Precious metal refining solid waste recycling method and device

Abstract

The invention provides a precious metal refining solid waste recycling method and device, and relates to the technical field of precious metal recycling, and the method comprises the steps that component detection is conducted on first refining solid waste of a first target batch, and component information of the first batch is obtained; on the basis of the recovery rate prediction model and the current technological parameter set of the target scene, precious metal refining recovery rate prediction is conducted, and a predicted recovery rate set is obtained; performing double-target recovery economical efficiency screening, and determining a to-be-extracted category set; performing multi-objective optimization to obtain refining process parameters; predicting first-batch residual component information of the refining filter residues according to the refining process parameters, correspondingly evaluating the reflux economy of the refining filter residues, and determining filter residue reflux parameters according to the reflux economy; and the refining process parameters and the filter residue backflow parameters are synthesized, and reutilization of the first refined solid waste is executed. The technical problem that in the precious metal refining process in the prior art, the solid waste reutilization rate is not high is solved.

Description

Technical Field

[0001] This invention relates to the field of precious metal recycling and reuse, and in particular to a method and apparatus for reusing solid waste from precious metal refining. Background Technology

[0002] In the precious metal smelting industry, anode mud, as a key byproduct generated during electrolytic refining, is enriched with various rare and precious metals such as gold, silver, platinum, palladium, and rhodium, making it a core resource for recovering rare and precious metals and possessing extremely high economic value.

[0003] However, existing precious metal refining processes mostly rely on fixed empirical parameters or a single process route, making it difficult to dynamically adapt to the compositional differences of different batches of anode mud. Fixed parameters can easily lead to fluctuations in the recovery rate of the target precious metal, or cause a surge in costs due to excessive reagent input. At the same time, existing technologies lack a systematic quantitative assessment of the residual value of refining filter residue, failing to fully explore the potential value of unrecovered precious metals in the filter residue, resulting in low solid waste reuse rates and causing unnecessary economic losses.

[0004] Therefore, there is an urgent need for an intelligent and precise method for reusing solid waste from precious metal refining to improve resource recycling efficiency and overall economic efficiency. Summary of the Invention

[0005] This invention addresses the technical problem of low solid waste recycling rate in existing precious metal refining processes by providing a method and apparatus for recycling solid waste from precious metal refining.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a method for recycling solid waste from precious metal refining, comprising: The composition of the first batch of refined solid waste in the first target batch was analyzed to obtain the composition information of the first batch, wherein the first batch of refined solid waste is anode mud; Based on the pre-constructed recovery rate prediction model and the current process parameter set of the target scenario, the precious metal refining recovery rate is predicted by combining the composition information of the first batch, and a set of predicted recovery rates is obtained. Based on the predicted recovery rate set, a dual-objective recovery economic screening is performed to determine the set of species to be extracted; By combining the recovery rate prediction model, the current process parameter set, and the set of species to be extracted, multi-objective optimization is performed to obtain refining process parameters. Based on the refining process parameters, the residual component information of the first batch of refining filter residue is predicted, the recirculation economy of the refining filter residue is evaluated accordingly, and the filter residue recirculation parameters are determined based on the recirculation economy. By combining the refining process parameters and the filter residue reflux parameters, the first refining solid waste is reused.

[0007] Secondly, the present invention provides a device for reusing solid waste from precious metal refining, comprising: a memory for storing computer software programs; and a processor for reading and executing the computer software programs, thereby realizing a method for reusing solid waste from precious metal refining.

[0008] The beneficial effects of this invention are: Compared to existing technologies, this application first performs component analysis on the first batch of refined solid waste in the first target batch to obtain the component information of the first batch, providing a reliable data foundation for subsequent recovery rate prediction and process optimization. Secondly, based on a pre-constructed recovery rate prediction model and the current process parameter set of the target scenario, combined with the component information of the first batch, the recovery rate of precious metal refining is predicted, obtaining a predicted recovery rate set. The recovery rate prediction model quantifies the correlation between component information, process parameters, and recovery rate, avoiding the subjectivity of traditional manual experience estimation and providing reliable key data support for subsequent analysis. Thirdly, based on the predicted recovery rate set, a dual-objective recovery economic screening is performed to determine the set of extractable species, avoiding ineffective extraction of low-content or high-content low-yield materials, thus improving the economy and efficiency of resource utilization. Furthermore, multi-objective optimization is performed by combining the recovery rate prediction model, the current process parameter set, and the set of extractable species to obtain refining process parameters, providing accurate operational basis for the efficient recovery of the precious metals to be extracted. Furthermore, based on the refining process parameters, the residual composition information of the first batch of refining filter residue is predicted, and the economics of recycling the filter residue is evaluated accordingly. Based on this economics, filter residue recycling parameters are determined. By quantitatively evaluating the economics of recycling the refining filter residue, reasonable recycling parameters are established to achieve a balance between maximizing resource utilization and cost control. Finally, by combining the refining process parameters and the filter residue recycling parameters, the first stage of refining solid waste is reused, improving the overall resource utilization rate and treatment economics of solid waste.

[0009] Through the above technical solution, this application obtains the first batch composition information through component detection, predicts the precious metal refining recovery rate using a recovery rate prediction model combined with current process parameters, identifies a high-value extraction category set through dual-objective recovery economic screening, determines suitable refining process parameters through multi-objective optimization, predicts the residual composition information of the refining filter residue based on the refining process parameters, evaluates the economics of recirculation to determine differentiated filter residue recirculation parameters, and finally integrates the refining process parameters and filter residue recirculation parameters to implement the reuse of the first batch of refining solid waste. In this way, it improves the single-batch recovery efficiency through precise process parameters and taps residual value through differentiated filter residue recirculation strategies, thereby increasing the resource utilization rate and economic value of refining solid waste and achieving efficient and economical reuse of precious metal refining solid waste. Attached Figure Description

[0010] Figure 1 A schematic flowchart of a method for recycling solid waste from precious metal refining provided by the present invention; Figure 2 This is a schematic diagram of a device for recycling solid waste from precious metal refining, provided by the present invention.

[0011] In the attached diagram, the components represented by each number are as follows: A device 200 for recycling solid waste from precious metal refining includes a memory 210, a processor 220, and a computer program 211. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0014] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0015] Example 1, as Figure 1 As shown, this embodiment of the invention provides a method for recycling solid waste from precious metal refining, comprising: S10: Perform component analysis on the first refined solid waste of the first target batch to obtain the component information of the first batch, wherein the first refined solid waste is anode mud.

[0016] In traditional precious metal smelting, anode slime is a key byproduct for enriching precious metals such as gold, silver, platinum, and palladium. Its recovery decisions often rely on human experience and are difficult to dynamically adapt to the compositional differences between different batches of anode slime. For example, the content of each precious metal fluctuates between different batches, and the proportion of impurities changes. This leads to large fluctuations in the precious metal recovery rate, and some recyclable resources are lost with the filter residue, resulting in low overall resource utilization and economic benefits.

[0017] To address the aforementioned issues, this application randomly selects a batch from the target batches as the first target batch, and performs component analysis on the first refined solid waste of the first target batch to obtain the component information of the first batch, wherein the first refined solid waste is anode mud.

[0018] Specifically, step S10 in the method includes: The first refined solid waste is pretreated, and samples are taken based on the pretreatment results to obtain the first sample solid. Based on the first sample solid, the precious metal recycling list of the target scene is traversed to perform content detection, forming a precious metal content composition vector of the first target batch, and the output is the composition information of the first batch.

