A gold mine beneficiation and smelting process optimization method, device and medium based on digital twinning

Abstract

This invention discloses a method, equipment, and medium for optimizing gold ore beneficiation and smelting processes based on digital twins, relating to the field of digital twin technology in gold ore beneficiation and smelting. The method includes: collecting process operation data, control variable data, and actuator status data for each process in a gold ore beneficiation and smelting production line; preprocessing and associating these data with the process topology to generate process-level operation data streams; analyzing the response relationship between changes in control variables and process operation results based on the operating status of the process-level digital twin, identifying key coupled control variables, and constructing a cross-process coupling relationship graph; verifying cross-process linkage consistency and resolving control conflicts using a set of candidate collaborative control strategies, and determining the combination of cross-process linkage control actions. This invention achieves full-process digital synchronization and high-fidelity simulation by generating a process-level digital twin, improving process visualization and decision support capabilities.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology for gold ore beneficiation and metallurgy, and in particular to a method, equipment and medium for optimizing gold ore beneficiation and metallurgy processes based on digital twins. Background Technology

[0002] In the field of gold ore beneficiation and metallurgical process optimization technology, traditional methods usually focus on independent modeling and local parameter optimization of individual processes from crushing, grinding, flotation to leaching. In recent years, with the development of intelligent manufacturing technology, digital twin technology has been introduced, aiming to achieve process monitoring and simulation analysis by constructing a virtual mapping of physical entities. Conventional solutions usually collect process operation data of each process by deploying sensor networks, and establish steady-state or dynamic models of each process based on historical data, and then implement rule-based predictive control adjustments to key control variables (such as grinding fineness and reagent addition).

[0003] However, existing optimization methods based on digital twins still have room for improvement in the collaborative optimization of the entire process; existing methods are insufficient in mining the topological relationships of materials, energy and information flows between processes, making it difficult to identify key coupling variables across processes and their interaction paths, thus limiting the global optimization capability; existing methods lack a mechanism for efficient simulation verification and iterative optimization of multivariable linkage control strategies in virtual space. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a gold ore beneficiation and smelting process optimization method based on digital twins to solve the problems of insufficient cross-process collaborative optimization and difficulty in efficiently generating control strategies.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for optimizing a gold ore beneficiation and smelting process based on digital twins. This method includes: collecting process operation data, control variable data, and actuator status data for each process in a gold ore beneficiation and smelting production line; preprocessing and associating these data with the process topology to generate a process-level operation data stream; based on the process-level operation data stream, constructing a virtual topology network of process objects corresponding to the control objects of each process in the gold ore beneficiation and smelting production line, and establishing a mapping relationship between the operating status of each process control object and real-time data to generate a process-level digital twin operating entity; based on the operating status of the process-level digital twin operating entity, analyzing the response relationship between changes in control variables of each process and the process operation results, identifying key coupled control variables, and constructing a cross-process coupling relationship graph; performing linkage relationship analysis on the cross-process coupling relationship graph to determine a set of cross-process linkage control actions, and combining and evolving the set of cross-process linkage control actions in the process-level digital twin operating entity to obtain a set of candidate collaborative control strategies; and performing cross-process linkage consistency verification and control conflict resolution on the candidate collaborative control strategy set to determine the combination of cross-process linkage control actions.

[0007] As a preferred embodiment of the gold ore beneficiation and smelting process optimization method based on digital twins described in this invention, the steps for generating process-level operational data streams are as follows: The preprocessing includes time alignment, outlier removal, and missing data completion. Based on the topological sequence and material transfer relationships between each process, the pre-processed process operation data, control variable data, and actuator status data are correlated and integrated to generate process-level operation data streams.

[0008] As a preferred embodiment of the gold ore beneficiation and smelting process optimization method based on digital twins described in this invention, the steps for constructing a virtual topology network of process objects corresponding to the control objects of each process in the gold ore beneficiation and smelting production line are as follows: Based on the process-level operation data flow, digital twin abstractions are performed on the physical control objects of each process to establish a set of twin object descriptions; Based on the topological order and material transfer relationship between each process, the corresponding twin object is instantiated from the twin object description set and associated with the process-level operation data flow to generate a set of twin object instances; The collection of twin object instances is logically linked and configured according to the gold ore beneficiation and smelting process to construct a virtual topology network of process objects.

[0009] As a preferred embodiment of the gold ore beneficiation and smelting process optimization method based on digital twins described in this invention, the steps for establishing the mapping relationship between the operating status and real-time data of each process control object to generate a process-level digital twin operating entity are as follows. By continuously receiving process-level operation data streams, the state updates and behavior evolution of each twin object instance in the virtual topology network of process objects are driven, and the operation state of each process control object is constrained to a uniformity based on the evolution results, thus forming a dynamic simulation operation environment. The dynamic simulation runtime environment is integrated and encapsulated with the virtual topology network of the process object to generate a process-level digital twin runtime.

[0010] As a preferred embodiment of the gold ore beneficiation and smelting process optimization method based on digital twins described in this invention, the steps for analyzing the response relationship between changes in control variables of each process and the process operation results, and identifying key coupled control variables, are as follows: In the process-level digital twin runtime, adjust the control variable settings of each process, synchronously simulate and record the resulting changes in process execution results, and generate a variable-result response dataset. Based on the variable-result response dataset, we analyze the range of processes affected by the adjustment of each control variable and determine the list of processes affected by each control variable. From the list of processes affected by each control variable, control variables whose influence spans across a single process are selected as key coupled control variables.

