A dynamic detection method, system and device for ion content in the copper extraction process

Abstract

The present invention relates to the technical field of copper-containing waste treatment, and particularly to a method, system and equipment for dynamically detecting ion content in the copper extraction process. The method includes: obtaining in real time the time-series data of ion concentration in the reaction system, and extracting the mass transfer kinetics correlation features of the time-series data of ion concentration; constructing an ion content prediction model based on the mass transfer kinetics correlation features, and the ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network; generating an initial extraction parameter set through parameter traversal, and performing process simulation based on the initial extraction parameter set using the ion content prediction model to obtain a dynamic prediction result; verifying the accuracy of the ion content prediction model by combining the dynamic prediction result with real-time detection data, and outputting a dynamic change curve of ion content. Through the present invention, the problems of large dynamic prediction deviation, lag in abnormal response and lack of cross-scale correlation in process optimization in the traditional copper extraction process are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of copper-containing waste treatment, and particularly to a method, system and device for dynamically detecting ion content in the copper extraction process. Background Art

[0002] In the hydrometallurgical process of copper, the extraction efficiency of copper liquid smelting slag directly affects the recovery rate of copper resources and production costs. Traditional ion content detection mainly relies on off-line laboratory analysis (such as ICP spectroscopy) or on-line monitoring by a single sensor. The former has a delay of several hours, resulting in a lag in process adjustment. The latter is prone to drift errors due to environmental corrosion and cannot analyze the mass transfer kinetic mechanism.

[0003] Existing prediction models mostly use static empirical formulas or isolated numerical simulations. Although they can partially reflect the flow field distribution or mass transfer law, they have defects such as cross-scale data fragmentation, low computational efficiency (single CFD simulation > 2 hours), and violation of physical conservation. Especially when the raw material composition fluctuates or the equipment ages, the model error increases significantly (> 8%), making it difficult to support real-time optimization decisions.

[0004] The information disclosed in this background art section is only intended to deepen the understanding of the overall background art of the present disclosure, and should not be regarded as an admission or any form of implication that this information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0005] The present invention provides a method, system and device for dynamically detecting ion content in the copper extraction process, which can effectively solve the problems in the background art.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for dynamically detecting ion content in the copper extraction process, the method comprising:

[0008] Obtaining in real time the time series data of ion concentration in the reaction system, and extracting the mass transfer kinetic correlation features of the time series data of ion concentration;

[0009] Constructing an ion content prediction model based on the mass transfer kinetic correlation features, the ion content prediction model characterizing the mass transfer relationship at the reaction interface through a topological network;

[0010] Generating an initial extraction parameter set through parameter traversal, and performing process simulation based on the initial extraction parameter set using the ion content prediction model to obtain a dynamic prediction result;

[0011] Combining the dynamic prediction result with real-time detection data to verify the accuracy of the ion content prediction model, and outputting the dynamic change curve of ion content.

[0012] Further, the ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network, including:

[0013] The reaction interface is discretized into spatial grid cells, and each of the spatial grid cells serves as a node of the topological network. The node attributes include the phase composition, surface activation energy, and local flow velocity gradient of the mineral particles within the cell;

[0014] Adjacent nodes are connected by edges, and the edge weights are determined by the weighted sum of the particle contact area and the diffusion flux between the two nodes;

[0015] A mass conservation residual constraint is added to the topological network to minimize the deviation between the total amount of copper in the input liquid and the total amount of copper in the output liquid and residue;

[0016] The node attributes of the topological network are compared with the initial extraction parameter set to generate a dynamic prediction result of the ion content.

[0017] Further, adding a mass conservation residual constraint to the topological network includes:

[0018] The flow rate of the input liquid, the flow rate of the output liquid, the copper content of the residue, and the corresponding copper ion concentration are obtained in real time, and the total amount of copper in the input liquid and the total amount of copper in the output liquid and residue are calculated;

[0019] Calculate the percentage deviation between the total amount of copper in the input liquid and the total amount of copper in the output liquid and residue. When the percentage deviation exceeds a fixed threshold, network optimization is triggered;

[0020] If the total amount of copper in the input liquid is higher than the total amount of copper in the output liquid and residue, increase the relevant weight parameter of the diffusion flux and reduce the connection weight between nodes in the high activation energy region;

[0021] If the total amount of copper in the input liquid is lower than the total amount of copper in the output liquid and residue, increase the relevant weight parameter of the particle contact area and strengthen the connection relationship of nodes in the low concentration gradient region;

[0022] The adjusted parameters are transmitted back to the topological network, the node attributes and the edge weight distribution are recalculated, and the comparison of the percentage deviation is repeated until the percentage deviation drops below the fixed threshold or reaches the preset maximum number of iterations.

[0023] Further, extracting the mass transfer kinetics correlation characteristics includes:

[0024] Perform first-order differential processing on the ion concentration time series data to calculate the concentration change rate characteristics;

[0025] Based on the numerical simulation of the flow field, flow field distribution parameters are generated, and the flow field distribution parameters include local flow velocity gradient and turbulent kinetic energy;

[0026] Construct a multivariate correlation matrix, perform spatio-temporal alignment on the concentration change rate characteristics and the flow field distribution parameters, and calculate the cross-correlation coefficient;

[0027] Extract the deep spatio-temporal characteristics of the multivariate correlation matrix based on the cross-correlation coefficient, and output the mass transfer kinetics feature vector with dimension compression.