[0019] In this embodiment, the first refined solid waste is pretreated, and samples are taken based on the pretreatment results to obtain a first sample solid. Pretreatment of the first refined solid waste is necessary because anode mud, as a byproduct of precious metal smelting, typically exhibits problems such as uneven particle size, uneven composition distribution, and the presence of moisture or viscous substances. Direct sampling would result in a sample that cannot represent the true composition of the entire batch of material. Therefore, pretreatment through steps such as crushing, grinding, homogenization, and drying is required to eliminate interference from uneven particle size and other issues, ultimately obtaining a pretreated result with relatively uniform physical properties and composition distribution. For example, sampling based on the pretreatment results can be performed using a multi-point random sampling and mixing method (i.e., taking multiple sub-samples from multiple different locations in the pretreatment results and then mixing them into a comprehensive sample to obtain a first sample solid that accurately reflects the compositional characteristics of the first target batch of anode mud, avoiding detection distortion caused by local component deviations).

[0020] Secondly, based on the first solid sample, the content of precious metals in the target scenario's precious metal recycling list is analyzed to form a precious metal content composition vector for the first target batch, which is then output as the first batch's composition information. The precious metal recycling list refers to a list of technically recyclable and economically valuable precious metals in different smelting scenarios. For example, a target scenario's precious metal recycling list might include gold, silver, platinum, palladium, rhodium, ruthenium, and iridium. This list is determined based on the scenario's technological capabilities, equipment conditions, and market demand, and represents the target range for composition analysis. The content analysis, which involves traversing the target scenario's precious metal recycling list, is performed on the first solid sample, analyzing the content of each precious metal listed in the list. This analysis utilizes high-precision instrumental analysis methods such as atomic absorption spectrometry and inductively coupled plasma mass spectrometry to ensure the accuracy and precision of the content analysis.

[0021] For example, based on the first sample solid, the precious metal recovery list (such as gold, silver, and platinum) of the target scenario is traversed, and the content of each precious metal is detected one by one by atomic absorption spectrometry to obtain the content of each precious metal, such as gold 2.3g / ton, silver 45g / ton, and platinum 0.7g / ton. The content is then integrated in the form of a vector to form the precious metal content composition vector of the first target batch, such as [gold 2.3g / ton, silver 45g / ton, platinum 0.7g / ton]. The output is the composition information of the first batch, which can intuitively and clearly present the composition information of the first target batch of anode mud.

[0022] In summary, compared to existing technologies, this application performs component analysis on the first batch of refined solid waste (anode mud) to obtain the component information for the first batch. Thus, through standardized pretreatment, sampling, and testing procedures, the obtained component information for the first batch is ensured to be accurate, comprehensive, and representative, providing a reliable data foundation for subsequent recovery rate prediction, process optimization, and other related processes.

[0023] S20: Based on the pre-built recovery rate prediction model and the current process parameter set of the target scenario, combined with the composition information of the first batch, the recovery rate of precious metal refining is predicted to obtain the predicted recovery rate set.

[0024] The composition of anode slime, including the content of various precious metals and the proportion of impurities, along with process parameters such as leaching temperature, reagent concentration, and reaction time, all affect the recovery rate of precious metals. Traditional methods often employ fixed process paths and rely on manual experience to judge process parameters, making it difficult to accurately quantify the combined impact of composition and process parameters on the recovery rate. This results in poor stability of precious metal recovery rates. For example, some low-content but recoverable precious metals may be lost with the filter residue due to insufficient process adaptability, ultimately leading to low resource utilization and limited overall economic benefits.

[0025] To address the aforementioned issues, this application uses a pre-constructed recovery rate prediction model and the current process parameter set of the target scenario, combined with the composition information of the first batch, to predict the recovery rate of precious metal refining and obtain a set of predicted recovery rates.

[0026] Specifically, step S20 in the method includes: The recovery rate prediction model includes a property-driven prediction model and a data-driven prediction model. The construction steps of the recovery rate prediction model include: Based on the first batch composition information, a precious metal component is randomly selected as the first prediction object, and the first sample data of the first prediction object is obtained accordingly. The first sample data includes sample refining process parameters, sample batch composition information and sample recovery rate. Based on the first sample data, the prior property model is matched and invoked, and the candidate property prediction model of the first prediction object is constructed accordingly. The prediction performance of the candidate property prediction model is verified and the prediction performance of the candidate property is obtained. If the performance of the candidate physical property prediction meets the target prediction performance requirements, the candidate physical property prediction model is output as the recovery rate prediction model, and the precious metal components are iteratively selected by traversing the first batch of component information.

[0027] In this embodiment, based on the first batch composition information, a precious metal component is randomly selected as the first prediction target, and the first sample data of the first prediction target is obtained accordingly. The first sample data includes sample refining process parameters, sample batch composition information, and sample recovery rate. The sample refining process parameters refer to the process conditions used in the historical extraction of the first prediction target, such as temperature, time, reagent concentration, pH value, etc. The sample batch composition information refers to the content of the first prediction target and the composition data of other associated elements and impurities in the historical batches. The sample recovery rate refers to the actual recovery rate under the above sample refining process parameters and sample batch composition information conditions. For example, based on the first batch composition information (e.g., [gold 2.3g / ton, silver 45g / ton, platinum 0.7g / ton]), a precious metal component (e.g., gold) is randomly selected as the first prediction object, and the corresponding sample refining process parameters (e.g., cyanide concentration 0.05mol / L, leaching temperature 60℃, stirring rate 300r / min, leaching time 4 hours), sample batch composition information (e.g., gold 2.1g / ton, silver 42g / ton, copper 1.5%, lead 0.8%) and sample recovery rate (e.g., 90%) are obtained as the first sample data.

[0028] Secondly, based on the first sample data, a priori property models are matched and invoked to construct alternative property prediction models for the first prediction object. The predictive performance of these alternative property prediction models is then verified, and their predictive performance is obtained. The priori property model refers to a prediction model established based on the inherent physicochemical properties of a substance, such as reaction thermodynamics, kinetic characteristics, phase equilibrium laws, and solubility, combined with long-accumulated theoretical knowledge and practical experience. It derives the correlation between process parameters and target results from a theoretical perspective using known property laws. In this application, the priori property model refers to a pre-generated basic model based on the physicochemical properties of noble metals, such as dissolution kinetics and reaction equilibrium constants. Examples include the dissolution rate model of gold in cyanide solution and the leaching equilibrium model of platinum in a specific acid. Priori property models do not rely on statistical fitting of large amounts of experimental data but derive results through existing theoretical knowledge. They are particularly suitable for reaction processes supported by mature property laws and can provide a theoretical basis for recovery prediction.

[0029] For example, based on the first sample data, a priori physical property model related to gold recovery is matched and invoked, such as the gold dissolution rate model in cyanide solution (this model presupposes the theoretical relationship between cyanide ion concentration, temperature, and gold dissolution rate based on the reaction mechanism, and the core formula includes parameters such as reaction rate constant and activation energy). Then, the parameters of the priori physical property model are calibrated. For example, the actual gold dissolution rate at a cyanide concentration of 0.05 mol / L and a temperature of 60°C in the first sample data is used to back-calculate and adjust the reaction rate constant in the priori physical property model based on the principle, such as from the theoretical value of 0.002 min. -1 Corrected to a more realistic 0.0018 min -1 Meanwhile, key parameters such as activation energy are calibrated to obtain alternative property prediction models for gold recovery.