[0011] As a preferred embodiment of the gold ore beneficiation and smelting process optimization method based on digital twins described in this invention, the steps for constructing the cross-process coupling relationship graph are as follows: Analyze the degree and direction of mutual influence among the changes of key coupled control variables, and generate a variable coupling relationship matrix; Using key coupling control variables as nodes and coupling relationships quantified by the variable coupling relationship matrix as edges, a cross-process coupling relationship graph is constructed.

[0012] As a preferred embodiment of the gold ore beneficiation and smelting process optimization method based on digital twins described in this invention, the step of combining and evolving the cross-process linkage control action set in the process-level digital twin runtime to obtain a candidate collaborative control strategy set includes the following steps: Topological analysis and pattern recognition are performed on the cross-process coupling relationship map to extract the strong coupling paths and linkage relationships between key coupling control variables, forming a set of linkage relationship patterns. The set of linkage relationship patterns is regularized, and the linkage timing logic and collaborative constraints of key coupled control variables are extracted to generate a set of cross-process linkage control rules. Based on the cross-process linkage control rule set, the control values ​​and adjustment methods of different key coupled control variables are combined to generate a cross-process linkage control action set; Based on the cross-process linkage control action set, the action set is combined, serialized and parameter adjusted in the process-level digital twin operation body to generate a candidate collaborative control strategy evolution set; In the process-level digital twin runtime, the evolution set of candidate collaborative control strategies is simulated and run, and the performance is evaluated and screened based on the simulation results to obtain the candidate collaborative control strategy set.

[0013] As a preferred embodiment of the gold ore beneficiation and smelting process optimization method based on digital twins described in this invention, the steps for determining the cross-process linkage control action combination are as follows: Extract the cross-process linkage control actions and control variable adjustment sequence of each collaborative control strategy from the candidate collaborative control strategy set to form a linkage action description; Based on the description of the linkage action, the adjustment direction, timing and range of the cross-process linkage control action are checked for consistency, control conflicts are identified and resolved, and a consistent linkage control action set is generated. Perform executability verification on the set of consistent linkage control actions and determine the combination of cross-process linkage control actions.

[0014] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the gold ore beneficiation and smelting process optimization method based on digital twins as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the gold ore beneficiation and smelting process optimization method based on digital twins as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: by generating a process-level digital twin operating entity, the digital synchronization and high-fidelity simulation of the entire process are realized, providing a virtual experimental environment for dynamic monitoring and predictive analysis, and improving process visualization and decision support capabilities; by acquiring a set of candidate collaborative control strategies, the automated generation and iterative optimization of multivariable control strategies are realized, and through efficient simulation verification in virtual space, the trial and error costs in actual production are reduced, and optimization efficiency and reliability are enhanced. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a gold ore beneficiation and smelting process optimization method based on digital twins.

[0019] Figure 2 A flowchart for constructing a virtual topology network for process objects.

[0020] Figure 3 A flowchart for identifying key coupling control variables and constructing a cross-process coupling relationship map.

[0021] Figure 4 A flowchart for obtaining a set of candidate collaborative control strategies.

[0022] Figure 5 A comparative graph showing the change in metal recovery rate over time for different control strategies.

[0023] Figure 6 This is a spatiotemporal evolution distribution map of concentrate grade under multiple control strategies. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0027] Reference Figures 1-6 This is one embodiment of the present invention, which provides a method for optimizing gold ore beneficiation and smelting processes based on digital twins, including the following steps: S1. Collect process operation data, control variable data and actuator status data of each process in the gold ore beneficiation and smelting production line, and preprocess and associate them with the process topology to generate process-level operation data stream.

[0028] S1.1: Process operation data includes feed characteristic data, slurry state data, and beneficiation result data.

[0029] It should be noted that the feed characteristic data describes the inherent properties of the raw ore entering the beneficiation process. This data is obtained through regular sampling and laboratory testing of the raw ore, including ore grade, ore hardness, mineral composition, and particle size distribution. The slurry state data reflects the physical and chemical properties of the slurry fluid in each process. This data is obtained through real-time monitoring using sensors (such as concentration meters, pH meters, and thermometers) installed in pipelines and tanks, including slurry concentration, pH value, temperature, and redox potential. The beneficiation result data are quality indicator data that evaluate the separation and smelting effects of each process. This data is obtained through regular sampling and laboratory testing of intermediate and final products (such as concentrate and tailings) produced by the process, including concentrate grade, tailings grade, metal recovery rate, and content of harmful elements.

[0030] S1.2: Control variable data include grinding and classification adjustment parameters, flotation adjustment parameters, and smelting adjustment parameters.

[0031] It should be noted that grinding and classification adjustment parameters are process settings used to regulate grinding particle size and classification efficiency, including ball mill feed rate, classifier overflow concentration, and hydrocyclone feed pressure; flotation adjustment parameters are process settings used to regulate flotation froth layer state and separation efficiency, including collector and frother addition rates, flotation machine aeration rate, and slurry level; smelting adjustment parameters are process settings used to regulate the efficiency of leaching and displacement chemical reaction processes, including sodium cyanide concentration, alkalinity (lime addition amount), and reaction temperature in the leaching tank.