[0028] Furthermore, generate flow field distribution parameters based on flow field numerical simulation, including:

[0029] Divide the internal flow field of the extraction reactor into unstructured grids, and the grid size is dynamically adjusted according to the stirring rate;

[0030] Set the inlet boundary condition as the measured inlet liquid flow rate and pressure pulsation data, and the outlet boundary condition as the free outflow condition;

[0031] Based on the unstructured grids and the boundary conditions, use a turbulence model combined with a multiphase flow interface tracking method to solve the flow field control equation;

[0032] Perform residual analysis on the equation solution results and the measured flow rate data. When the relative error of the local flow rate exceeds a fixed threshold, trigger grid adaptive refinement and restart the solution process;

[0033] If it is less than the fixed threshold, extract the flow field distribution parameters from the converged flow field data, and perform spatio-temporal alignment on the flow field distribution parameters and the ion concentration data.

[0034] Furthermore, calculate the concentration change rate characteristics, including:

[0035] Segment the ion concentration time series data according to a preset time window, and apply moving average filtering to each segment of data. The width of the filtering window is adjusted according to the real-time noise level;

[0036] Calculate the first-order differential value for the denoised data segment, and the initial width of the differential window is adjusted according to the concentration fluctuation frequency;

[0037] Compare the differential calculation results with the concentration change rate. If the relative error exceeds the fixed threshold, trace back to check the rationality of data segmentation, redefine the time window, and adjust the filtering window width and differential calculation strategy;

[0038] Attach a timestamp and a spatial position label to each differential value. The timestamp is aligned with the time step of the flow field numerical simulation, and the spatial position label corresponds to the preset grid coordinates in the extraction tank.

[0039] Furthermore, obtain the dynamic prediction result, including:

[0040] Input the initial extraction parameter set into the ion content prediction model and initialize the node attributes of the topological network;

[0041] Based on the topological network, synchronously perform the prediction of the transient change and steady-state distribution of the ion content and the identification of the mass transfer anomaly risk;

[0042] Compare the prediction results with the real-time detection data and generate an error signal, backpropagate to update the parameters of the topological network, and dynamically shorten the sliding window of the short-term prediction according to the error signal;

[0043] Generate the dynamic prediction results at each time node based on the sliding window of the short-term prediction along the time axis.

[0044] Furthermore, output the dynamic change curve of the ion content, including:

[0045] Fuse the results of the short-term prediction with the steady-state prediction trend to generate a continuous prediction curve, and superimpose and display it with the real-time detection data;

[0046] Highlight the mass transfer anomaly risk area on the continuous prediction curve, and divide the risk level according to the sudden drop amplitude of the edge weight of the topological network;

[0047] Generate a confidence interval for the predicted value, and the coverage interval probability is positively correlated with the stability of the process parameters;

[0048] Generate the dynamic change curve of the ion content based on the continuous prediction curve and the confidence interval of the predicted value.

[0049] A dynamic detection system for the ion content in the copper extraction process, the system includes:

[0050] A feature extraction module, which obtains the time-series data of the ion concentration in the reaction system in real time and extracts the mass transfer kinetics correlation features of the time-series data of the ion concentration;

[0051] A content prediction module, which constructs an ion content prediction model based on the mass transfer kinetics correlation features, and the ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network;

[0052] A process simulation module, which generates an initial extraction parameter set through parameter traversal, and performs process simulation based on the initial extraction parameter set using the ion content prediction model to obtain dynamic prediction results;

[0053] A result output module, which combines the dynamic prediction results with the real-time detection data to verify the accuracy of the ion content prediction model and outputs the dynamic change curve of the ion content.

[0054] A dynamic detection device for the ion content in the copper extraction process, which is used to implement the dynamic detection method for the ion content in the copper extraction process.

[0055] Through the technical solution of the present invention, the following technical effects can be achieved:

[0056] Improve the copper resource recovery efficiency and reduce the consumption of extractant, achieve dynamic prediction of high-precision ion concentration and credibility evaluation, enhance the timeliness of process anomaly warning and the response speed of parameter self-optimization, reduce the requirements for computing resources and manual intervention, and improve the process stability.

[0057] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically describes the specific embodiments of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0059] Figure 1 It is a schematic flow chart of a method for dynamically detecting ion content in the copper extraction process;

[0060] Figure 2 It is a schematic structural diagram characterizing the mass transfer relationship at the reaction interface;

[0061] Figure 3 It is a schematic flow chart for extracting the correlation characteristics of mass transfer kinetics;

[0062] Figure 4 It is a schematic structural diagram for obtaining the dynamic prediction result. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all of them.

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0065] Embodiment 1;

[0066] Such asFigure 1 As shown, the present application provides a method for dynamically detecting ion content in the copper extraction process, and the method includes:

[0067] S10: Obtain the time-series data of ion concentration in the reaction system in real time, and extract the mass transfer kinetic correlation characteristics of the time-series data of ion concentration;

[0068] S20: Construct an ion content prediction model based on the mass transfer kinetic correlation characteristics, and the ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network;

[0069] S30: Generate an initial extraction parameter set through parameter traversal, and perform process simulation based on the initial extraction parameter set using the ion content prediction model to obtain dynamic prediction results;

[0070] S40: Combine the dynamic prediction results with the real-time detection data to verify the accuracy of the ion content prediction model, and output the dynamic change curve of ion content.