[0030] For example, to verify the predictive performance of the candidate property prediction model, the sample refining process parameters and sample batch composition information in the first sample data can be input into the candidate property prediction model, and the predicted sample recovery rate (e.g., 88%) can be output. The prediction performance can be evaluated by calculating the error between the predicted sample recovery rate and the actual sample recovery rate, such as calculating the absolute error, mean square error, etc. For example, the absolute error between the predicted sample recovery rate of 88% and the actual sample recovery rate of 90% is 2%, which can be used as the candidate property prediction performance.

[0031] Finally, if the performance of the candidate physical property prediction meets the target prediction performance requirement, the candidate physical property prediction model is output as the recovery rate prediction model, and the precious metal components are iteratively selected by traversing the first batch of component information. The target prediction performance requirement is a predefined quantitative standard based on the actual application scenario, such as an absolute error ≤3%. Those skilled in the art can dynamically set this standard according to recovery accuracy requirements, process stability requirements, etc. For example, if the performance of the candidate physical property prediction (e.g., an absolute error of 2%) meets the target prediction performance requirement (e.g., an absolute error ≤3%), the candidate physical property prediction model is determined as the recovery rate prediction model for the first prediction object. Then, the above complete steps are repeated for other precious metals (e.g., silver, platinum) in the first batch of component information: randomly selecting prediction objects, obtaining corresponding sample data, matching and calling prior physical property models, prediction performance, etc., ultimately forming a set of candidate physical property prediction models covering all precious metals in the first batch of component information and all meeting the target prediction performance requirement, which together serve as the recovery rate prediction model.

[0032] Furthermore, the construction steps of the recovery rate prediction model also include: If the performance of the alternative property prediction does not meet the target prediction performance requirement, then based on the data-driven method, using the first sample data as supervision and combined with the first preset model parameter constraints, an alternative data prediction model is constructed and trained. Verify whether the prediction performance of the candidate data prediction model meets the target prediction performance requirements; If the conditions are not met, the first preset model parameter constraints are adjusted, and the alternative data prediction model is iteratively constructed and trained until the prediction performance of the alternative data prediction model meets the target prediction performance requirements, and the output is the recovery rate prediction model.

[0033] In this embodiment, if the performance of the candidate physical property prediction does not meet the target prediction performance requirements (possibly due to the complexity of the reaction mechanism of precious metals, making it difficult for the physical property-driven prediction model to accurately describe it), a candidate data prediction model is constructed and trained based on a data-driven method, using the first sample data as supervision and combined with the first preset model parameter constraints. The first preset model parameter constraints are pre-set restrictions on model parameters when constructing the candidate data prediction model based on the data-driven method. For example, restrictions may include limiting the number of neural network layers to ≤5, the depth of the decision tree to ≤10, and the maximum number of iterations to ≤5000. This ensures that the candidate data prediction model conforms to the process logic, avoids overfitting or meaningless parameter values, and improves the stability and generalization ability of model training. The first preset model parameter constraints are pre-set by those skilled in the art based on the process characteristics, data distribution characteristics, and model type (such as machine learning models, statistical models, etc.) of precious metal recycling.

[0034] For example, a candidate data prediction model can be constructed based on a neural network architecture, combined with first preset model parameter constraints (such as the number of neural network layers ≤ 5 layers and the maximum number of iterations ≤ 5000). The candidate data prediction model mainly consists of an input layer, a hidden layer, an output layer, and a regularization module. The input layer contains N neurons, where N is set according to the number of key features of the first sample data. The hidden layer is set to 3 layers, with 64, 32, and 16 neurons in each layer, respectively. Each layer uses the ReLU activation function to enhance nonlinear expression and alleviate the gradient vanishing problem. The output layer has 1 neuron, which is mapped to the 0-100% range through the Sigmoid function as the prediction recovery rate. The regularization module introduces L2 regularization to prevent overfitting.

[0035] For example, during training, the first sample data is divided into a training set, a validation set, and a test set in a 7:1.5:1.5 ratio. The data in the training set is used as supervision, and the Adam optimizer (learning rate) is employed. [0.0001, 0.01]) Minimize the mean squared error loss function. Every 50 iterations, evaluate the model performance using the validation set and dynamically adjust the learning rate. At the same time, verify the generalization ability on the test set. Combined with the constraint of the maximum number of iterations ≤ 5000 in the first preset model parameter constraint, when the test set loss fluctuation is less than 1e-5 in 100 consecutive iterations, the model is considered to have converged, and the candidate data prediction model is obtained.

[0036] Secondly, verify whether the predictive performance of the candidate data prediction model meets the target predictive performance requirements. For example, to verify the predictive performance of the candidate data prediction model, several samples from the test set that were not involved in model training, including refining process parameters and batch composition information, can be input into the candidate data prediction model. The model will output several predicted sample recovery rates. The errors between these predicted sample recovery rates and the actual sample recovery rates (e.g., calculating absolute error, mean square error, etc.) are then calculated, and the mean error is used as the predictive performance of the candidate data prediction model. Finally, verify whether the predictive performance of the candidate data prediction model meets the target predictive performance requirements. If it does, the candidate data prediction model is output as the recovery rate prediction model for the first prediction object.

[0037] Finally, if the conditions are not met, the first preset model parameter constraints are adjusted, for example, by relaxing the model complexity restrictions, increasing the number of neural network layers, adjusting the regularization parameters, etc. Then, based on the adjusted first preset model parameter constraints, the candidate data prediction model is iteratively constructed and trained until the prediction performance of the candidate data prediction model meets the target prediction performance requirements, and the output is the recovery rate prediction model.

[0038] Thus, by using a dual-path modeling approach driven by physical properties and data, a high-precision recovery rate prediction model can be built for different precious metals, solving the problems of poor adaptability and insufficient accuracy of traditional prediction methods.

[0039] Furthermore, based on the pre-built recovery rate prediction model and the current process parameter set of the target scenario, combined with the first batch composition information, the recovery rate of precious metal refining is predicted to obtain a set of predicted recovery rates. Specifically, the recovery rate prediction model is a specific prediction model built for different precious metals. By inputting the current process parameter set of the target scenario and the first batch composition information, the recovery rate of precious metal refining can be predicted. For example, the prediction process is carried out one by one according to the type of precious metal. For example, for gold, the pre-trained recovery rate prediction model is called, and the current process parameters (such as cyanide concentration 0.05mol / L, leaching temperature 60℃, stirring rate 300r / min, leaching time 4 hours) and the first batch composition information (such as gold 2.3g / ton, silver 45g / ton, platinum 0.7g / ton) are input, and the predicted recovery rate of gold is output, such as 91%. The same method is used to predict the recovery rate of other precious metal elements, and several predicted recovery rates are obtained. These are combined to obtain a set of predicted recovery rates.

[0040] In summary, compared to existing technologies, this application uses a pre-constructed recovery rate prediction model and the current set of process parameters for the target scenario, combined with the composition information of the first batch, to predict the recovery rate of precious metal refining and obtain a set of predicted recovery rates. In this way, the correlation between composition information, process parameters, and recovery rate is quantified through the recovery rate prediction model, avoiding the subjectivity of traditional manual estimation and providing reliable key data support for subsequent analysis.

[0041] S30: Based on the predicted recovery rate set, perform a dual-objective recovery economic screening to determine the set of species to be extracted.

