation degree and grade deviation increases, its weight coefficient will be increased from 0.3 to 0.4. The model will then prioritize the adjustment of citric acid ratio based on barite dissociation degree in subsequent recommendations. The updated model's prediction error is re-evaluated using a validation set (for example, reducing it from 0.2:0.2 to 0.15:0.15), ensuring that the optimized recommendation logic is more aligned with actual flotation results.

[0098] Specifically, the method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore of the present invention further includes:

[0099] Modified starch helps enhance the inhibitory effect of modified water glass on fluorite by covering the calcium active sites on the surface of fluorite;

[0100] Modified water glass covers the calcium active sites on the surface of fluorite through chemical adsorption, reducing the affinity between fluorite and collector, and keeping the fluorite surface hydrophilic;

[0101] Citric acid attaches to the barium active sites on the barite surface through electrostatic interaction or hydrogen bonding, changing the hydrophobicity of the barite surface and regulating its binding ability with the collector;

[0102] The three act synergistically in the homogeneous inhibitor system formed in step 5, differentially inhibiting the adsorption capacity of fluorite and barite on the collector, so that one of the minerals floats preferentially and the other remains in the slurry for selective separation.

[0103] In the high-calcium fluorite and barite paragenetic mineral combination inhibitor described in the present invention, modified starch, modified water glass and citric acid achieve precise regulation of the adsorption capacity of the two minerals through differentiated action sites and synergistic mechanisms. The specific action mode and synergistic process need to be explained in combination with the mineral surface characteristics and the physicochemical properties of the inhibitor components.

[0104] Modified starch, as a supplementary inhibitor, enhances the shielding of calcium active sites on the fluorite surface through physical coating. Crystalline lattice fractures expose numerous calcium ions on the fluorite surface, creating positively charged active sites that readily bind to collectors (such as oleate ions) through electrostatic interactions. Modified starch (such as carboxymethyl starch), containing multiple hydroxyl (-OH) and carboxyl (-COOH) groups in its molecular chain, can adsorb to the Ca²⁺ sites on the fluorite surface through hydrogen bonding, forming a polymer film approximately 5-10 nm thick. While this film alone has limited inhibitory capacity (reducing collector adsorption by only approximately 10%), it can form a superimposed coating with modified water glass. Modified water glass (such as sodium silicate-modified products) contains silicate ions, which chemically adsorb with Ca²⁺ to form a stable calcium silicate complex with a coating thickness of approximately 2-5 nm. The modified starch polymer film coats the surface of the calcium silicate layer, further hindering contact between the collector and Ca²⁺, reducing collector adsorption on the fluorite surface by approximately 30%, significantly enhancing the inhibitory effect.

[0105] The core function of modified water glass is to reduce the affinity of fluorite for collectors through chemical adsorption. The binding energy of its silicate ions to the Ca²⁺ on the fluorite surface is approximately -80 kJ / mol (calculated by density functional theory), far higher than the binding energy of the collector's oleate ions (-COO⁻) to Ca²⁺ (approximately -50 kJ / mol). Therefore, it preferentially occupies the Ca²⁺ sites. After adsorption, the fluorite surface changes from its original hydrophobicity (contact angle of approximately 75°) to a hydrophilicity (contact angle of approximately 30°), increasing its surface energy by approximately 20 mN / m. This makes it difficult for bubbles to attach to it and float up, remaining in the slurry.

[0106] Citric acid modulates the surface hydrophobicity of barite through electrostatic interactions and hydrogen bonding. The barite surface is primarily composed of positively charged barium ions. Citric acid molecules contain three carboxyl groups (-COOH), which become negatively charged after dissociation (at pH 7, the degree of dissociation is approximately 80%). These citric acid molecules can be adsorbed to the Ba²⁺ sites through electrostatic attraction. After adsorption, the hydrophobic carbon chains of the citric acid molecules face outward, increasing the contact angle of the barite surface from hydrophilic (approximately 40°) to slightly hydrophobic (approximately 60°), and reducing the surface energy by approximately 15 mN / m. At this point, the barite's binding affinity with collectors (such as alkyl sulfonates) lies between that of fluorite (strongly hydrophilic) and uninhibited minerals (strongly hydrophobic). During flotation, the barite floats later than the uninhibited mineral but earlier than the strongly suppressed fluorite, achieving selective separation.