[0071] Specifically, the time-series data of ion concentration in the copper extraction process is obtained in real time through sensors or on-line detection devices. The time-series data of ion concentration includes the time-series data of the change of ion concentration in the reaction system over time. The time-series data of ion concentration is the input of this method, which reflects the dynamic change of copper ions in the reaction. The high-frequency sampling method can be used to ensure the timeliness and accuracy of the data, so as to reflect the changes in the reaction system in real time. The time-series data of ion concentration obtained in real time is processed to extract the mass transfer kinetic correlation features therein. First, the ion concentration data is processed mathematically to obtain the concentration change rate features. At the same time, through the numerical simulation of the flow field and the space-time alignment method, the deep-level features of mass transfer kinetics are extracted according to the concentration change rate features. The method of topological network is used to construct the ion content prediction model. First, the reaction interface is discretized. The adjacent nodes of the topological network are connected by edges, and the weight of the edge is determined by the weighted sum of the particle contact area and the diffusion flux. The structure of the topological network characterizes the mass transfer relationship of the reaction interface, thus providing a model basis for the prediction of ion content. After the ion content prediction model is constructed, the initial extraction parameter set is generated by parameter traversal. The initial extraction parameter set includes multiple parameters that affect the copper ion content, such as reaction temperature, reaction time, stirring speed, etc. The initial extraction parameter set provides a basis for the subsequent process simulation, ensuring that the model can cover all possible extraction conditions. According to the initial extraction parameter set, the process simulation is carried out based on the ion content prediction model. The purpose of the process simulation is to predict the change of ion content in the copper extraction process according to the current parameters and generate the dynamic prediction results. The process simulation takes into account factors such as mass transfer kinetic characteristics and flow field distribution to ensure that the simulation results accurately reflect the actual situation. The dynamic prediction results are compared and verified with the real-time detection data. If the error between the prediction results and the real-time detection data exceeds the set threshold, the prediction model needs to be adjusted to improve its accuracy. The parameters of the topological network are adjusted according to the error size to ensure that the prediction model can accurately predict the change of ion content. By integrating the dynamic prediction results and the real-time detection data, the dynamic change curve of ion content is generated. The dynamic change curve of ion content shows the change of ion content in the copper extraction process and can be used for process optimization and abnormal warning. The generation of the dynamic change curve of ion content takes into account the combination of short-term prediction and steady-state prediction to ensure the continuity and accuracy of the prediction results.

[0072] Through the technical solution of the present invention, the copper resource recovery efficiency is improved and the extractant consumption is reduced, the high-precision dynamic prediction of ion concentration and the credibility evaluation are realized, the timeliness of process abnormal warning and the response speed of parameter self-optimization are enhanced, the requirements for computing resources and manual intervention are reduced, and the process stability is improved.

[0073] Furthermore, as Figure 2 shown, the ion content prediction model characterizes the mass transfer relationship of the reaction interface through a topological network, including:

[0074] The reaction interface is discretized into spatial grid cells, and each spatial grid cell serves as a node of the topological network. The node attributes include the phase composition, surface activation energy, and local flow velocity gradient of the mineral particles within the cell;

[0075] Adjacent nodes are connected by edges, and the edge weights are determined by the weighted sum of the particle contact area and diffusion flux between the two nodes;

[0076] A mass conservation residual constraint is added to the topological network to minimize the deviation between the total amount of input liquid copper and the total amount of output liquid and residual copper;

[0077] Compare the node attributes of the topological network with the initial extraction parameter set to generate a dynamic prediction result of the ion content.

[0078] As a preference of the above embodiments, first, the reaction interface (such as the liquid-solid interface in the extraction tank) is discretized into spatial grid cells. This discretization process divides the reaction interface into multiple small spatial units, and each unit corresponds to a node in the topological network. The attributes of each grid cell include the phase composition of mineral particles (such as mineral species, particle size, etc.), surface activation energy (representing the reaction activity of mineral particles), and local flow velocity gradient (describing the flow characteristics of the fluid in this area, such as turbulence intensity and flow velocity distribution); adjacent grid cells (i.e., spatially adjacent units on the reaction interface) are connected by edges, and the weights of these edges are determined by the weighted sum of the following two factors: one is the particle contact area, which represents the degree of particle contact between two adjacent units and affects the efficiency of mass transfer, and the other is the diffusion flux, which describes the diffusion rate of substances between adjacent grid cells and is usually related to the local flow velocity and the diffusion ability of particles. By weighting the two factors, the weight of the edge can reflect the mass transfer relationship between adjacent units, thus providing an accurate description of mass transfer in the topological network; to ensure the material conservation of the model, it is necessary to add a mass conservation residual constraint to the topological network, that is, the deviation between the total amount of input liquid copper and the total amount of output liquid copper and residual copper should be minimized. The mass conservation residual constraint can be achieved by setting an optimization objective function, which measures the difference between input, output, and residue, and minimizes the deviation by adjusting the attributes of each node in the topological network (such as mineral particle phase, surface activation energy, etc.), ensuring an accurate reflection of material conservation in the model; during the dynamic prediction process, the node attributes in the topological network are compared using the initial extraction parameter set (such as reaction temperature, reaction time, stirring speed, etc.). This comparison process involves matching the values of the initial extraction parameter set with the phase composition, surface activation energy, and flow velocity gradient and other attributes of each node in the topological network. According to the comparison results, the node attributes in the topological network are updated, enabling the network model to more accurately predict the dynamic changes in copper ion content based on the current extraction conditions; according to the above steps, in the topological network of the reaction interface, the relationship between node attributes and edge weights already reflects the mass transfer dynamics of the reaction system. By adjusting and optimizing the node attributes, the model will generate a dynamic prediction result of the change in copper ion content over time. This dynamic prediction result is based on the model's accurate description of the mass transfer process, reaction interface, and flow field conditions and can be used to predict the concentration change trend of copper ions under different extraction conditions.