[0042] In the process of precious metal recycling, there are two situations where there is no actual recycling value: some precious metals may have too low a content (such as far below the industry-recognized minimum economic content), and even if 100% recycling is achieved, the value corresponding to the total amount cannot cover the extraction cost, thus having no recycling value; other precious metals may have no actual recycling significance even if the content meets the standard, because the total cost in the recycling process (such as equipment depreciation, labor, chemicals, energy, etc.) exceeds the economic value generated by recycling, resulting in negative net income or failure to reach the preset profit threshold.

[0043] To address the aforementioned issues, this application employs a dual-objective recovery economics screening based on the predicted recovery rate set to determine the set of species to be extracted.

[0044] Specifically, step S30 in the method includes: With the minimum economic content as the primary objective, an initial economic screening was conducted based on the component information of the first batch. Obtain the constant and variable costs of multiple refining process production lines, and combine the predicted recovery rate with the total batch size of the first target batch to perform an initial economic calculation on the economic screening results. If the result of the economic calculation is greater than the preset economic constraint, the corresponding precious metal component will be added to the set of types to be extracted. The minimum economic content, the constant cost, and the variable cost are prior data.

[0045] In this embodiment, the primary objective is to determine the minimum economic content, and an initial economic screening is conducted based on the composition information of the first batch. The minimum economic content refers to the lowest concentration critical value at which a certain precious metal is worth extracting. If the content is below the minimum economic content, even with 100% recovery, its value cannot cover the extraction cost. The minimum economic content is prior data and can be obtained based on industry consensus, historical measured data, expert definitions, etc. For example, by analyzing loss-making cases from the past three years, the minimum economic content is determined to be 0.5 g / ton for gold, 5 g / ton for silver, and 0.4 g / ton for platinum. For example, the content of each precious metal in the first batch of component information is compared with the minimum economic content. Precious metals with a content less than the minimum economic content are eliminated, and precious metals with a content greater than or equal to the minimum economic content are retained to obtain the initial economic screening results. For example, comparing the content of each precious metal in the first batch of component information [gold 2.3g / ton, silver 45g / ton, platinum 0.7g / ton] with the minimum economic content, since gold 2.3g / ton ≥ 0.5g / ton, silver 45g / ton ≥ 5g / ton, and platinum 0.7g / ton ≥ 0.4g / ton, gold, silver, and platinum are retained. In this way, precious metals with too low a content in the first batch of component information that would inevitably result in a loss in extraction are eliminated by using the minimum economic content.

[0046] Secondly, the fixed and variable costs of multiple refining process lines are obtained, and combined with the predicted recovery rate and the total batch size of the first target batch, an initial economic feasibility study is conducted to assess the economic viability of the preliminary screening results. Fixed costs are prior data, such as depreciation of production line equipment and fixed labor costs, which do not change with the processing volume. These costs can be obtained from the company's financial data. For example, if a precious metal recycling line has an average monthly fixed cost of 500,000 yuan, the fixed cost for processing the first target batch is 50,000 yuan, allocated according to the processing volume ratio. Variable costs are also prior data, such as reagent consumption and energy costs, which change with the processing volume. These costs can be obtained from historical process data. For example, the variable cost of extracting 1g of gold is 80 yuan. The initiation economic accounting is based on the predicted recovery rate, total batch volume, precious metal content, and precious metal price to calculate the expected recovery value. Expected recovery value = predicted recovery rate × total batch volume × precious metal content × precious metal price. Combined with the total cost calculated by constant cost and variable cost, the total cost = constant cost + variable cost is obtained by subtracting the constant cost from the variable cost. The initiation economic accounting result = expected recovery value - total cost. The initiation economic accounting result can reflect the net benefit of recovering the precious metal. For example, if the predicted recovery rate is 91%, the total batch size of the first target batch is 100 tons, and the gold content is 2.3g / ton, if the fixed cost is 50,000 yuan, the variable cost is 91% × 100 × 2.3 × 80 = 16,744 yuan, and the price of gold is 500 yuan / g, then the expected recovery value is 91% × 100 × 2.3 × 500 = 104,650 yuan, and the total cost is 50,000 + 16,744 = 66,744 yuan. Therefore, the result of the initial economic calculation is 104,650 - 66,744 = 37,906 yuan. The initial economic calculation result reflects the net profit from recovering the precious metal. Thus, the actual profitability of recovering the precious metal can be verified through the initial economic calculation result, avoiding the problem of insufficient recovery profit.

[0047] Finally, if the initial economic calculation result exceeds the preset economic constraint, the corresponding precious metal component is added to the set of extractable categories. The preset economic constraint is a quantitative threshold for determining whether precious metal recycling has actual profit value. It can be dynamically adjusted based on actual needs such as resource priority, investment return requirements, and market fluctuations in the target scenario. For example, a net profit of 30,000 yuan can be used as the preset economic constraint. Those skilled in the art can flexibly adjust this constraint based on variables such as real-time metal prices, production line load, and cost structure, thereby ensuring that resources are prioritized for precious metals with higher recycling returns, avoiding the occupation of limited resources such as equipment and manpower by low-return extraction activities. For example, if the initial economic calculation result for a precious metal exceeds the preset economic constraint, the corresponding precious metal component is added to the set of extractable categories; conversely, if the initial economic calculation result is less than or equal to the preset economic constraint, it is excluded. For instance, if the initial economic calculation result for gold is 37,906 yuan and the preset economic constraint is 30,000 yuan, gold is added to the set of extractable categories because the initial economic calculation result for gold exceeds the preset economic constraint.

[0048] In summary, compared to existing technologies, this application uses a dual-objective economic screening method based on the predicted recovery rate set to determine the set of extractable species. Thus, through initial economic screening, initiating economic accounting, and setting preset economic constraints, the extractable species are selected, avoiding ineffective extraction of low-content or high-content low-yield species, thereby improving the economy and efficiency of resource utilization.

[0049] S40: Combine the recovery rate prediction model, the current process parameter set, and the set of types to be extracted to perform multi-objective optimization to obtain refining process parameters.

[0050] Traditional refining process parameters are mostly fixed empirical values. They cannot be dynamically adjusted according to the compositional differences of different batches of anode mud, nor can they be predicted by quantitative models to determine the impact of process parameter changes on the recovery rate. This often leads to problems such as poor recovery rate stability, uncontrolled costs, and large fluctuations in revenue in actual production. Furthermore, it is impossible to achieve a precise balance between recovery rate, cost, and revenue.

[0051] To address the aforementioned issues, this application combines the recovery rate prediction model, the current process parameter set, and the set of species to be extracted to perform multi-objective optimization to obtain refining process parameters.

[0052] Specifically, step S40 in the method includes: Construct a process parameter optimization function, wherein the process parameter optimization function includes at least a recovery rate factor, a relative cost factor, and a relative profit factor; Using the set of categories to be extracted as an index, the current set of process parameters is traversed to extract process parameters, and the recovery rate prediction model is traversed to call the model. Combining the process parameter optimization function with the model call results, the extracted process parameters are iteratively optimized in multiple objectives until the function value of the process parameter optimization function converges, and the corresponding multi-objective optimization result is output as the refining process parameters.