[0032] S1.3: The status data of the actuator includes the start / stop status of the actuator, the execution output level, and the execution position status.

[0033] It should be noted that the start / stop status of the actuator is a basic status signal reflecting whether the motor and pump power equipment are energized and running, and the status includes running, stopped, and fault; the execution output level is a feedback signal reflecting the current actual output of the regulating valve and the variable frequency drive continuous regulating mechanism, including the valve opening percentage, the frequency converter output frequency, and the real-time speed of the motor; the execution position status is a signal reflecting whether the opening and closing mechanism of the gate valve and the baffle has reached the limit of the mechanical position, including fully open, fully closed, and not fully closed.

[0034] S1.4: Preprocessing includes time alignment, outlier removal, and missing data completion.

[0035] It should be noted that time alignment refers to the forced synchronization of process operation data and control variable data collected from different processes using a unified clock source to ensure millisecond-level time consistency across all data records; outlier removal refers to the method of comparing sensor readings (such as concentration and pH value) and actuator status data (such as valve opening) in real-time process operation data with reasonable operating ranges based on a process knowledge base, directly marking and removing values ​​that exceed the reasonable operating range; missing data completion refers to the method of using linear interpolation based on adjacent valid data points to generate continuous values ​​to fill in the gaps caused by temporary data loss during transmission and acquisition delays.

[0036] It should be noted that the process knowledge base is based on process flow documents, equipment parameters and operating procedures to form the basic constraint range of control variables for each process. It also combines historical production data and test data to extract typical value ranges and normal fluctuation ranges under stable operating conditions. At the same time, it integrates the physicochemical mechanisms of the beneficiation and metallurgy process (such as mass balance and reaction kinetics) to form structured rule entries and unify variable names and units. Finally, after on-site verification, it is formed, including a dictionary of process control variables and monitoring points, reasonable operating ranges and normal fluctuation ranges of control variables for each process, operating condition labeling rules, anomaly judgment rules and inter-process constraint relationships.

[0037] S1.5: Based on the topological sequence and material transfer relationship between each process, the pre-processed process operation data, control variable data and actuator status data are correlated and integrated to generate process-level operation data flow.

[0038] Furthermore, based on the process topology and material transfer sequence in the gold ore beneficiation and smelting production line from crushing, grinding, flotation to leaching, the pre-processed process operation data, control variable data, and actuator status data are matched and spliced ​​according to the process nodes through which the material flows. Taking the process as the basic node, the matched process operation data, control variable data, and actuator status data are vertically integrated, and the data of each time slice is labeled with the corresponding process, forming a process-level operation data flow that is organized according to the process topology, is time-continuous, and intrinsically related.

[0039] S2. Based on the process-level operation data flow, construct a virtual topology network of process objects corresponding to the control objects of each process in the gold ore beneficiation and smelting production line, and establish a mapping relationship between the operation status and real-time data of each process control object to generate a process-level digital twin operation body.

[0040] S2.1: Based on the process-level operation data flow, perform digital twin abstraction of the physical control objects of each process and establish a set of twin object descriptions.

[0041] Furthermore, based on the process labels of crushing, grinding, flotation, and leaching, subsets of process operation data, control variable data, and actuator status data corresponding to each process are separated from the process-level operation data stream. For each process's physical control object (such as a jaw crusher in the crushing process or an overflow ball mill in the grinding process), a standardized description template containing status attributes, control interfaces, and behavioral rules is created (taking a jaw crusher as an example, the status attribute refers to the feed particle size, the control interface refers to the discharge port size, and the behavioral rules refer to the logic for calculating the product particle size). A mapping relationship is established between the status attributes in each standardized description template and the real-time process operation data and actuator status data in the corresponding data subset; a mapping relationship is established between the control interface and the control variable data in the data subset; and the corresponding calculation logic is configured for the behavioral rules based on the process principles. All standardized description templates that have been mapped and configured for all processes are collected to form a twin object description set.

[0042] It should be noted that the calculation logic of the behavior rules is based on empirical formulas established according to the process mechanism. By using real-time data mapped from state attributes and control interfaces as input parameters for mathematical operations, the target process parameters are calculated. For example, the behavior rules of a jaw crusher calculate the mass percentage of each particle size in the product using empirical formulas that include discharge port size and feed particle size.

[0043] S2.2: Based on the topological order and material transfer relationship between each process, instantiate the corresponding twin object from the twin object description set, and associate it with the process-level operation data flow to generate a set of twin object instances.

[0044] Furthermore, based on the topological sequence of the gold ore beneficiation process from crushing, grinding, flotation to leaching, standardized description templates for the corresponding processes are extracted sequentially from the twin object description set, and a unique instance identifier is assigned to each standardized description template to complete the twin object instantiation. The state attributes and control interfaces defined in each twin object instance are bound one-to-one with the real-time process operation data, actuator status data and control variable data under the corresponding process label in the process-level operation data stream through the instance identifier to establish a data-driven relationship. The twin object instances finally obtained from each process are summarized according to the process topological sequence to generate a twin object instance set.