[0079] Furthermore, adding a mass conservation residual constraint to the topological network includes:

[0080] Obtain the input liquid flow rate, output liquid flow rate, residual copper content, and the corresponding copper ion concentration in real time, and calculate the total amount of input liquid copper and the total amount of output liquid and residual copper;

[0081] Calculate the percentage deviation between the total amount of copper in the input liquid and the total amount of copper in the output liquid and residue. When the percentage deviation exceeds a fixed threshold, trigger network optimization;

[0082] If the total amount of copper in the input liquid is higher than the total amount of copper in the output liquid and residue, increase the relevant weight parameters of the diffusion flux and reduce the connection weights between nodes in the high activation energy region;

[0083] If the total amount of copper in the input liquid is lower than the total amount of copper in the output liquid and residue, increase the relevant weight parameters of the particle contact area and strengthen the connection relationship of nodes in the low concentration gradient region;

[0084] Transmit the adjusted parameters back to the topological network, recalculate the node attributes and edge weight distribution, and repeat the comparison of the percentage deviation until the percentage deviation drops below the fixed threshold or reaches the preset maximum number of iterations.

[0085] As a preference of the above embodiments, first, the input liquid flow rate, output liquid flow rate, the content of residual copper, and the copper ion concentration are obtained in real time through on-line monitoring devices. These data are usually collected by flow meters, copper ion sensors, and residue measurement devices to ensure the timeliness and accuracy of the data. For example, the input liquid flow rate can be measured by a flow meter, the output liquid flow rate is measured by a similar device, the residual copper content is obtained through chemical analysis or a sensor, and the copper ion concentration is monitored in real time by an on-line copper ion sensor; after obtaining the above data in real time, calculate the total amount of copper in the input liquid, which can be obtained by multiplying the input liquid flow rate by the copper ion concentration in the input liquid. Then, calculate the total amount of copper in the output liquid and the residue. The amount of copper in the output liquid can be obtained by multiplying the output liquid flow rate by the copper ion concentration in the output liquid, and the amount of copper in the residue is obtained by a similar method or other detection methods. The sum of the two is the total amount of copper in the output liquid and the residue; according to the total amount of copper in the input liquid and the total amount of copper in the output liquid and the residue, calculate the percentage deviation between them. This deviation reflects the loss or accumulation of copper during the input and output processes. If this percentage deviation exceeds a fixed threshold (e.g., set to 5%), then trigger the network optimization step; if the total amount of copper in the input liquid is higher than the total amount of copper in the output liquid and the residue, this indicates that the transfer efficiency of copper ions is low, and there may be insufficient diffusion flux or insufficient mass transfer. Therefore, increase the relevant weight parameters of the diffusion flux, so that the mass transfer speed between the corresponding nodes in the topological network increases. In addition, reduce the connection weight between the nodes in the high activation energy region because the mass transfer speed in the high activation energy region is slow, and reducing its weight helps to balance the mass transfer process; if the total amount of copper in the input liquid is lower than the total amount of copper in the output liquid and the residue, this indicates that the loss of copper is large, and there may be accumulation of copper in the reaction region or too low reaction efficiency. Therefore, increase the relevant weight parameters of the particle contact area, enhance the contact between particles, and improve the reaction efficiency. In addition, strengthen the connection relationship of the nodes in the low concentration gradient region to promote the diffusion of copper ions in the low concentration region and improve the overall mass transfer effect; after the optimization adjustment, transmit the updated parameters (such as diffusion flux, particle contact area, connection weight, etc.) back to the topological network. These adjustments will affect the node attributes and the weight distribution of the edges in the topological network, thereby making a more accurate correction for the dynamic prediction of ion content; after the updated parameters are transmitted back to the topological network, recalculate the node attributes and the weight distribution of the edges. The updated information such as diffusion flux and particle contact area will affect the mass transfer process in the reaction system, and thus affect the transmission characteristics of the entire network; after completing the network optimization and node attribute update, recalculate the percentage deviation between the total amount of copper in the input liquid and the total amount of copper in the output liquid and the residue. By comparing the current deviation with the set threshold, judge whether the optimization process is accurate enough. This process will be repeatedly executed until the percentage deviation drops below the fixed threshold or reaches the preset maximum number of iterations. If the deviation drops within the acceptable range, the optimization process ends; otherwise, continue with the network optimization.

[0086] Furthermore, as Figure 3 shown, extract the mass transfer kinetics correlation features, including:

[0087] Perform a first-order differential process on the ion concentration time series data to calculate the concentration change rate feature;

[0088] Generate flow field distribution parameters based on the numerical simulation of the flow field. The flow field distribution parameters include local velocity gradient and turbulent kinetic energy;

[0089] Construct a multi-variable correlation matrix, align the concentration change rate feature and the flow field distribution parameters in space-time, and calculate the cross-correlation coefficient;

[0090] Extract the deep space-time features of the multi-variable correlation matrix based on the cross-correlation coefficient, and output the mass transfer kinetics feature vector with compressed dimensions.