[0053] In this embodiment, a process parameter optimization function is first constructed. This function includes at least a recovery rate factor, a relative cost factor, and a relative profit factor. The recovery rate factor reflects the difference between the actual recovery and the theoretical optimal level. Recovery rate factor = predicted recovery rate / theoretical maximum recovery rate. For example, if the predicted recovery rate under a certain process parameter is 92%, and the theoretical maximum recovery rate under that process parameter is 98%, then the recovery rate factor = 92% / 98% ≈ 0.94. The larger the recovery rate factor, the closer the actual recovery is to the theoretical optimal level. The relative cost factor measures the deviation between the actual cost and the industry benchmark cost. Relative cost factor = actual cost / benchmark cost. For example, if the actual cost under a certain process parameter is 1... The cost factor is 1.5 million yuan, and the benchmark cost for similar process parameters is 1.2 million yuan. Therefore, the relative cost factor = 1.5 / 1.2 = 1.25. The smaller the relative cost factor, the better the cost control. The relative profit factor reflects the difference between the actual net profit and the ideal net profit. The relative profit factor = actual net profit / ideal net profit. For example, if the actual net profit of a batch is 120,000 yuan, and the ideal net profit under the theoretical optimal condition is 150,000 yuan, then the relative profit factor = 120,000 / 150,000 = 0.8. The higher the relative profit factor, the closer the actual net profit is to the ideal state. The recovery rate factor, relative cost factor, and relative profit factor are all dimensionless data, ensuring that indicators of different dimensions can be directly calculated. For example, the process parameter optimization function can be obtained by weighted integration of the recovery rate factor, relative cost factor, and relative profit factor. For instance, the process parameter optimization function = α × recovery rate factor + β × (1 / relative cost factor) + γ × relative profit factor, where α, β, and γ are the weights of the recovery rate factor, relative cost factor, and relative profit factor, respectively. These weights can be dynamically adjusted according to actual production priorities; for example, α can be set to 0.3, β to 0.3, and γ to 0.4. In this way, the process parameter optimization function transforms the three objectives of improving recovery rate, controlling costs, and increasing profits into calculable quantitative indicators.

[0054] Secondly, using the set of species to be extracted as an index, the current set of process parameters is traversed to extract the process parameters for each species. Then, the recovery prediction models are traversed and called upon. For example, using the set of species to be extracted (e.g., {gold, silver}) as an index, process parameters related to gold and silver are extracted from the current set of process parameters, such as cyanide concentration and leaching time for gold, and nitric acid concentration and stirring rate for silver. Then, the recovery prediction models corresponding to the species to be extracted are matched, such as the cyanide dissolution model for gold and the nitric acid leaching model for silver, to ensure that the recovery rates under different process parameters are covered.

[0055] Finally, combining the process parameter optimization function and the model call results, iterative multi-objective optimization is performed on the extracted process parameters until the function value of the process parameter optimization function converges, and the corresponding multi-objective optimization result is output as the refining process parameters. For example, a set of process parameters is randomly selected from the extracted process parameters. For instance, for gold elements in the extraction set, the randomly selected process parameters are cyanide 0.05 mol / L, temperature 60℃, stirring rate 300 r / min, and reaction time 4 hours. This set of process parameters and the first batch composition information are input into the gold element recovery rate prediction model, which outputs the predicted recovery rate. Then, the recovery rate factor, relative cost factor, and relative benefit factor are calculated according to the above method. Combined with the process parameter optimization function, the function value of the process parameter optimization function is obtained, such as 0.82. Then, intelligent optimization algorithms, such as particle swarm optimization, are applied. Simulated annealing algorithms, etc., extract process parameters from the process parameter extraction results, recalculate the function value of the process parameter optimization function, repeat this process, continuously optimize the process parameters until the function value fluctuation of the process parameter optimization function is ≤0.001 in 30 consecutive iterations, then convergence is determined, and the process parameters at this time are output as refining process parameters. For example, the refining process parameters for gold are cyanide concentration 0.055mol / L, leaching temperature 62℃, stirring rate 320r / min, and reaction time 3.5 hours, while the refining process parameters for silver are nitric acid concentration 12%, pH value 1.2, and reaction time 2.8 hours.

[0056] In summary, compared to existing technologies, this application combines the recovery rate prediction model, the current process parameter set, and the set of extractable species for multi-objective optimization to obtain refining process parameters. Thus, through iterative optimization, the final refining process parameters can find the optimal balance between recovery rate, cost, and benefit, providing precise operational guidance for the efficient recovery of precious metals.

[0057] S50: Based on the refining process parameters, predict the residual component information of the first batch of refining filter residue, evaluate the recirculation economy of the refining filter residue accordingly, and determine the filter residue recirculation parameters based on the recirculation economy.

[0058] In the process of recovering refining filter residue, problems such as the low value of residual precious metals and the high cost of recirculation may lead to a recovery benefit lower than the cost of recirculation, resulting in ineffective resource input. Traditional methods often adopt a crude treatment mode of full recirculation or full disposal, leading to ineffective recovery or resource waste.

[0059] To address the aforementioned issues, this application predicts the residual component information of the first batch of refining filter residue based on the refining process parameters, evaluates the economic efficiency of recirculating the refining filter residue accordingly, and determines the filter residue recirculation parameters based on the economic efficiency of recirculation.

[0060] The first batch of residual component information can be obtained through a pre-trained residual component information prediction model. By inputting refining process parameters, the residual component information prediction model predicts and outputs the first batch of residual component information in the refining filter residue.

[0061] For example, the residual component information prediction model can be built based on a random forest architecture, mainly consisting of an input layer, an ensemble learning module composed of 100 CART decision trees, a feature importance evaluation module, and an output layer. The input layer is responsible for standardizing refining process parameters and selecting features. The ensemble learning module learns in parallel using 100 decision trees (each tree randomly selects 70% of the samples and 60% of the features for training) to reduce the risk of overfitting. The feature importance evaluation module quantifies the impact of refining process parameters on the prediction results. The output layer transforms the model's raw output into structured data. During training, data pairs containing refining process parameters and residual component information are extracted from historical batch production data as a sample set. The sample set is divided into a training set and a validation set in an 8:2 ratio. Key hyperparameters are optimized through grid search to determine the optimal parameter combination, such as a tree depth of 12, a minimum number of split samples of 8, and a minimum number of samples per leaf node of 3. Convergence is achieved when the mean squared error on the validation set is stably controlled within 0.05, thus obtaining the residual component information prediction model.

[0062] For example, by inputting refining process parameters (gold: cyanide concentration 0.055 mol / L, leaching temperature 62℃, stirring rate 320 r / min, reaction time 3.5 hours; silver: nitric acid concentration 12%, pH value 1.2, reaction time 2.8 hours) into the residual component information prediction model, the model predicts and outputs structured information on the first batch of residual components: residual amount of precious metals (gold 0.1 g / ton, silver 2.45 g / ton, platinum 0.03 g / ton), residual impurity percentage (copper 0.32%, lead 0.08%, iron 0.15%), and physicochemical properties (pH=9.3, moisture content 4.8%, density 1.17 g / cm³). The information on the first batch of residual components provides basic data for calculating the residual value of precious metals and also provides key parameters for the subsequent calculation of the conditioning effect coefficient.

[0063] Specifically, step S50 in the method includes: Based on the residual component information of the first batch, the residual value of precious metals, the cost of using the reflux equipment, the cost of reflux operation, and the conditioning effect coefficient of the refined filter residue are calculated respectively. The residual value of the precious metal is calculated based on real-time market prices; the cost of using the reflux equipment includes the cost of equipment wear and tear and operating time; the cost of reflux operation includes the cost of labor; and the conditioning effect coefficient includes the conditioning substitution rate of pH, density, and moisture content. The economics of reflux are calculated based on the residual value of the precious metal, the cost of using the reflux equipment, the cost of reflux operation, and the conditioning effect coefficient. When the reflux economy is greater than the first reflux threshold, the reflux ratio of the refining filter residue is determined to be 1; When the economic efficiency of the reflux is between the first reflux threshold and the second reflux threshold, the reflux ratio is calculated based on the conditioning demand and the conditioning mapping model. When the reflux economy is less than the second reflux threshold, the reflux ratio of the refining filter residue is determined to be 0; Based on the reflux ratio, the filter residue reflux parameters are generated.