[0045] S2.3: Logically link and configure the set of twin object instances according to the gold ore beneficiation and smelting process to construct a virtual topology network of process objects.

[0046] Furthermore, based on the topological sequence of the gold ore beneficiation process from crushing, grinding, flotation to leaching, each twin object instance in the twin object instance set is traversed sequentially; based on the material transfer relationship between processes, logical links are established between twin object instances of adjacent processes (such as linking the "product" output interface of the twin object instance of the previous process with the "feed" input interface of the twin object instance of the next process); the material type, flow range and transmission logic carried by the link are configured synchronously, thereby forming a virtual topology network of process objects that reflects the complete material flow and logical dependencies of the gold ore beneficiation production line.

[0047] It should be noted that the process object virtual topology network is the core structure used for full-process simulation, analysis and collaborative optimization of the physical production line of gold ore beneficiation in digital space. Through nodes and links, it accurately represents the logical connection relationship, material transfer path and state and control dependency relationship between the twin object instances corresponding to each process from crushing, grinding, flotation to leaching. This forms the digital basic framework for subsequent process-level simulation, linkage analysis and control strategy verification.

[0048] S2.4: Through the continuous inflow of process-level operation data flow, the state update and behavior evolution of each twin object instance in the virtual topology network of process objects are driven, and the operation state of each process control object is constrained in a consistent manner based on the evolution results, forming a dynamic simulation operation environment.

[0049] Furthermore, the continuously flowing process-level operational data stream is injected into the virtual topology network of process objects in real time. This drives each twin object instance in the virtual topology network to execute the behavioral rules defined in its standardized description template according to the bound data, so that the state attributes of the twin object instances are updated synchronously with the time series data. Based on the updated state attributes, the process behavior of each twin object instance is gradually calculated and evolved according to the configured behavioral rules in the twin object instances, resulting in simulation results that reflect the changes in the process operation. Based on the simulation results, the operating status of adjacent twin object instances is verified and coordinated according to the inter-process constraints defined in the process knowledge base (such as material balance and quality transfer between upstream and downstream processes) to ensure that it conforms to the process logic of actual production. Through continuous data-driven, state evolution and consistency constraint processes, a dynamic simulation operation environment that can be dynamically updated with the process-level operational data stream is formed.

[0050] S2.5: Integrate and encapsulate the dynamic simulation runtime environment with the virtual topology network of the process object to generate a process-level digital twin runtime.

[0051] Furthermore, the dynamic simulation runtime environment is integrated with the virtual topology network of the process object, combining the real-time data-driven and state evolution capabilities of the dynamic simulation runtime environment with the structured logical relationships of the virtual topology network of the process object. This is encapsulated into a computing entity with a unified input / output interface that can run independently, thereby generating a process-level digital twin runtime that can map the running status of the entire gold ore beneficiation and smelting process in real time and support simulation and deduction.

[0052] S3. Based on the operating status of the process-level digital twin, analyze the response relationship between the changes in control variables of each process and the process operation results, identify key coupled control variables, and construct a cross-process coupling relationship map.

[0053] S3.1: In the process-level digital twin runtime, adjust the control variable settings of each process, synchronously simulate and record the resulting changes in process execution results, and generate a variable-result response dataset.

[0054] Furthermore, within the process-level digital twin runtime, based on the reasonable operating range of each control variable defined in the process knowledge base, the control variable settings for each process from crushing, grinding, flotation to leaching are modified sequentially (e.g., the grinding fineness setting for the grinding process and the reagent addition setting for the flotation process). After each adjustment, the process-level digital twin runtime is driven to perform full-process simulation calculations based on the new control variable settings, and the simulation-generated process operation result change data (e.g., final concentrate grade and metal recovery rate) are recorded synchronously. The control variable identifiers, adjusted settings, and corresponding process operation result change data are associated and stored to generate a variable-result response dataset.

[0055] It should be noted that, Figure 5This study demonstrates the overall trend of metal recovery rate over time under different control strategies (such as fixed control, single-process regulation, and cross-process linkage control) under the same ore grade disturbance and feeding conditions. The upper part is an overview curve covering the entire time range, which visually shows the long-term recovery rate evolution level of each control strategy and their mutual approximation or differentiation relationships throughout the entire process. It can be observed that cross-process linkage control maintains a higher recovery rate level in most time periods, while fixed control and single-process regulation exhibit more obvious fluctuations and lag response characteristics. The lower part refines and magnifies the local time window of the intermediate operation stage (such as 120–180s), which more clearly shows the significant relative rise and fall differences of the three recovery rate curves in the intermediate operation stage. The lower point corresponds to the critical moment when the difference in the suppression and recovery capabilities of different control strategies is most significant when the ore grade changes suddenly or the process disturbance is amplified. This visually demonstrates the advantages of cross-process linkage control in terms of recovery rate stability and recovery speed under dynamic disturbance conditions, providing key node basis for subsequent digital twin simulation and collaborative control effect evaluation.

[0056] S3.2: Based on the variable-result response dataset, analyze the range of processes affected by the adjustment of each control variable, and determine the list of processes affected by each control variable.