[0091] As a preference of the above embodiments, first, collect real-time sequential data of ion concentration to reflect the change of ion concentration over time during the copper extraction process. Perform a first-order differential processing on the sequential data of ion concentration to calculate the rate of change of ion concentration over time (concentration change rate feature). By calculating the ratio of the concentration change amount to the time interval, obtain the rate of concentration change at each moment. This feature can help understand the change speed of ion concentration during the reaction process. Use numerical simulation of the flow field (such as CFD simulation) to generate distribution parameters of the flow field, including local flow velocity gradient and turbulent kinetic energy. These parameters reflect the flow characteristics of the fluid in the reactor. The local flow velocity gradient reflects the spatial change of the fluid velocity, and the turbulent kinetic energy describes the turbulence intensity in the fluid flow, which is an important factor affecting the mass transfer process. Combine the ion concentration change rate feature and the flow field distribution parameters (local flow velocity gradient and turbulent kinetic energy) to construct a multivariate correlation matrix. Each row of the multivariate correlation matrix corresponds to the feature data at a moment, and the columns represent different types of features (such as concentration change rate, flow velocity gradient, turbulent kinetic energy, etc.). In this way, the concentration change feature and the flow field parameters can be aligned one by one according to time and spatial position to form a high-dimensional data matrix, characterizing the mass transfer behavior at different time points and spatial positions. In the multivariate correlation matrix, correspond the ion concentration change rate and the flow field parameters (such as local flow velocity gradient and turbulent kinetic energy) through spatio-temporal alignment, and calculate the cross-correlation coefficient between them. The cross-correlation coefficient is used to measure the relationship between different variables, characterizing the linear relationship between the ion concentration change rate and the flow field parameters. By calculating the cross-correlation coefficient, the influence of the flow field characteristics on the ion concentration change can be analyzed. According to the calculated cross-correlation coefficient, further extract deep spatio-temporal features from the multivariate correlation matrix, and perform feature compression through a dimensionality reduction method (such as PCA, etc.) to reduce the feature dimension. After the dimensionality reduction process, a mass transfer kinetics feature vector containing the main information can be obtained. This feature vector effectively represents the kinetics characteristics in the mass transfer process and can provide optimized input for the ion content prediction model.

[0092] Furthermore, based on the numerical simulation of the flow field, generate the flow field distribution parameters, including:

[0093] Divide the internal flow field of the extraction reactor into unstructured grids, and the grid size is dynamically adjusted according to the stirring rate;

[0094] Set the inlet boundary condition as the measured inlet liquid flow velocity and pressure pulsation data, and the outlet boundary condition as the free outflow condition;

[0095] Based on the unstructured grids and boundary conditions, use the turbulence model combined with the multiphase flow interface tracking method to solve the flow field control equation;

[0096] Perform residual analysis on the equation solution results and the measured flow velocity data. When the relative error of the local flow velocity exceeds a fixed threshold, trigger grid adaptive refinement and restart the solution process;

[0097] If it is less than the fixed threshold, extract the flow field distribution parameters from the converged flow field data, and perform spatio-temporal alignment on the flow field distribution parameters and the ion concentration data.

[0098] As a preference of the above embodiments, first, the flow field inside the extraction reactor is divided into unstructured grids. Unstructured grids have greater flexibility and can better adapt to complex fluid regions, such as the stirring zone inside the reactor, the region near the liquid surface, etc. The size of the grids is dynamically adjusted according to the stirring rate. When the stirring rate is high, the fluid disturbance increases, and the local flow field changes are more complex, so finer grids need to be used to more accurately simulate the behavior of the fluid; while when the stirring rate is low, larger grid sizes can be used to improve the calculation efficiency. In numerical simulation, the setting of boundary conditions is very important. When setting the inlet boundary conditions, the actually measured inlet liquid flow velocity and pressure pulsation data are used. These data reflect the changes in the inlet liquid flow rate and pressure fluctuations, which helps to accurately simulate the inlet conditions of the fluid. The outlet boundary conditions are set as free outflow conditions, which means that the fluid can freely flow out of the outlet region without being affected by additional constraint conditions. This setting is usually applicable to the outlet region of the extraction reactor and can better reflect the natural flow of the fluid. Under the set unstructured grids and boundary conditions, a turbulence model (such as the RANS model, etc.) is used to simulate the turbulence flow characteristics. In addition, due to the existence of multiphase flow in the reactor, a multiphase flow interface tracking method is used to handle the interaction between liquid-solid or liquid-liquid. By solving the flow field control equations (such as the Navier-Stokes equations), information such as the velocity distribution, pressure field, and turbulence intensity of the entire flow field is obtained. These information provide the necessary basis for the subsequent simulation of the mass transfer process. After obtaining the numerical solution of the flow field, residual analysis is carried out, and the numerical calculation results are compared with the actually measured flow velocity data. By calculating the relative error of the local flow velocity, the accuracy of the simulation results is judged. When the relative error of the local flow velocity exceeds a fixed threshold, it indicates that the current flow field simulation results are not accurate enough. Therefore, the grids need to be adaptively refined. Adaptive refinement means using finer grids in regions with large flow field changes (such as high turbulence intensity regions) to improve the calculation accuracy. If the relative error of the local flow velocity is less than the set fixed threshold, it indicates that the flow field simulation has converged, and the obtained flow field data is accurate enough. From the converged flow field data, flow field distribution parameters are extracted, mainly including local flow velocity gradient, turbulent kinetic energy, etc. These parameters reflect the movement of the fluid in the reactor and directly affect the ion mass transfer process. The extracted flow field distribution parameters are spatially and temporally aligned with the ion concentration data, that is, ensuring that the time and space positions of the flow field distribution parameters match those of the ion concentration, so as to perform effective correlation analysis in the subsequent model.

[0099] Furthermore, calculate the concentration change rate characteristics, including:

[0100] Segment the ion concentration time series data according to a preset time window, and apply moving average filtering processing to each segment of data. The width of the filtering window is adjusted according to the real-time noise level;

[0101] Calculate the first-order differential value for the denoised data segment, and the initial width of the differential window is adjusted according to the concentration fluctuation frequency;

[0102] Compare the differential calculation result with the concentration change rate. If the relative error exceeds a fixed threshold, trace back to check the rationality of data segmentation, redefine the time window and adjust the filtering window width and differential calculation strategy;

[0103] Attach a timestamp and a spatial position label to each differential value. The timestamp is aligned with the time step of the flow field numerical simulation, and the spatial position label corresponds to the preset grid coordinates in the extraction tank.