[0064] In this embodiment, based on the first batch of residual component information, the residual value of precious metals, the cost of using the reflux equipment, the cost of reflux operation, and the conditioning effect coefficient of the refined filter residue are calculated. The residual value of precious metals is calculated based on real-time market prices; the cost of using the reflux equipment includes the cost of equipment wear and tear and operating time; the cost of reflux operation includes labor costs; and the conditioning effect coefficient includes the conditioning substitution rate for pH, density, and moisture content. For example, based on the first batch of residual component information, such as 0.2 g / ton of gold, 3.5 g / ton of silver, pH=9, and 15% moisture content in the filter residue, the residual value of precious metals, the cost of using the reflux equipment, the cost of reflux operation, and the conditioning effect coefficient are calculated.

[0065] Among them, the residual value of precious metals refers to the economic value of the precious metals that have not been fully extracted from the refining filter residue at the real-time market price. It can be obtained by summing the products of the residual amount of each precious metal in the refining filter residue and the real-time market price. It can reflect the potential recovery value of the residual precious metals in the refining filter residue. For example, if the total amount of refining filter residue is 50 tons, the current gold price is 500 yuan / g, and the current silver price is 10 yuan / g, then the residual value of precious metals = 50 × 0.2 × 500 + 50 × 3.5 × 10 = 6750 yuan.

[0066] The cost of using the recirculation equipment includes the cost of equipment wear and tear and the cost of operating time. The cost of equipment wear and tear can be calculated based on the equipment depreciation rate and the loss coefficient of a single recirculation (preliminary data, such as 0.01% / time). For example, if the original value of the equipment is 1 million yuan, the cost of equipment wear and tear per recirculation = 1 million yuan × 0.01% = 100 yuan. The cost of operating time can be obtained by multiplying the energy consumption per unit time of the equipment by the recirculation processing time. For example, if the energy consumption per unit time is 200 yuan / hour, and the recirculation processing time is 2 hours, the cost of operating time = 200 × 2 = 400 yuan. Therefore, the cost of using the recirculation equipment = 100 + 400 = 500 yuan.

[0067] Among them, the cost of reflux operation refers to the labor cost in the process of filter residue reflux. It can be calculated based on the number of people involved in the operation, working hours and wages. For example, if the hourly wage is 30 yuan / hour and 3 people work for 2 hours, then the cost of reflux operation = 3 × 2 × 30 = 180 yuan.

[0068] The conditioning effect coefficient measures the value of refining filter residue in regulating and replacing subsequent processes. It means that the pH, density, and moisture content of the filter residue can reduce the proportion of new reagents and energy input. The conditioning effect coefficient is a unit factor between 0 and 1. For example, if the pH of the refining filter residue is 9, it can replace 50% of the alkali additive, and if the moisture content is 5%, it can reduce the dilution water by 20%. Then the conditioning effect coefficient = (50% + 20%) / 2 = 0.35. The higher the conditioning effect coefficient, the greater the conditioning value.

[0069] Secondly, the economic efficiency of recirculation is calculated based on the residual value of precious metals, the cost of using the recirculation equipment, the cost of recirculation operation, and the conditioning effect coefficient. The recirculation economic efficiency is calculated as follows: (Residual value of precious metals + Conditioning effect coefficient × Baseline conditioning cost) - (Cost of using recirculation equipment + Cost of recirculation operation). The baseline conditioning cost refers to the comprehensive cost of reagents, energy, etc., required to achieve the process conditions without filter residue recirculation. This cost can be obtained from historical data, such as 1000 yuan for a certain batch. For example, if the residual value of precious metals is 6750 yuan, the conditioning effect coefficient is 0.35, the cost of using the recirculation equipment is 500 yuan, the cost of recirculation operation is 180 yuan, and the baseline conditioning cost is 1000 yuan, then the recirculation economic efficiency is (6750 + 0.35 × 1000) - (500 + 180) = 6420. The recirculation economic efficiency is a quantitative indicator that comprehensively measures the benefits and costs of filter residue recirculation. The higher the recirculation economic efficiency, the more significant the overall benefits of filter residue recirculation.

[0070] Secondly, when the economic benefits of recirculation exceed the first recirculation threshold, the recirculation ratio of the refining filter residue is set to 1. The first recirculation threshold is a pre-set critical standard used to determine whether the full recirculation of the filter residue has significant economic value. It is a priori data and can be determined by combining actual factors such as the fixed cost of equipment in the production line, the target profit level, and historical recirculation benefits. It can provide a clear quantitative boundary for recirculation decisions. For example, based on nearly 500 batches of full-volume data, a production line calculated that when the net profit of full-volume recirculation is less than 6,000 yuan, the equipment occupancy cost of full-volume recirculation will exceed the actual profit. Therefore, the first recirculation threshold is set to 6,000 yuan. For example, if the economic benefit of recirculation is 6420 yuan and the first recirculation threshold is 6000 yuan, since 6420 yuan > 6000 yuan, it means that the total benefit of recirculating the refining filter residue exceeds the total cost of recirculation. Therefore, the recirculation ratio of the refining filter residue is 1, that is, the entire amount is recirculated into the next batch of processes to fully recover the residual precious metals and give full play to their conditioning effect, thereby maximizing resource utilization and benefits.

[0071] Furthermore, when the economic efficiency of recirculation is between the first recirculation threshold and the second recirculation threshold, the recirculation ratio is calculated based on the conditioning requirements and the pre-constructed conditioning mapping model. The second recirculation threshold is a pre-set critical value used to define whether the filter residue is suitable for partial recirculation. It belongs to a priori data. The second recirculation threshold is lower than the first recirculation threshold and can be determined based on the balance between the marginal cost of partial recirculation (such as the energy consumption of some equipment startup and the tiered labor cost) and the potential benefits. The second recirculation threshold can provide a precise boundary for recirculation decisions in the middle range. For example, a production line found by analyzing historical data of partial recirculation that when the economic efficiency of recirculation is less than 1,000 yuan, even if only a small amount of filter residue is recirculated, its benefits are difficult to cover the costs. Therefore, the second recirculation threshold is set to 1,000 yuan.

[0072] For example, the conditioning mapping model can be constructed based on the conditioning requirements of the next batch of processes (such as target pH value, moisture content range, density threshold, etc.), the physicochemical properties of the current refining filter residue (such as actual pH value, moisture content, density, etc.), and the optimal reflux ratio. For example, three sets of data can be obtained from historical data: the conditioning requirements of the next batch of processes, the physicochemical properties of the current refining filter residue, and the optimal reflux ratio. The mapping relationship between the three can be established, and then the conditioning mapping model can be trained. By inputting the current conditioning requirements and the physicochemical properties of the refining filter residue, the reflux ratio can be output.

[0073] For example, if the first reflux threshold is 6,000 yuan and the second reflux threshold is 1,000 yuan, when the reflux economics is 2,000 yuan, since the reflux economics are between the first and second reflux thresholds, it means that although the filter residue reflux has not reached the benefit standard of full reflux, it still has some recovery value and conditioning effect. Then, the conditioning requirements of the next batch of processes and the physicochemical properties of the current refining filter residue are input into the conditioning mapping model, and the reflux ratio is mapped out, such as 0.4, that is, only 40% of the refining filter residue is refluxed into the next batch of processes.