[0107] The synergistic effect of the three components relies on the homogeneous inhibitor system formed in step 5. Modified water glass (liquid) and citric acid (powder) are uniformly dispersed under mechanical stirring. Modified starch (powder) is pre-dissolved (e.g., stirring in 40°C water for 30 minutes to form a 1% solution) before addition. The three components are then mixed at the molecular level in a homogeneous system (dispersed particle size <1μm). When the inhibitor is added to the slurry, the modified water glass preferentially adsorbs on the fluorite surface (90% adsorption within 2 minutes), followed by coverage by the modified starch (complete within 5 minutes), and the citric acid simultaneously adsorbs on the barite surface (complete within 3 minutes). This sequential adsorption and complementary spatial coverage make the fluorite surface significantly more hydrophilic than the barite (with a contact angle difference of approximately 30°). Ultimately, during flotation, the barite, due to its weak hydrophobicity, preferentially attaches to air bubbles and floats up (5-8 minutes), while the fluorite, due to its strong hydrophilicity, remains in the slurry (no significant floating after 10 minutes), achieving efficient separation of the two minerals.

[0108] The present invention solves the problem of selective separation of high-calcium fluorite and barite due to the similar surface properties through the synergistic effect of the components of the combined inhibitor. Modified water glass preferentially covers the calcium active sites on the surface of fluorite through chemical adsorption, reduces its affinity with the collector, and keeps the fluorite surface in a hydrophilic state; citric acid attaches to the barium active sites on the surface of barite through electrostatic action or hydrogen bonding, regulating its hydrophobicity to a weakly hydrophobic state; modified starch is used as an auxiliary component to enhance the shielding effect of the calcium sites on the surface of fluorite through physical covering. The three components are adsorbed sequentially in the homogeneous inhibitor system, forming a surface difference of strong hydrophilicity of fluorite and weak hydrophobicity of barite, so that the barite floats preferentially while the fluorite remains in the slurry, thereby achieving selective separation.

[0109] To address the incompatible formats and inefficient integration of multi-source data in process mineralogy, this invention achieves unified management through structured data processing and database construction. Tabular data output by an X-ray fluorescence spectrometer is converted into numerical database columns through field mapping. Graphical data from an automatic mineral analyzer is extracted through image recognition to extract key parameters (such as particle size range and percentage of dissociation) and stored as numerical fields. All data is then appended with sample number and timestamp metadata to form a structured process mineralogy database. This database eliminates format differences between text tables and graphical statistics, providing a unified data input foundation for subsequent correlation analysis.

[0110] Data-driven correlation model training and dynamic optimization further enhance the adaptability of separation results. Based on historical data from the process mineralogy database, a random forest algorithm is used to train the correlation model, learning the correlation between mineral properties such as calcite content, fluorite particle size distribution, and barite monomer dissociation degree and inhibitor ratios. Real-time mineral property data of the ore to be processed is input into the model, which then outputs a dynamically recommended ratio. Flotation test results are fed back into the model, and the recommendation logic is optimized by adjusting the feature weight coefficients. This allows the inhibitor ratio to dynamically adapt to the differences in properties between different batches of ore, ultimately improving the separation efficiency of high-calcium fluorite and barite.

[0111] Example 1: For a sample of a high-calcium fluorite and barite paragenetic ore, an X-ray fluorescence spectrometer (PANalytical Axios) was used to obtain tabular data showing a calcite content of 18%, a fluorite content of 45%, and a barite content of 22%. Scanning and analysis using a mineral automatic analyzer (FEI MLA650) revealed the particle size distribution data for each mineral (30% for the 0.074-0.15 mm range, 50% for the 0.15-0.3 mm range, and 20% for the 0.3-0.6 mm range) and intergrowth relationship data (25% for fluorite-barite intergrowth and 15% for fluorite-calcite intergrowth). Using the particle size distribution histogram and intergrowth relationship heat map output by the mineral automatic analyzer, optical characterization (OCR) technology was used to extract the values ​​and corresponding proportions of each particle size range in the histogram. An image segmentation algorithm was then used to identify the intergrowth percentage values ​​(25% and 15%) in the heat map. All data are appended with the sample number S20250515-01 and the collection time of May 15, 2025, at 10:00, and stored in a MySQL database through field mapping to form a structured process mineralogy database containing fields such as “calcite content,” “fluorite particle size distribution,” and “barite monomer dissociation degree.”