[0104] As an optimization of the above embodiments, the ion concentration time series data is segmented according to a preset time window. The ion concentration data within each time period will be processed as an independent data block. A moving average filter is applied to each segment of data. The function of the moving average filter is to smooth the data, remove random noise in a short period of time, and retain the long-term trend of the data. The width of the filter window will be dynamically adjusted according to the noise level monitored in real time. If the noise is large, the window width is increased to better smooth the data; if the noise is small, the window width can be appropriately reduced to avoid over-smoothing. For each data segment after filtering, the rate of change of the concentration is calculated. Specifically, the first-order differential value is calculated. This operation can reveal the speed of change of the ion concentration and reflect the dynamic behavior in the mass transfer process. The initial width of the differential window will be adjusted according to the concentration fluctuation frequency. If the concentration changes rapidly, a shorter window width is adopted; if the concentration changes slowly, a larger window width can be selected to capture the change trend over a longer period of time. The rate of change of the concentration obtained by the differential calculation (differential result) is compared with the rate of change of the concentration. The rate of change of the concentration is obtained by the first-order differential calculation, and here it is used to check whether the differential calculation accurately reflects the change of the concentration. If the relative error between the differential result and the actual rate of change of the concentration exceeds a fixed threshold (for example, set to 5%), the rationality of the data segmentation is traced back to ensure that there is no problem with the delimitation of the time window. If it is found that the segmentation is unreasonable, the time window needs to be re-divided, and the width of the filter window and the differential calculation strategy need to be adjusted. If the data segmentation is unreasonable or there is a large error in the differential calculation, the time window and the filtering strategy need to be readjusted. The length of the time window can be dynamically adjusted according to the change law of the concentration fluctuation, and the width of the moving average filter window can be adjusted to better adapt to the new data characteristics. The differential calculation strategy also needs to be adjusted according to the specific characteristics of the data, such as increasing or decreasing the width of the calculation window to ensure that the obtained differential result is more accurate. To ensure the alignment of the data with the results of the flow field numerical simulation, each differential calculation result (rate of change of the concentration) needs to be appended with a timestamp and a spatial position label. The timestamp is aligned with the time step of the flow field numerical simulation to ensure the time synchronization of the ion concentration data sequence and the flow field simulation. The spatial position label corresponds to the preset grid coordinates in the extraction tank to ensure that the position of each data point can be accurately mapped to the physical position in the extraction reactor. This helps to correctly correlate the flow field characteristics with the ion concentration changes in subsequent analysis for spatio-temporal alignment.

[0105] Furthermore, as Figure 4 shown, obtaining the dynamic prediction result includes:

[0106] Inputting the initial extraction parameter set into the ion content prediction model and initializing the node attributes of the topological network;

[0107] Based on the topological network, synchronously execute the prediction of transient changes and steady-state distributions of ion content and the identification of mass transfer anomaly risks;

[0108] Compare the prediction results with the real-time detection data and generate an error signal, backpropagate to update the parameters of the topological network, and dynamically shorten the sliding window of short-term prediction according to the error signal;

[0109] Generate dynamic prediction results for each time node along the time axis based on the sliding window of short-term prediction.

[0110] As a preference of the above embodiments, first, an initial extraction parameter set (such as reaction temperature, stirring speed, reaction time, etc.) is input into the ion content prediction model. The initial extraction parameter set represents the preliminary settings of various process conditions during the extraction process. Subsequently, the node attributes of the topological network are initialized according to these parameters. These attributes may include the phase composition of mineral particles, surface activation energy, local flow velocity gradient, etc., which are set according to the actual conditions of the extraction process. When initializing the topological network, ensure that the attributes of each node are consistent with the current process conditions to provide a basis for subsequent prediction and optimization. Based on the initialized topological network, perform the prediction of the transient change of ion content, that is, according to the node attributes in the topological network and the current extraction conditions, simulate the process of ion content changing with time. This process focuses on predicting the short-term change trend of ion concentration; perform the steady-state distribution prediction, simulate the distribution of ion concentration after long-term stability during the extraction process, and predict the copper ion concentration distribution in the equilibrium state; identify the risk of mass transfer anomalies, identify possible mass transfer anomaly risk areas through the topological network, such as areas with too low local flow velocity or too poor mass transfer efficiency, and timely detect potential problems in the mass transfer process to provide comprehensive data support for process optimization and fault prediction. Compare and analyze the prediction results of the transient change of ion content, steady-state distribution, and mass transfer risk obtained through the topological network with the real-time detection data. The real-time detection data can monitor the concentration of copper ions in real time through an online sensor. During the comparison process, calculate the error signal according to the difference between the prediction result and the actual monitoring data. The error signal reflects the deviation between the model prediction and the actual measurement. According to the generated error signal, use the backpropagation algorithm (similar to the backpropagation mechanism in a neural network) to update the parameters in the topological network. This process will adjust the attributes of the topological network nodes, such as the surface activation energy of mineral particles, flow velocity gradient, etc., to reduce the prediction error. At the same time, dynamically adjust the sliding window of the short-term prediction according to the size of the error signal. When the prediction error is large, shorten the time length of the sliding window, so that the ion content prediction model can adjust the parameters more sensitively and improve the accuracy of short-term prediction. Use the adjusted sliding window and the updated topological network parameters to generate dynamic prediction results at each time node on the time axis. These results describe the change trend of copper ion concentration at each time point. Due to the dynamic adjustment of the sliding window, the ion content prediction model can flexibly adjust the prediction time span according to different error signals, making the prediction more accurate and real-time. The prediction results at each time node will be gradually output to form a continuous dynamic prediction curve of ion content during the copper extraction process.