[0074] Furthermore, when the economic benefits of recirculation are less than the second recirculation threshold, it indicates that the recirculation revenue is insufficient to cover the costs and the conditioning value is low. Therefore, the recirculation ratio of the refining filter residue is determined to be 0, meaning that all the refining filter residue is discarded to avoid ineffective investment.

[0075] Finally, filter residue reflux parameters are generated based on the reflux ratio. For example, based on the determined reflux ratio, filter residue reflux parameters including the reflux ratio (e.g., 1, 0.4, 0, etc.) and the reflux timing (before the next batch of leaching begins) are generated. The reflux ratio (e.g., 1, 0.4, 0, etc.) ensures differentiated treatment of the refined filter residue, including full reflux (reflux ratio of 1), partial reflux (e.g., reflux ratio of 0.4), or full disposal (reflux ratio of 0). In this way, through dynamic adjustment, the resource utilization rate is improved by reusing high-value refined filter residue, while the ineffective cost is reduced by disposing of low-value refined filter residue. At the same time, the conditioning effect is used to reduce costs, thereby achieving synergistic benefits of resource recovery and cost control.

[0076] In summary, compared to existing technologies, this application predicts the residual composition information of the first batch of refining filter residue based on the refining process parameters, accordingly evaluates the economic viability of recirculating the refining filter residue, and determines the filter residue recirculation parameters based on the economic viability of recirculation. Thus, by quantitatively evaluating the economic viability of recirculating the refining filter residue, reasonable filter residue recirculation parameters are determined, achieving a balance between maximizing resource utilization and controlling costs.

[0077] S60: Combining the refining process parameters and the filter residue reflux parameters, the first refining solid waste is reused.

[0078] The aforementioned steps determined the refining process parameters suitable for the types to be extracted through iterative optimization, and obtained differentiated filter residue reflux parameters based on reflux economic analysis. Based on this, the first batch of refined solid waste can be accurately refined and recycled, and the refined filter residue can be co-processed with the next batch of refined solid waste according to the differentiated filter residue reflux parameters and mixing process parameters.

[0079] To address the aforementioned issues, this application integrates the refining process parameters and the filter residue reflux parameters to implement the reuse of the first refining solid waste.

[0080] Specifically, step S60 in the method includes: Based on the refining process parameters, the first refining solid waste is refined and recycled to output the refining and recycled product and the refining filter residue. According to the filter residue reflux parameters and mixing process parameters, the refined filter residue is mixed with the next batch of refined solid waste. The mixing process parameters include stirring intensity and mixing time.

[0081] In this embodiment, the first refined solid waste is first refined and recovered based on refining process parameters, outputting refined recovered products and refined filter residue. For example, if the refining process parameters for gold are a cyanide concentration of 0.055 mol / L, a leaching temperature of 62°C, a stirring rate of 320 r / min, and a reaction time of 3.5 hours, and the refining process parameters for silver are a nitric acid concentration of 12%, a pH value of 1.2, and a reaction time of 2.8 hours, then for gold, the cyanide concentration is controlled at 0.055 mol / L, the leaching temperature at 62°C, and the stirring rate at 320 r / min, and cyanide leaching is carried out in a reactor for 3.5 hours to dissolve the gold into the liquid phase; for silver, the nitric acid concentration is adjusted to 12%, the pH value to 1.2, and nitric acid leaching is carried out for 2.8 hours to dissolve the silver; then, after solid-liquid separation (such as plate and frame filtration, with a pressure of 0.3 MPa), refined recovered products (such as gold mud or silver powder) and refined filter residue (containing undissolved residual components and impurities) are output.

[0082] Secondly, based on the filter residue reflux parameters and mixing process parameters, the refined filter residue is mixed with the next batch of refined solid waste. The mixing process parameters include stirring intensity and mixing time. Stirring intensity determines the uniformity of material dispersion, and mixing time ensures that the filter residue and new material are in full contact. Together, they ensure that the residual value and conditioning effect of the refluxed filter residue are effectively exerted. For example, based on the filter residue reflux parameters determined in step S50, and combined with the mixing process parameters, the current batch of refined filter residue is mixed with the next batch of refined solid waste according to the reflux ratio. For instance, if the reflux ratio is 1, all the refined filter residue in this batch is fully mixed with the next batch of refined solid waste according to the mixing process parameters with a certain stirring intensity (e.g., 400 r / min) and mixing time (e.g., 20 minutes). If the reflux ratio is 0.4, 40% of the refined filter residue in this batch is fully mixed with the next batch of refined solid waste according to the mixing process parameters with a certain stirring intensity (e.g., 300 r / min) and mixing time (e.g., 15 minutes). If the reflux ratio is 0, the entire batch of refined filter residue is discarded.

[0083] In summary, compared to existing technologies, this application integrates the refining process parameters and the filter residue reflux parameters to perform the reuse of the first refined solid waste. Thus, by relying on the refining process parameters, efficient extraction of precious metals from a single batch is achieved. Furthermore, by using the filter residue reflux parameters and mixing process parameters, the residual resource value and process adjustment value of the refined filter residue are incorporated into the entire process, improving the overall resource utilization rate and treatment economy of the solid waste.

[0084] In summary, the embodiments of this application have at least the following technical effects: Compared to existing technologies, this application first performs component analysis on the first batch of refined solid waste from the first target batch to obtain the component information of the first batch, wherein the first batch of refined solid waste is anode mud. Thus, through standardized pretreatment, sampling, and testing procedures, the obtained component information of the first batch is ensured to be accurate, comprehensive, and representative, providing a reliable data foundation for subsequent recovery rate prediction, process optimization, and other stages.

[0085] Secondly, this application uses a pre-constructed recovery rate prediction model and the current set of process parameters for the target scenario, combined with the composition information of the first batch, to predict the recovery rate of precious metal refining and obtain a set of predicted recovery rates. In this way, the correlation between composition information, process parameters, and recovery rate is quantified through the recovery rate prediction model, avoiding the subjectivity of traditional manual experience estimation and providing reliable key data support for subsequent analysis.

[0086] Furthermore, this application uses the predicted recovery rate set to perform a dual-objective recovery economic screening to determine the set of categories to be extracted. Thus, through initial economic screening, initiating economic accounting, and setting preset economic constraints, the categories to be extracted are selected, avoiding ineffective extraction of low-content or high-content low-yield items, thereby improving the economy and efficiency of resource utilization.

[0087] Furthermore, this application combines the recovery rate prediction model, the current process parameter set, and the set of types to be extracted to perform multi-objective optimization to obtain refining process parameters. Thus, through iterative optimization, the final refining process parameters can find the optimal balance between recovery rate, cost, and benefit, providing precise operational guidance for the efficient recovery of precious metals.

[0088] Furthermore, this application predicts the residual component information of the first batch of refining filter residue based on the refining process parameters, evaluates the economics of recirculating the refining filter residue accordingly, and determines the filter residue recirculation parameters based on the economics of recirculation. Thus, by quantitatively evaluating the economics of recirculating the refining filter residue, reasonable filter residue recirculation parameters are determined, achieving a balance between maximizing resource utilization and controlling costs.

[0089] Finally, this application integrates the refining process parameters and the filter residue reflux parameters to perform the reuse of the first refined solid waste. In this way, efficient extraction of precious metals in a single batch is achieved by relying on the refining process parameters, and the residual resource value and process adjustment value of the refined filter residue are incorporated into the entire process utilization through the filter residue reflux parameters and mixing process parameters, thereby improving the overall resource utilization rate and treatment economy of solid waste.