[0057] Furthermore, the system iterates through all the adjustments recorded in the variable-response dataset. For each control variable, it extracts the process operation result change data corresponding to the change in the control variable set value, and calculates the numerical offset of the process operation data (such as concentrate grade, tailings grade, and recovery rate) corresponding to each process operation in the process operation result change data. Processes whose offset exceeds the normal fluctuation range corresponding to the "reasonable operating range" defined for the corresponding process in the process knowledge base are identified as processes affected by the control variable adjustment. The system summarizes all affected process identifiers and generates a list of processes affected by the control variable.

[0058] It should be noted that, Figure 6Using time as the horizontal axis and different control strategies (such as fixed control, single-process regulation, and cross-process linkage control) as the vertical axis, each color block represents the concentrate grade level of the corresponding control strategy at that moment. In terms of color meaning, darker colors (blue-purple) indicate lower concentrate grade, while lighter colors (yellow-green) indicate higher concentrate grade. Therefore, the changes in color blocks over time can intuitively reflect the dynamic evolution of concentrate grade under different control strategies. From the overall color block distribution, it can be seen that concentrate grade is not statically stable, but exhibits obvious phased increases and decreases under the influence of ore disturbances and process responses. Cross-process linkage control shows a larger range of lighter-colored areas in multiple time intervals, indicating that it can maintain a higher grade level continuously under dynamic conditions. Fixed control and single-process regulation, on the other hand, show darker banded areas in certain time periods, reflecting that grade deterioration is more likely to occur when disturbances are amplified or process mismatches occur. Figure 6 The spatiotemporal evolution trajectory of concentrate grade was characterized by time-series thermodynamics, which enabled the rapid location of key time intervals where "grade is unstable or changes significantly", thus providing a visual basis for the screening and optimization of collaborative control strategies in digital twin environments.

[0059] S3.3: Select control variables whose influence spans a single process from the list of processes affected by each control variable, and use them as key coupling control variables.

[0060] Furthermore, the list of processes affected by each control variable is examined sequentially, and the number of process identifiers contained in the list of processes affected by each control variable is counted; control variables containing two or more different process identifiers are determined to have an impact range that spans a single process; all control variables that meet the conditions are selected and identified as key coupled control variables.

[0061] S3.4: Analyze the degree and direction of mutual influence between changes in key coupled control variables, and generate a variable coupling relationship matrix.

[0062] Furthermore, for each pair of different key coupled control variables, records in which the set values ​​of both change simultaneously are extracted from the variable-result response dataset. The degree of mutual influence of the changes is quantified by calculating the Pearson correlation coefficient, and the direction of influence is indicated by the positive or negative Pearson correlation coefficient (positive value indicates unidirectional influence, negative value indicates inverse influence). The Pearson correlation coefficient values ​​between all pairs of key coupled control variables are organized into a square matrix in which both rows and columns are variable identifiers to generate a variable coupling relationship matrix.

[0063] The expression for calculating the Pearson correlation coefficient is: ; in, Indicates key coupling control variables and The Pearson correlation coefficient between the two is a statistic that measures the degree of linear correlation between the changing trends of two key coupled control variables. This represents the key coupling control variables extracted from the "variable-result response dataset". and The total number of records in which the set value changes effectively at the same time; and They represent the first time. Among the valid records, key coupled control variables and The change in the set value relative to its previous value; and These represent the key coupling control variables. and In all The average change in each valid record.

[0064] S3.5: Using key coupling control variables as nodes and coupling relationships quantified by the variable coupling relationship matrix as edges, construct a cross-process coupling relationship graph.

[0065] Furthermore, graph nodes are created using each key coupled control variable and its associated process information as elements. Based on the Pearson correlation coefficient value and influence direction of each pair of key coupled control variables recorded in the variable coupling relationship matrix, connection edges with Pearson correlation coefficient and direction attributes are created between the corresponding nodes. By combining all nodes and edges, a cross-process coupling relationship graph that can intuitively display the mutual influence relationship between key coupled control variables is formed.

[0066] S4. Analyze the linkage relationship of the cross-process coupling relationship graph, determine the cross-process linkage control action set, and combine and evolve the cross-process linkage control action set in the process-level digital twin runtime to obtain the candidate collaborative control strategy set.

[0067] S4.1: Perform topology analysis and pattern recognition on the cross-process coupling relationship graph to parse out the strong coupling paths and linkage relationships between key coupling control variables, forming a set of linkage relationship patterns.

[0068] Furthermore, based on the topological structure of the cross-process coupling relationship graph, variable transmission paths formed by two or more key coupling control variable nodes connected by edges are identified. For each variable transmission path, the comprehensive coupling strength is calculated (e.g., taking the average of the absolute values ​​of the Pearson correlation coefficients of each connecting edge in the path), and compared with a preset strong coupling judgment threshold. Variable transmission paths exceeding the strong coupling judgment threshold are judged as strong coupling paths. At the same time, the linkage direction relationship between each key coupling control variable in the strong coupling path is recorded (e.g., if all connecting edges have positive Pearson correlation coefficients, it indicates linkage in the same direction; if some connecting edges have negative Pearson correlation coefficients, it indicates linkage in the opposite direction). Based on the node connection order, comprehensive coupling strength, and linkage direction characteristics of the strong coupling paths, the strong coupling paths are classified and summarized to form a set of linkage relationship patterns.