[0112] Example 2: A random forest algorithm was used to train an association model based on 100 sets of historical data from the Process Mineralogy Database. The data was divided into a training set (70 sets) and a validation set (30 sets) in a 7:3 ratio. The training set included calcite content (15%-22%), fluorite particle size distribution (0.074-0.6 mm), barite monomer dissociation degree (85%-95%), and corresponding inhibitor ratios (modified water glass: citric acid = 2.5:1.5 to 3.5:0.5). The validation set was used to evaluate the model's prediction error. During the training phase, the model learned the association that for every 1% increase in calcite content, the modified water glass ratio increased by 0.1. During the validation phase, by adjusting the random forest tree depth (from 5 to 7 layers) and the feature sampling ratio (from 0.7 to 0.8), the model's prediction error decreased from 0.3:0.3 to 0.2:0.2, improving stability.

[0113] Example 3: Real-time mineral property data for the ore to be processed was collected: X-ray fluorescence spectrometry revealed a calcite content of 20% (higher than the historical average threshold of 17% in the database), an automatic mineral analyzer determined a fluorite particle size distribution of 55% in the 0.1-0.5 mm range (with a median particle size of 0.25 mm, less than the historical median of 0.3 mm), and a barite monomer dissociation degree of 88% (less than the historical baseline of 90%). This data was fed into a trained association model. Based on the correlation between calcite content and fluorite particle size, the model outputted a ratio for increasing the modified sodium silicate ratio (3.2:0.8). Based on the correlation between the barite dissociation degree and the citric acid ratio, the model outputted a ratio for increasing the citric acid ratio (2.5:1.5).

[0114] Example 4: 320g of modified water glass and 80g of citric acid were weighed in the recommended ratio of 3.2:0.8 and stirred at 120 rpm for 18 minutes at room temperature (25°C) using an IKA RW 20 mechanical stirrer (anchor blade, 5cm diameter). Within the first 5 minutes of stirring, the viscous modified water glass and citric acid powder separated into separate layers. After 5-15 minutes, the paddles formed a vortex in the liquid, gradually dissolving and dispersing the citric acid powder. After 15 minutes, the color and viscosity of the liquid became consistent, forming a homogeneous translucent liquid. The dispersed particle size was determined to be less than 1μm, completing the preparation of the combined inhibitor.

[0115] Example 5: The prepared combined depressant (3.2:0.8) was added at a dosage of 100g / t to an XFD single-cell flotation machine (slurry volume 1L, 30% concentration, aeration rate 0.15m³ / h, agitator speed 1800 rpm). Test records show that fluorite showed no significant flotation after 10 minutes, while barite began to float after 5-8 minutes. The concentrate contained 93% calcium fluoride (target 92%), 89% barium sulfate (target 90%), and 2% residual calcite (target 3%). Comparing the recommended mix ratio with the test results, the barite grade deviation was -1%. Combined with the barite dissociation degree of 88%, a feedback adjustment signal was generated, adjusting the barite dissociation degree feature weight in the model from 0.3 to 0.4. After the update, the model's prediction error for subsequent samples decreased to 0.15:0.15, improving the match between the recommended mix ratio and actual flotation results.

[0116] Example 6: Analysis of the mechanism of action of the combined inhibitor: Modified water glass (silicate ions) chemically adsorbs Ca⁺ sites on the fluorite surface (binding energy -80 kJ / mol), reducing the fluorite contact angle from 75° to 30° and enhancing its surface hydrophilicity. Citric acid (negatively charged after dissociation of the carboxyl group) electrostatically adsorbs Ba⁺ sites on the barite surface, increasing the barite contact angle from 40° to 60° and rendering the surface slightly hydrophobic. Modified starch (carboxymethyl starch) hydrogen bonds with the fluorite surface, forming a polymer film above the calcium silicate layer, further reducing collector adsorption by approximately 30%. The three agents were sequentially adsorbed in a homogeneous system (modified water glass for 2 minutes, modified starch for 5 minutes, and citric acid for 3 minutes), creating a surface difference between fluorite with a strongly hydrophilic (contact angle of 30°) and barite with a weakly hydrophobic (contact angle of 60°), enabling the preferential floating of barite and the selective separation of fluorite from the ore pulp.