[0090] Through the above technical solution, this application obtains the first batch composition information through component detection, predicts the precious metal refining recovery rate using a recovery rate prediction model combined with current process parameters, identifies a high-value extraction category set through dual-objective recovery economic screening, determines suitable refining process parameters through multi-objective optimization, predicts the residual composition information of the refining filter residue based on the refining process parameters, evaluates the economics of recirculation to determine differentiated filter residue recirculation parameters, and finally integrates the refining process parameters and filter residue recirculation parameters to implement the reuse of the first batch of refining solid waste. In this way, it improves the single-batch recovery efficiency through precise process parameters and taps residual value through differentiated filter residue recirculation strategies, thereby increasing the resource utilization rate and economic value of refining solid waste and achieving efficient and economical reuse of precious metal refining solid waste.

[0091] Example 2, as Figure 2 As shown, this embodiment of the invention provides a device for reusing solid waste from precious metal refining. The device includes: a memory 210 for storing computer software program 211; and a processor 220 for reading and executing the computer software program 211, thereby realizing a method for reusing solid waste from precious metal refining.

[0092] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0093] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0094] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0095] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0096] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0097] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0098] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for recycling solid waste from precious metal refining, characterized in that, include: The composition of the first batch of refined solid waste in the first target batch was analyzed to obtain the composition information of the first batch, wherein the first batch of refined solid waste is anode mud; Based on the pre-constructed recovery rate prediction model and the current process parameter set of the target scenario, the precious metal refining recovery rate is predicted by combining the composition information of the first batch, and a set of predicted recovery rates is obtained. Based on the predicted recovery rate set, a dual-objective recovery economic screening is performed to determine the set of species to be extracted; By combining the recovery rate prediction model, the current process parameter set, and the set of species to be extracted, multi-objective optimization is performed to obtain refining process parameters. Based on the refining process parameters, the residual component information of the first batch of refining filter residue is predicted, the recirculation economy of the refining filter residue is evaluated accordingly, and the filter residue recirculation parameters are determined based on the recirculation economy. By combining the refining process parameters and the filter residue reflux parameters, the first refining solid waste is reused.

2. The method as described in claim 1, characterized in that, Compositional analysis was performed on the first batch of refined solid waste from the first target batch to obtain the compositional information for the first batch, including: The first refined solid waste is pretreated, and samples are taken based on the pretreatment results to obtain the first sample solid. Based on the first sample solid, the precious metal recycling list of the target scene is traversed to perform content detection, forming a precious metal content composition vector of the first target batch, and the output is the composition information of the first batch.

3. The method for recycling solid waste from precious metal refining as described in claim 2, characterized in that, Based on the predicted recovery rate set, a dual-objective recovery economics screening is performed to determine the set of species to be extracted, including: With the minimum economic content as the primary objective, an initial economic screening was conducted based on the component information of the first batch. Obtain the constant and variable costs of multiple refining process production lines, and combine the predicted recovery rate with the total batch size of the first target batch to perform an initial economic calculation on the economic screening results. If the result of the economic calculation is greater than the preset economic constraint, the corresponding precious metal component will be added to the set of types to be extracted. The minimum economic content, the constant cost, and the variable cost are prior data.

4. A method for recycling solid waste from precious metal refining as described in claim 3, characterized in that, By combining the recovery rate prediction model, the current process parameter set, and the target extraction species set, multi-objective optimization is performed to obtain refining process parameters, including: Construct a process parameter optimization function, wherein the process parameter optimization function includes at least a recovery rate factor, a relative cost factor, and a relative profit factor; Using the set of categories to be extracted as an index, the current set of process parameters is traversed to extract process parameters, and the recovery rate prediction model is traversed to call the model. Combining the process parameter optimization function with the model call results, the extracted process parameters are iteratively optimized in multiple objectives until the function value of the process parameter optimization function converges, and the corresponding multi-objective optimization result is output as the refining process parameters.

5. A method for recycling solid waste from precious metal refining as described in claim 4, characterized in that, Based on the refining process parameters, the residual composition information of the first batch of refining filter residue is predicted, the recirculation economics of the refining filter residue is evaluated accordingly, and the filter residue recirculation parameters are determined based on the recirculation economics, including: Based on the residual component information of the first batch, the residual value of precious metals, the cost of using the reflux equipment, the cost of reflux operation, and the conditioning effect coefficient of the refined filter residue are calculated respectively. The residual value of the precious metal is calculated based on real-time market prices; the cost of using the reflux equipment includes the cost of equipment wear and tear and operating time; the cost of reflux operation includes the cost of labor; and the conditioning effect coefficient includes the conditioning substitution rate of pH, density, and moisture content. The economics of reflux are calculated based on the residual value of the precious metal, the cost of using the reflux equipment, the cost of reflux operation, and the conditioning effect coefficient. When the reflux economy is greater than the first reflux threshold, the reflux ratio of the refining filter residue is determined to be 1; When the economic efficiency of the reflux is between the first reflux threshold and the second reflux threshold, the reflux ratio is calculated based on the conditioning demand and the conditioning mapping model. When the reflux economy is less than the second reflux threshold, the reflux ratio of the refining filter residue is determined to be 0; Based on the reflux ratio, the filter residue reflux parameters are generated.

6. The method for recycling solid waste from precious metal refining as described in claim 1, characterized in that, Combining the refining process parameters and the filter residue reflux parameters, the reuse of the first refining solid waste is carried out, including: Based on the refining process parameters, the first refining solid waste is refined and recycled to output the refining and recycled product and the refining filter residue. According to the filter residue reflux parameters and mixing process parameters, the refined filter residue is mixed with the next batch of refined solid waste. The mixing process parameters include stirring intensity and mixing time.

7. A method for recycling solid waste from precious metal refining as described in claim 1, characterized in that, The recovery rate prediction model includes a property-driven prediction model and a data-driven prediction model. The construction steps of the recovery rate prediction model include: Based on the first batch composition information, a precious metal component is randomly selected as the first prediction object, and the first sample data of the first prediction object is obtained accordingly. The first sample data includes sample refining process parameters, sample batch composition information and sample recovery rate. Based on the first sample data, the prior property model is matched and invoked, and the candidate property prediction model of the first prediction object is constructed accordingly. The prediction performance of the candidate property prediction model is verified and the prediction performance of the candidate property is obtained. If the performance of the candidate physical property prediction meets the target prediction performance requirements, the candidate physical property prediction model is output as the recovery rate prediction model, and the precious metal components are iteratively selected by traversing the first batch of component information.

8. A method for recycling solid waste from precious metal refining as described in claim 7, characterized in that, Also includes: If the performance of the alternative property prediction does not meet the target prediction performance requirement, then based on the data-driven method, using the first sample data as supervision and combined with the first preset model parameter constraints, an alternative data prediction model is constructed and trained. Verify whether the prediction performance of the candidate data prediction model meets the target prediction performance requirements; If the conditions are not met, the first preset model parameter constraints are adjusted, and the alternative data prediction model is iteratively constructed and trained until the prediction performance of the alternative data prediction model meets the target prediction performance requirements, and the output is the recovery rate prediction model.

9. A device for recycling solid waste from precious metal refining, characterized in that, include: Memory, used to store computer software programs; A processor is used to read and execute the computer software program, thereby implementing the method for recycling solid waste from precious metal refining as described in any one of claims 1-8.

Images