[0069] It should be noted that the strong coupling judgment threshold is set by analyzing the distribution of the absolute values ​​of the Pearson correlation coefficients of all key coupled control variable pairs in the historical production data accumulated over a long period of time, selecting a specific quantile (such as the top 25% quantile) as the initial empirical value, and then correcting it by combining the qualitative knowledge of inter-process constraints and process mechanisms in the process knowledge base. An exemplary value range is 0.5 to 0.8. If it is higher than 0.8, it is easy to miss some actual significant coupling relationships, causing the set of linkage relationship patterns to miss important control paths. If it is lower than 0.5, it is easy to include a large number of weakly correlated or accidentally correlated key coupled control variable pairs in the strong coupling path, resulting in an overly complex set of linkage relationship patterns and the introduction of noise interference.

[0070] S4.2: Regularize the set of linkage relationship patterns, extract the linkage timing logic and collaborative constraints of key coupled control variables, and generate a set of cross-process linkage control rules.

[0071] Furthermore, the strongly coupled paths recorded in the set of linkage relationship patterns are read one by one. According to the node connection order of the key coupled control variables in the strongly coupled path, the order of action of each key coupled control variable in cross-process linkage control is clarified (e.g., "grinding fineness" must precede "flotation reagent dosage"). The linkage direction is determined based on the sign of the Pearson correlation coefficient corresponding to the connection edge (e.g., a positive Pearson correlation coefficient is defined as "same-direction adjustment," and a negative one is defined as "reverse-direction adjustment"). Combining the reasonable operating range boundary values ​​of each key coupled control variable in the process knowledge base, the linkage adjustment ratio coefficient between paired key coupled control variables is calculated. The permissible adjustment limits are defined; the clear sequence of actions and linkage directions are specifically refined into a timing logic rule that stipulates that subsequent related variables must perform adjustments in the same or opposite direction after a specific time delay, triggered by a certain adjustment amount of a key coupled control variable; the calculated adjustment ratio coefficient and the permissible adjustment limits are refined into a numerical constraint rule that stipulates that when key coupled control variables are linked, the specific adjustment amount of one key coupled control variable must correspond to the change of another key coupled control variable within a specific numerical range; the timing logic rules and numerical constraint rules are integrated to generate a cross-process linkage control rule set.

[0072] S4.3: Based on the cross-process linkage control rule set, combine the control values ​​and adjustment methods of different key coupled control variables to generate a cross-process linkage control action set.

[0073] Furthermore, each timing logic rule in the cross-process linkage control rule set is analyzed to extract a complete action sequence triggered by a specific adjustment amount of the key coupled control variable, including the identifier of the subsequent associated key coupled control variable, adjustment direction, and delay requirements. Simultaneously, the numerical constraint rules in the cross-process linkage control rule set are analyzed to obtain the linkage adjustment ratio coefficient and allowable adjustment boundary between each pair of key coupled control variables. Based on the order of the action sequence, the adjustment requirements of different key coupled control variables are combined. For each key coupled control variable requiring adjustment in the action sequence, the specific control value that meets the boundary conditions is calculated based on the linkage adjustment ratio coefficient and allowable adjustment boundary provided by the numerical constraint rules. Based on the adjustment direction and delay requirements specified by the timing logic rules, the adjustment method and execution time of each control value are determined. The control values, adjustment methods, and execution times of all key coupled control variables are integrated into a unified, independently sendable control command, generating a cross-process linkage control action set.

[0074] The expression for calculating the control value is: ; in, This indicates the controlled values ​​of the related variables that need to be calculated; and These represent the related variables. The lower and upper limits of the allowable adjustment amount; Representing related variables Initial values ​​before performing the adjustment action; This indicates the relationship between the trigger variable and the associated variable. The linkage adjustment ratio coefficient; This indicates a specific adjustment amount for the trigger variable.

[0075] It should be noted that when calculating the control values, the linkage adjustment ratio coefficient is used. The values ​​already include the necessary unit conversion factors to ensure... The calculation result and the initial value of the associated variable They have the same physical dimensions.

[0076] S4.4: Based on the cross-process linkage control action set, the action set is combined, serialized and parameter adjusted and evolved in the process-level digital twin operation body to generate a candidate collaborative control strategy evolution set.

[0077] Furthermore, within the process-level digital twin runtime, multiple control instructions are selected from the cross-process linkage control action set and grouped according to different combination logics to form various action combination schemes. For the multiple control instructions in each action combination scheme, they are sorted and timestamped according to the production cycle and process sequence to complete the serialization and arrangement of control instructions. Based on the serialization and arrangement, the control parameters in the control instructions are iteratively adjusted in multiple rounds according to the fixed parameter adjustment step size (such as 5% to 10% of the allowable adjustment range of key coupled control variables) and the evolutionary algebra (such as 3 to 5) under the dual constraints of the allowable adjustment boundary defined by the cross-process linkage control rule set and the reasonable operating range of each key coupled control variable in the process knowledge base. Through the above combination, serialization and parameter adjustment evolution operations, a large number of candidate collaborative control strategies containing complete control instruction sequences and specific parameters that can be used for simulation testing are generated, forming a candidate collaborative control strategy evolution set.

[0078] S4.5: In the process-level digital twin runtime, simulate the evolution set of candidate collaborative control strategies, and perform performance evaluation and screening based on the simulation results to obtain the candidate collaborative control strategy set.