Claims

1. A method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore, characterized in that: include: Step 1: Collect process mineralogy data of high-calcium fluorite and barite paragenetic ore, the process mineralogy data including element content data output by X-ray fluorescence spectrometer, particle size distribution data and monomer dissociation degree correlation data output by mineral automatic analyzer, and corresponding data of inhibitor component, concentrate grade and recovery rate in historical flotation tests; Step 2: Convert the process mineralogy data into a structured format. The tabular data in the process mineralogy data is stored as database columns through field mapping. The key parameters of the graphic data are extracted through image recognition and stored as numerical fields. The sample number and acquisition timestamp metadata are added to form a process mineralogy database. Step 3: Using a machine learning algorithm to train an association model based on historical data on mineral properties and inhibitor composition in a process mineralogy database, the mineral property data includes calcite content, fluorite particle size distribution, and barite monomer dissociation degree; Step 4: receiving real-time mineral property data of the ore to be processed, inputting the real-time mineral property data of the ore to be processed into the correlation model, and obtaining a recommended inhibitor composition ratio; Step 5: Prepare a combined depressant by mixing raw materials according to the recommended ratio, record the flotation test results, and feed the test results back to the correlation model to update the model parameters. The flotation test results include concentrate grade, recovery rate, and mineral floating order data.

2. The method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to claim 1, wherein: The step 1 comprises: Calcite, fluorite and barite content distribution data output by X-ray fluorescence spectrometer; The mineral automatic analyzer outputs the data on the proportion of each mineral particle size range and the intergrowth ratio of fluorite, barite and calcite; The graphic data output by the mineral automatic analyzer is used to extract the particle size interval value and dissociation percentage parameters through image recognition.

3. The method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to claim 2, characterized in that: The step 3 comprises: The random forest algorithm was used to train the association model, and the input variables were the calcite content, fluorite particle size distribution, and barite monomer dissociation degree data from the process mineralogy database; The historical data in the process mineralogy database were divided into a training set and a validation set in a ratio of 7:

3. The training set was used to learn the correlation law between mineral characteristics and inhibitor ratios in the correlation model. The validation set is used to evaluate the prediction error of the association model and optimize the model stability by adjusting the tree depth and feature sampling ratio hyperparameters.

4. The method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to claim 3, characterized in that: The step 4 comprises: Real-time mineral property data of the ore to be processed include calcite content, fluorite particle size distribution range and barite monomer dissociation degree; Input the real-time mineral property data into the association model trained in step 3. If the calcite content is higher than the average threshold of the historical data in the process mineralogy database and the fluorite particle size is smaller than the median particle size of the historical data, the association model outputs a ratio for increasing the proportion of modified water glass; If the dissociation degree of the barite monomer is lower than the dissociation degree benchmark value of the historical data, the correlation model outputs a ratio of increasing the proportion of citric acid.

5. The method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to claim 4, characterized in that: The step 5 comprises: The raw materials of the combined inhibitor are modified water glass and citric acid, and the raw material ratio is determined based on the ratio of modified water glass to citric acid recommended by the correlation model in step 4; The raw materials were mixed by mechanical stirring at a speed of 100-150 rpm to uniformly disperse the modified water glass and citric acid; The stirring time is 15-20 minutes, and the mixing environment is at room temperature, so that the raw materials are fully mixed to form a homogeneous inhibitor system.

6. The method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to claim 5, characterized in that: Also includes: Apply the combined depressant prepared in step 5 to a flotation test, and record the floating order of fluorite and barite, the calcium fluoride content, the barium sulfate content, and the residual calcite in the concentrate during the test; Compare and analyze the flotation result data recorded in the test with the ratio recommended by the correlation model in step 4 to generate a feedback adjustment signal for adjusting the characteristic weight coefficient of the mineral characteristics to the inhibitor ratio in the correlation model in step 3; After the correlation model is updated, the inhibitor ratio recommendation logic for subsequent ore to be processed is optimized based on the updated feature weight coefficient to improve the matching degree between the recommended ratio and the actual flotation effect.

7. The method for preparing a combined inhibitor of high-calcium fluorite and barite paragenetic ore according to claim 6, characterized in that: Also includes: Modified starch is used to assist in enhancing the inhibitory effect of modified water glass on fluorite by covering the calcium active sites on the surface of fluorite; Modified water glass covers the calcium active sites on the surface of fluorite through chemical adsorption, which is used to reduce the affinity between fluorite and the collector, so that the fluorite surface remains hydrophilic; Citric acid is attached to the barium active sites on the barite surface through electrostatic interaction or hydrogen bonding, which is used to change the hydrophobicity of the barite surface.

Images