Beneficiation production whole-process process index optimization decision method and system

Abstract

This invention discloses a method and system for optimizing process indicators throughout the entire mineral processing production process, specifically relating to the field of industrial control system technology. It addresses the problem of difficulty in optimizing global process indicators due to independent control of each process step in existing mineral processing production processes. By acquiring real-time process data from multiple processes throughout the mineral processing production process, a state-space recursive graph is constructed and its complex network characteristics are analyzed to determine the existence of key transmission hubs throughout the process. If such hubs exist, the impact of dynamic interaction effects between processes centered on these key transmission hubs on the transmission of process indicator targets is analyzed. Based on the dynamic interaction effects, the key transmission path causing nonlinear distortion of upstream process optimization instructions after execution in downstream processes is traced. The degree of game-theoretic conflict between the control objectives of each process is assessed based on the key transmission path. According to the degree of game-theoretic conflict, the local control objectives of each process are adjusted through the industrial control system to generate optimization decisions for the entire mineral processing production process.

Description

Technical Field

[0001] This invention relates to the field of industrial control system technology, and more specifically, to a method and system for optimizing process parameters throughout the entire mineral processing production process. Background Technology

[0002] In the field of mineral processing, the application of industrial control systems has penetrated into various processes such as crushing, grinding, sorting and dewatering. Existing technologies generally adopt automated control strategies for single processes or local units. By setting independent process indicators within each process and using feedback control mechanisms, production stability is maintained. This local optimization mode based on processes has a certain effect in ensuring the stability of operating parameters of specific processes and constitutes the basis of current process control in the mineral processing industry.

[0003] However, due to the close physical and chemical coupling and material transfer relationship between the various processes in the mineral processing, it is difficult to coordinate the conflict between the control objectives of different processes by independently optimizing and controlling local processes. As a result, the optimization and adjustment of local process indicators cannot be effectively transmitted and contribute to the improvement of overall production efficiency, making it difficult to optimize the process indicators of the entire process. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for optimizing the process indicators of the entire mineral processing production process to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Decision-making methods for optimizing process parameters throughout the entire mineral processing production process include: S1. Obtain real-time process data for multiple processes throughout the entire mineral processing production process; S2. Based on real-time process data, by constructing a state-space recursive graph and analyzing its complex network characteristics, determine whether there are key transmission hubs in the entire process. S3. When a key transmission hub exists, analyze the impact of the dynamic interaction effect between processes centered on the key transmission hub on the transmission of process indicators. S4. Based on the impact of dynamic interaction effects on the transmission of process indicators, trace the key transmission path that causes nonlinear distortion of upstream process optimization instructions after execution in downstream processes. S5. Based on the key transmission path, assess the degree of game-theoretic conflict between the control objectives of each process. S6. Based on the degree of game conflict, adjust the local control objectives of each process through the industrial control system to generate optimization decisions for the entire mineral processing production process indicators.

[0006] Furthermore, real-time process data for multiple steps in the entire mineral processing production process are obtained, including: Collect particle size parameters for the crushing process, concentration and load parameters for the grinding process, separation operation parameters for the sorting process, and dewatering effect parameters for the dewatering process. Particle size parameters, concentration and load parameters, sorting operation parameters, and dewatering effect parameters are combined into real-time process data; real-time process data is collected by sensors and process control systems deployed in each process step.

[0007] Furthermore, based on real-time process data, by constructing a state-space recursive graph and analyzing its complex network characteristics, it is determined whether there are key transmission hubs in the entire process, including: A state-space recursion graph is constructed based on real-time process data, and the density distribution of recursion points in the state-space recursion graph is calculated: real-time process data is converted into multi-dimensional state vectors through phase space reconstruction, grid partitioning is used to statistically analyze the distribution frequency of recursion points in the state space, and the kernel density estimation method is used to calculate the density distribution of recursion points. Complex network characteristics are extracted from the state-space recursive graph, and node centrality indices are calculated, including betweenness centrality and proximity centrality. Based on the numerical distribution of node centrality indicators, identify the process nodes that play a key transmission role in the entire process; Based on the number and distribution characteristics of process nodes with key transmission functions, determine whether there are key transmission hubs in the entire process.

[0008] Furthermore, based on the number and distribution characteristics of process nodes with key transmission functions, determining whether there is a key transmission hub in the entire process includes: setting a centrality threshold for node centrality indicators, counting the number of process nodes that exceed the centrality threshold, analyzing the topological positional relationship of process nodes in the entire process network, and determining the existence of a key transmission hub when process nodes that exceed the centrality form a connected subgraph and are located on the critical path.