[0079] Furthermore, in the process-level digital twin runtime, a full-process simulation is performed on each candidate collaborative control strategy in the candidate collaborative control strategy evolution set to obtain process index data from the simulation results of each candidate collaborative control strategy. Based on the process optimization objectives (such as the highest total recovery rate and the lowest energy consumption), the process index data is converted into a comprehensive performance evaluation value for each candidate collaborative control strategy through weighted calculation, and all candidate collaborative control strategies are sorted in descending order based on the comprehensive performance evaluation value. The top-ranked (e.g., the top 10) candidate collaborative control strategies with the highest comprehensive performance evaluation values ​​are selected to generate a candidate collaborative control strategy set.

[0080] It should be noted that when calculating the comprehensive performance evaluation value, the preliminary importance ranking of process indicators such as total recovery rate, concentrate grade, and energy consumption is determined based on the process priority and production target defined in the process knowledge base. Combined with historical production data, the impact of the actual fluctuations of each process indicator on the overall economic benefits is statistically analyzed, and the preliminary importance ranking is quantitatively corrected. Finally, a multi-objective decision-making method (such as the analytic hierarchy process) is adopted to combine qualitative ranking with quantitative impact analysis to calculate the specific weight coefficient values ​​of each process indicator.

[0081] S5. Verify the consistency of cross-process linkage and resolve control conflicts for the candidate collaborative control strategy set, and determine the cross-process linkage control action combination.

[0082] S5.1: Extract the cross-process linkage control actions and control variable adjustment sequence of each collaborative control strategy from the candidate collaborative control strategy set to form a linkage action description.

[0083] Furthermore, the control instruction sequence of each candidate collaborative control strategy in the candidate collaborative control strategy set is read one by one, and the target variable identifier, control value, adjustment direction and execution time parameter in each control instruction are parsed. According to the value of the execution time parameter, all control instructions are sorted in chronological order, and the sorted control instruction sequence is described in a standardized format of "at [execution time], [target variable] is adjusted to [control value] in [adjustment direction]", generating a description of the linkage action of each candidate collaborative control strategy.

[0084] S5.2: Based on the description of the linkage action, perform consistency verification on the adjustment direction, timing and range of the cross-process linkage control action, identify and resolve control conflicts, and generate a consistent linkage control action set.

[0085] Furthermore, the adjustment direction of each target variable in the linkage action description is compared with the corresponding linkage direction in the cross-process linkage control rule set to check for any directional contradictions. If a contradiction exists, it is forcibly corrected to the linkage direction specified in the cross-process linkage control rule set. Based on the linkage timing logic defined in the cross-process linkage control rule set, the execution time order of different target variables in the linkage action description is checked to see if it conforms to the pre-order dependency relationship between processes. If it does not conform, the execution time is reordered according to the process priority defined in the cross-process linkage control rule set. The control value is compared with the reasonable operating range of the corresponding target variable in the process knowledge base to check if the control value exceeds the limit. If it exceeds the limit, the control value is corrected in conjunction with the linkage adjustment ratio coefficient defined in the cross-process linkage control rule set. All linkage action descriptions that have passed the verification are summarized to generate a consistent linkage control action set.

[0086] S5.3: Verify the executability of the set of consistent linkage control actions and determine the combination of cross-process linkage control actions.

[0087] Furthermore, based on the description of each linkage action in the consistent linkage control action set, it is checked whether the adjustment parameters in the linkage action description conform to the reasonable operating range defined in the process knowledge base, and the execution time in the linkage action description is determined to ensure that each linkage action is ordered according to the priority order of the process and the dependency relationship in the process flow, and that the adjustment order will not cause process conflicts; it is verified whether the adjustment amount in the linkage action description is within the allowable adjustment range defined in the cross-process linkage control rule set, and to ensure that the change range of the control variable meets the process safety requirements; after passing all the above checks, the set of linkage action descriptions with the highest execution efficiency is selected, and all control actions and precise parameters contained therein are determined as the cross-process linkage control action combination.

[0088] This embodiment also provides a computer device applicable to the gold ore beneficiation and smelting process optimization method based on digital twins, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the gold ore beneficiation and smelting process optimization method based on digital twins as proposed in the above embodiment.

[0089] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0090] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the gold ore beneficiation and smelting process optimization method based on digital twins as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0091] In summary, this invention achieves full-process digital synchronization and high-fidelity simulation by generating a process-level digital twin runtime, providing a virtual experimental environment for dynamic monitoring and predictive analysis, and enhancing process visualization and decision support capabilities; by acquiring a set of candidate collaborative control strategies, it realizes the automated generation and iterative optimization of multivariable control strategies; and by conducting efficient simulation verification in virtual space, it reduces trial-and-error costs in actual production and enhances optimization efficiency and reliability.