[0111] Furthermore, output the dynamic change curve of ion content, including:

[0112] Fuse the results of short-term prediction and the steady-state prediction trend to generate a continuous prediction curve, and superimpose and display it with the real-time detection data;

[0113] Highlight the mass transfer anomaly risk area on the continuous prediction curve, and divide the risk level according to the sudden drop amplitude of the edge weight of the topological network;

[0114] Generate a confidence interval for the predicted value, and the coverage probability of the interval is positively correlated with the stability of the process parameters;

[0115] Generate a dynamic change curve of ion content based on the continuous prediction curve and the confidence interval of the predicted value.

[0116] As a preference of the above embodiment, fuse the short-term prediction results after sliding window adjustment with the steady-state prediction trend. The short-term prediction results reflect the short-term ion concentration changes during the copper extraction process, while the steady-state prediction trend predicts the long-term trend after the system reaches equilibrium. By combining these two parts of prediction results, a continuous prediction curve is formed. This curve shows the changing trend of ion concentration over time, including short-term fluctuations and long-term equilibrium. Then, superimpose the continuous prediction curve with the real-time detection data. The real-time detection data is obtained in real-time through an on-line monitoring device and is compared and displayed with the prediction curve for facilitating an intuitive evaluation of the difference between the prediction result and the actual measurement. On the continuous prediction curve, use highlighter marks to highlight the possible mass transfer anomaly risk areas. These risk areas may be regions where the mass transfer efficiency is low or abnormal during the reaction process, which may lead to unexpected changes in the copper ion concentration. Divide the risk level according to the sudden drop amplitude of the edge weight in the topological network. When the weight of certain edges drops significantly, it indicates that the mass transfer efficiency in this area may be affected, so it needs to be identified and marked as a risk area. The risk level can be divided into multiple levels, such as low risk, medium risk, and high risk, representing different degrees of mass transfer anomalies. Based on the continuous prediction curve, generate a confidence interval for the predicted value. The confidence interval represents the uncertainty range of the prediction result, usually given at a certain confidence level (such as 95% confidence interval). The width of the confidence interval is positively correlated with the stability of the process parameters. When the process parameters (such as stirring speed, temperature, etc.) are relatively stable, the confidence interval will be relatively narrow; while when the process parameters fluctuate greatly, the confidence interval will become wider, thus indicating an increase in the uncertainty of the prediction. This confidence interval can help evaluate the credibility of the prediction result as a quantitative measure of the prediction accuracy. Finally, based on the fused continuous prediction curve and the corresponding confidence interval of the predicted value, generate the final dynamic change curve of ion content. The dynamic change curve of ion content not only shows the changing trend of copper ion concentration over time, but also reflects the credibility and uncertainty of the prediction through the confidence interval. The generation of the dynamic change curve of ion content takes into account short-term fluctuations, long-term trends, and possible risk areas, and can intuitively display the process stability and prediction reliability, providing real-time monitoring and optimization guidance for the production process.

[0117] Embodiment 2;

[0118] Based on the same inventive concept as the method for dynamically detecting ion content in a copper extraction process in the foregoing embodiments, the present invention also provides a system for dynamically detecting ion content in a copper extraction process, the system comprising:

[0119] A feature extraction module that obtains the ion concentration time-series data of the reaction system in real time and extracts the mass transfer kinetics correlation features of the ion concentration time-series data;

[0120] A content prediction module that constructs an ion content prediction model based on the mass transfer kinetics correlation features, and the ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network;

[0121] A process simulation module that generates an initial extraction parameter set through parameter traversal, and performs process simulation based on the initial extraction parameter set using the ion content prediction model to obtain a dynamic prediction result;

[0122] A result output module that combines the dynamic prediction result with the real-time detection data to verify the accuracy of the ion content prediction model and outputs the dynamic change curve of the ion content.

[0123] The above adjustment system in the present invention can effectively implement a method for dynamically detecting ion content in a copper extraction process, and the technical effects that can be achieved are as described in the foregoing embodiments, which will not be elaborated here.

[0124] Embodiment III;

[0125] Based on the same inventive concept as the method for dynamically detecting ion content in a copper extraction process in the foregoing embodiments, the present invention also provides a device for dynamically detecting ion content in a copper extraction process, which is used to implement the method for dynamically detecting ion content in a copper extraction process.

[0126] The above device in the present invention can effectively implement the method for dynamically detecting ion content in a copper extraction process, and the technical effects that can be achieved are as described in the foregoing embodiments, which will not be elaborated here.

[0127] Although the present application has been described in combination with specific features and their embodiments, it is obvious that various modifications and combinations can be made without departing from the spirit and scope of the present application. Accordingly, the present specification and the drawings are merely exemplary descriptions of the present application as defined by the appended claims, and are considered to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A method for dynamically detecting ion content in the copper extraction process, characterized in that, The method includes: Obtaining real-time ion concentration time-series data of the reaction system and extracting the mass transfer kinetics correlation features of the ion concentration time-series data; Constructing an ion content prediction model based on the mass transfer kinetics correlation features, where the ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network; Generating an initial extraction parameter set through parameter traversal, and performing process simulation based on the initial extraction parameter set using the ion content prediction model to obtain a dynamic prediction result; Combining the dynamic prediction result with real-time detection data to verify the accuracy of the ion content prediction model and outputting an ion content dynamic change curve; The ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network, including: Discretizing the reaction interface into spatial grid cells, with each spatial grid cell serving as a node of the topological network, and the node attributes including the phase composition, surface activation energy, and local flow velocity gradient of the mineral particles within the cell; Adjacent nodes are connected by edges, and the edge weights are determined by the weighted sum of the particle contact area and diffusion flux between the two nodes; Adding a mass conservation residual constraint to the topological network to minimize the deviation between the total amount of copper in the input liquid and the total amount of copper in the output liquid and residue; Comparing the node attributes of the topological network with the initial extraction parameter set to generate a dynamic prediction result of the ion content.