[0009] Furthermore, when key transmission hubs exist, the impact of the dynamic interaction effects between processes centered on these key transmission hubs on the transmission of process performance targets is analyzed, including: Based on the key transmission nodes containing key transmission functions, the dynamic interaction effect between process nodes is calculated. The dynamic interaction effect is quantified by the mutual information or transmission entropy of the changes in process parameters between process nodes. Analyze the correlation strength between key transmission function process nodes and adjacent process nodes; Determine the impact of dynamic interaction effects on the transmission of process indicators based on the correlation strength: Establish a mapping relationship between correlation strength and the transmission of process indicators, divide the correlation strength into level intervals, determine the impact of each level interval through statistical analysis, and calculate the final impact using a weighted average method; Establish the correspondence between key transmission function process nodes and process indicator targets.

[0010] Furthermore, based on the impact of dynamic interaction effects on the transmission of process performance targets, the key transmission paths that lead to nonlinear distortions in upstream process optimization instructions after execution in downstream processes are traced, including: Based on the impact of dynamic interaction effects on the transmission of process target indicators, the deviation between upstream process optimization instructions and downstream process execution results is identified. Calculate the degree of nonlinear distortion of the upstream process optimization instructions during the execution of the downstream process; Process connections whose nonlinear distortion exceeds a preset distortion threshold are selected based on the degree of nonlinear distortion. Analyze the path characteristics of process parameter target transmission in process connections that exceed the preset distortion threshold; The key transmission paths leading to nonlinear distortions are identified based on path characteristics.

[0011] Furthermore, calculating the degree of nonlinear distortion includes: establishing a transfer function model of the upstream process optimization command and the downstream process execution result, calculating the gain change of the transfer function under different operating conditions, analyzing the nonlinear characteristics of the gain change, and using the standard deviation and coefficient of variation of the gain change to quantify the degree of nonlinear distortion.

[0012] Furthermore, based on the key transmission path, the degree of game-theoretic conflict between the control objectives of each process is assessed, including: Identify the process control objectives corresponding to the process nodes on the critical transmission path; Calculate the conflict index values ​​between process control objectives. The conflict index values ​​are obtained through variance analysis of process control objectives. The degree of game-related conflict between process control objectives is determined based on the conflict index value, and the degree of game-related conflict is quantified by the Nash equilibrium distance; Establish the correspondence between key transmission paths and the degree of game conflict.

[0013] Furthermore, based on the degree of game conflict, the local control objectives of each process are adjusted through the industrial control system to generate optimization decisions for the entire mineral processing production process indicators, including: The adjustment weights of local control objectives for each process are determined based on the degree of game conflict, and the adjustment weights of local control objectives are positively correlated with the degree of game conflict. Calculate the adjustment amount of the local control target for each process based on the adjustment weight of the local control target; The local control target adjustment is executed by the industrial control system, and the adjustment of the local control target is achieved by the control algorithm of the industrial control system. Verify the synergistic effect between the adjusted local control objectives of each process; Based on the verification results, optimization decisions are generated for the entire mineral processing production process indicators.

[0014] On the other hand, the present invention provides a decision-making system for optimizing process indicators throughout the entire mineral processing production process, including: The data acquisition module is used to acquire real-time process data for multiple processes throughout the entire mineral processing production process; The hub identification module is used to determine whether there are key transmission hubs in the entire process based on real-time process data by constructing a state-space recursive graph and analyzing its complex network characteristics. The impact analysis module is used to analyze the impact of the dynamic interaction effect between processes centered on the key transmission hub on the transmission of process indicators when a key transmission hub exists. The path tracing module is used to trace the key transmission paths that cause nonlinear distortions in the execution of upstream process optimization instructions after downstream processes, based on the impact of dynamic interaction effects on the transmission of process indicator targets. The game theory evaluation module is used to assess the degree of game-theoretic conflict between the control objectives of each process based on key transmission paths. The decision generation module is used to generate optimization decisions for the entire mineral processing production process by adjusting the local control objectives of each process through the industrial control system based on the degree of game conflict.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By acquiring real-time process data from multiple processes throughout the mineral processing production process and constructing a state-space recursion graph based on this data to analyze the characteristics of complex networks, the system can accurately identify key transmission hubs in the entire process. This allows the system to deeply understand the intrinsic connections and dependencies between processes, providing a solid data foundation for subsequent optimization decisions. By analyzing the impact of dynamic interaction effects between processes centered on key transmission hubs on the transmission of process indicators, the system can effectively capture complex behavioral patterns under process coupling, avoiding overall performance degradation caused by local optimization. Furthermore, by tracing the key transmission paths that generate nonlinear distortions after the upstream process optimization instructions are executed in the downstream process, the system can accurately locate bottlenecks affecting the stability of the entire process, providing a basis for targeted adjustments. Based on the key transmission paths, the system assesses the degree of game-theoretic conflict between the control objectives of each process, enabling the system to quantify the competition and coordination relationships between processes, thus laying the foundation for global optimization. Finally, based on the degree of game-theoretic conflict, the system adjusts the local control objectives of each process through the industrial control system, generating optimization decisions for the entire process indicators, achieving collaborative control from local to global, and improving overall production efficiency.