[0092] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A digital-twin-based gold ore dressing process optimization method, characterized in that: include, Collect process operation data, control variable data, and actuator status data of each process in the gold ore beneficiation and smelting production line, and preprocess and associate them with the process topology to generate process-level operation data streams; Based on the process-level operation data flow, a virtual topology network of process objects corresponding to the control objects of each process in the gold ore beneficiation and smelting production line is constructed, and a mapping relationship between the operation status and real-time data of each process control object is established to generate a process-level digital twin operation body. Based on the operational status of the process-level digital twin, the response relationship between the changes in control variables of each process and the process operation results is analyzed, key coupled control variables are identified, and a cross-process coupling relationship map is constructed. The linkage relationship of the cross-process coupling relationship map is analyzed to determine the cross-process linkage control action set, and the cross-process linkage control action set is combined and evolved in the process-level digital twin operation to obtain the candidate collaborative control strategy set; The candidate collaborative control strategy set is subjected to cross-process linkage consistency verification and control conflict resolution to determine the cross-process linkage control action combination.

2. The digital-twin-based gold mine beneficiation process optimization method according to claim 1, characterized in that: The steps for generating the process-level operational data stream are as follows: The preprocessing includes time alignment, outlier removal, and missing data completion. Based on the topological sequence and material transfer relationships between each process, the pre-processed process operation data, control variable data, and actuator status data are correlated and integrated to generate process-level operation data streams.

3. The digital-twin-based gold mine beneficiation process optimization method according to claim 1, characterized in that: The steps for constructing the virtual topology network of process objects corresponding to the control objects of each process in the gold ore beneficiation and smelting production line are as follows: Based on the process-level operation data flow, digital twin abstractions are performed on the physical control objects of each process to establish a set of twin object descriptions; Based on the topological order and material transfer relationship between each process, the corresponding twin object is instantiated from the twin object description set and associated with the process-level operation data flow to generate a set of twin object instances; The collection of twin object instances is logically linked and configured according to the gold ore beneficiation and smelting process to construct a virtual topology network of process objects.

4. The digital-twin-based gold mine beneficiation process optimization method according to claim 1, characterized in that: The steps for establishing the mapping relationship between the operating status and real-time data of each process control object, and generating a process-level digital twin operating entity, are as follows: By continuously receiving process-level operation data streams, the state updates and behavior evolution of each twin object instance in the virtual topology network of process objects are driven, and the operation state of each process control object is constrained to a uniformity based on the evolution results, thus forming a dynamic simulation operation environment. The dynamic simulation runtime environment is integrated and encapsulated with the virtual topology network of the process object to generate a process-level digital twin runtime.

5. The digital-twin-based gold mine beneficiation process optimization method according to claim 1, characterized in that: The analysis of the response relationship between changes in control variables at each process stage and the process operation results, and the identification of key coupled control variables, are carried out in the following steps. In the process-level digital twin runtime, adjust the control variable settings of each process, synchronously simulate and record the resulting changes in process execution results, and generate a variable-result response dataset. Based on the variable-result response dataset, we analyze the range of processes affected by the adjustment of each control variable and determine the list of processes affected by each control variable. From the list of processes affected by each control variable, control variables whose influence spans across a single process are selected as key coupled control variables.

6. The digital-twin-based gold mine beneficiation process optimization method according to claim 1, characterized in that: The steps for constructing the cross-process coupling relationship graph are as follows: Analyze the degree and direction of mutual influence among the changes of key coupled control variables, and generate a variable coupling relationship matrix; Using key coupling control variables as nodes and coupling relationships quantified by the variable coupling relationship matrix as edges, a cross-process coupling relationship graph is constructed.

7. The digital-twin-based gold mine beneficiation process optimization method according to claim 1, characterized in that: The steps for combining and evolving cross-process linkage control action sets within the process-level digital twin runtime to obtain a candidate collaborative control strategy set are as follows: Topological analysis and pattern recognition are performed on the cross-process coupling relationship map to extract the strong coupling paths and linkage relationships between key coupling control variables, forming a set of linkage relationship patterns. The set of linkage relationship patterns is regularized, and the linkage timing logic and collaborative constraints of key coupled control variables are extracted to generate a set of cross-process linkage control rules. Based on the cross-process linkage control rule set, the control values ​​and adjustment methods of different key coupled control variables are combined to generate a cross-process linkage control action set; Based on the cross-process linkage control action set, the action set is combined, serialized and parameter adjusted in the process-level digital twin operation body to generate a candidate collaborative control strategy evolution set; In the process-level digital twin runtime, the evolution set of candidate collaborative control strategies is simulated and run, and the performance is evaluated and screened based on the simulation results to obtain the candidate collaborative control strategy set.

8. The digital-twin-based gold mine beneficiation process optimization method according to claim 1, characterized in that: The steps for determining the cross-process linkage control action combination are as follows: Extract the cross-process linkage control actions and control variable adjustment sequence of each collaborative control strategy from the candidate collaborative control strategy set to form a linkage action description; Based on the description of the linkage action, the adjustment direction, timing and range of the cross-process linkage control action are checked for consistency, control conflicts are identified and resolved, and a consistent linkage control action set is generated. Perform executability verification on the set of consistent linkage control actions and determine the combination of cross-process linkage control actions. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is characterized in that: When the processor executes the computer program, it implements the steps of the gold ore beneficiation and smelting process optimization method based on digital twins as described in any one of claims 1 to 8.

10. A computer readable storage medium having stored thereon a computer program, characterized in that: When the computer program is executed by the processor, it implements the steps of the gold ore beneficiation and smelting process optimization method based on digital twins as described in any one of claims 1 to 8.

Images