2. The method for dynamically detecting the ion content in the copper extraction process according to claim 1, wherein, Adding a mass conservation residual constraint to the topological network includes: Obtaining the input liquid flow rate, output liquid flow rate, residue copper content, and corresponding copper ion concentration in real time, and calculating the total amount of copper in the input liquid and the total amount of copper in the output liquid and residue; Calculating the percentage deviation between the total amount of copper in the input liquid and the total amount of copper in the output liquid and residue, and triggering network optimization when the percentage deviation exceeds a fixed threshold; If the total amount of copper in the input liquid is higher than the total amount of copper in the output liquid and residue, enhancing the relevant weight parameters of the diffusion flux and reducing the connection weight between nodes in the high activation energy region; If the total amount of copper in the input liquid is lower than the total amount of copper in the output liquid and residue, increasing the relevant weight parameters of the particle contact area and strengthening the connection relationship between nodes in the low concentration gradient region; Transmitting the adjusted parameters back to the topological network, recalculating the node attributes and edge weight distribution, and repeating the comparison of the percentage deviation until the percentage deviation drops below the fixed threshold or reaches the preset maximum number of iterations.

3. The dynamic detection method for ionic content in the copper extraction process according to claim 1, wherein Extracting the mass transfer kinetics correlation features includes: Performing a first-order differential process on the ion concentration time-series data to calculate the concentration change rate feature; Generating flow field distribution parameters based on fluid flow numerical simulation, where the flow field distribution parameters include local flow velocity gradient and turbulent kinetic energy; Constructing a multi-variable correlation matrix, aligning the concentration change rate feature and the flow field distribution parameters in space and time, and calculating the cross-correlation coefficient; Extracting the deep space-time features of the multi-variable correlation matrix based on the cross-correlation coefficient and outputting a dimension-compressed mass transfer kinetics feature vector.

4. The method for dynamically detecting the ion content in the copper extraction process according to claim 3, characterized in that, Generating flow field distribution parameters based on fluid flow numerical simulation includes: Dividing the internal flow field of the extraction reactor into unstructured grids, and dynamically adjusting the grid size according to the stirring rate; Set the inlet boundary conditions as the measured inlet liquid flow rate and pressure pulsation data, and the outlet boundary conditions as the free outflow condition; Based on the unstructured grid and the boundary conditions, use a turbulence model combined with a multiphase flow interface tracking method to solve the flow field control equations; Perform residual analysis on the equation solution results and the measured flow rate data. When the relative error of the local flow rate exceeds a fixed threshold, trigger grid adaptive refinement and restart the solution process; If it is less than the fixed threshold, extract the flow field distribution parameters from the converged flow field data, and align the flow field distribution parameters with the ion concentration data in space and time.

5. The method for dynamically detecting ion content in the copper extraction process according to claim 3, characterized in that Calculate the characteristic of the concentration change rate, including: Segment the ion concentration time series data according to a preset time window, and apply a moving average filter to each segment of data. The width of the filter window is adjusted according to the real-time noise level; Calculate the first-order differential value for the denoised data segment. The initial width of the differential window is adjusted according to the concentration fluctuation frequency; Compare the differential calculation results with the concentration change rate. If the relative error exceeds the fixed threshold, trace back to check the rationality of the data segmentation, redefine the time window, and adjust the filter window width and differential calculation strategy; Attach a timestamp and a spatial position label to each differential value. The timestamp is aligned with the time step of the flow field numerical simulation, and the spatial position label corresponds to the preset grid coordinates in the extraction tank.

6. The dynamic detection method for ion content in the copper extraction process according to claim 1, characterized in that Obtain the dynamic prediction results, including: Input the initial extraction parameter set into the ion content prediction model and initialize the node attributes of the topological network; Based on the topological network, synchronously perform the prediction of the transient change, steady-state distribution of the ion content, and the identification of mass transfer anomaly risks; Compare the prediction results with the real-time detection data and generate an error signal, backpropagate to update the parameters of the topological network, and dynamically shorten the sliding window of the short-term prediction according to the error signal; Generate the dynamic prediction results at each time node based on the sliding window of the short-term prediction along the time axis.

7. The method for dynamically detecting the ion content in the copper extraction process according to claim 6, characterized in that, Output the dynamic change curve of the ion content, including: Fuse the results of the short-term prediction with the steady-state prediction trend to generate a continuous prediction curve, and superimpose it with the real-time detection data for display; Highlight the mass transfer anomaly risk area on the continuous prediction curve, and divide the risk level according to the sudden drop amplitude of the edge weights of the topological network; Generate a confidence interval for the predicted value, and the coverage interval probability is positively correlated with the stability of the process parameters; Generate the dynamic change curve of the ion content based on the continuous prediction curve and the confidence interval of the predicted value.

8. A dynamic detection system for ion content in the copper extraction process, characterized in that, Adopt the method for dynamically detecting the ion content in the copper extraction process as described in Claim 1. The system includes: A feature extraction module that obtains the ion concentration time series data of the reaction system in real time and extracts the mass transfer kinetics correlation features of the ion concentration time series data; A content prediction module that constructs an ion content prediction model based on the mass transfer kinetics correlation features. The ion content prediction model characterizes the mass transfer relationship at the reaction interface through a topological network; A process simulation module that generates an initial extraction parameter set through parameter traversal, and performs process simulation based on the initial extraction parameter set using the ion content prediction model to obtain dynamic prediction results; The result output module combines the dynamic prediction results with the real-time detection data to verify the accuracy of the ion content prediction model and outputs the dynamic change curve of the ion content.

9. An ion content dynamic detection device for a copper extraction process is used to implement the ion content dynamic detection method for a copper extraction process as described in claims 1-7.

Images