[0016] 2. The full-process dynamic coordination mechanism implemented through the industrial control system, through state-space recursion graphs and complex network characteristic analysis, enables the system to monitor the transmission hubs between processes in real time, ensuring timely intervention at key nodes. Dynamic interaction effect analysis helps the system understand the transmission law of process indicators between processes, reducing the chain reaction of adverse reactions caused by local adjustments. Tracing the nonlinear distortion path allows the system to identify potential fault points in advance, enhancing the robustness of production. The assessment of the degree of game conflict provides a scientific basis for adjusting control objectives, avoiding the uncertainty brought about by subjective decision-making. Finally, the industrial control system executes optimization decisions to ensure that the control objectives of each process reach a balance in the conflict, achieving continuous optimization of the process indicators throughout the entire process, improving resource utilization and production stability. This industrial control method based on data-driven and system analysis effectively overcomes the limitations of traditional local optimization and provides an efficient and reliable global solution for mineral processing production. Attached Figure Description

[0017] Figure 1 This is a flowchart of the decision-making method for optimizing process indicators throughout the mineral processing production process of the present invention. Figure 2 This is a schematic diagram of the structure of the mineral processing production process index optimization decision system of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: Figure 1 This invention presents a method for optimizing process parameters throughout the entire mineral processing production process, including: S1. Obtain real-time process data for multiple processes throughout the entire mineral processing production process; S2. Based on real-time process data, by constructing a state-space recursive graph and analyzing its complex network characteristics, determine whether there are key transmission hubs in the entire process. S3. When a key transmission hub exists, analyze the impact of the dynamic interaction effect between processes centered on the key transmission hub on the transmission of process indicators. S4. Based on the impact of dynamic interaction effects on the transmission of process indicators, trace the key transmission path that causes nonlinear distortion of upstream process optimization instructions after execution in downstream processes. S5. Based on the key transmission path, assess the degree of game-theoretic conflict between the control objectives of each process. S6. Based on the degree of game conflict, adjust the local control objectives of each process through the industrial control system to generate optimization decisions for the entire mineral processing production process indicators.

[0020] S1. Obtain real-time process data for multiple processes throughout the entire mineral processing production process, specifically implemented as follows: Real-time process data is collected through sensors and process control systems deployed at each stage. In the crushing stage, particle size parameters are collected, specifically feed particle size and discharge particle size. Feed particle size refers to the size distribution of the material entering the crusher, which is detected online using a laser particle size analyzer mounted on the feed belt. For example, the detection range covers particle sizes from 0 mm to 300 mm, and the results are output as a particle size distribution curve, with the characteristic particle size d50 recorded as a key parameter. Discharge particle size refers to the size distribution of the material discharged from the crusher, which is monitored in real-time using a machine vision inspection system installed at the discharge port. The system acquires material flow images and extracts particle contour features using image processing algorithms to calculate the particle size distribution of the discharged material. The sampling frequency for both particle size parameters is set to, for example, once every 10 seconds, and the collected data is transmitted to the process control system via industrial Ethernet.

[0021] In the grinding process, concentration and load parameters are collected, specifically including feed concentration and mill load. Feed concentration is measured using a gamma-ray concentration meter installed on the mill feed pipeline. The meter calculates the percentage of solids in the slurry by detecting the attenuation of gamma rays after they pass through it; for example, the measurement range covers solids concentrations from 20% to 80%. Mill load is comprehensively monitored using power and vibration sensors installed on the mill drive system. The power sensor monitors the drive power of the mill motor in real time, while the vibration sensor collects vibration spectrum data of the mill cylinder. The power data and vibration spectrum characteristics are input into a pre-established load assessment model to calculate the mill's real-time load status, expressed as a percentage, for example, 100% represents the maximum design capacity.

[0022] In the sorting process, sorting operation parameters are collected, specifically including the reagent addition amount and the flotation cell level. The reagent addition amount is measured using an electromagnetic flowmeter installed on the reagent delivery pipeline. The flowmeter detects the volumetric flow rate of the reagent solution in real time, and calculates the actual added reagent mass based on the reagent concentration. The measurement result is recorded in kilograms per minute. The flotation cell level is monitored using an ultrasonic level gauge installed at the top of the flotation cell. The level gauge measures the distance between the liquid surface and the probe by emitting and receiving ultrasonic signals, thereby calculating the slurry level height in the cell. The level height is expressed in millimeters; for example, the measurement range is from 0 mm at the bottom of the cell to the overflow weir height.

[0023] In the dewatering process, parameters for the dewatering effect are collected, specifically including feed concentration and filter cake moisture content. Feed concentration is detected online using a concentration meter installed in the feed pipe of the dewatering equipment. A nuclear concentration meter is used to measure the solids content of the slurry, based on the gamma-ray absorption method, with a measurement range covering solids concentrations from 5% to 70%. Filter cake moisture content is detected online using an infrared moisture meter installed at the filter cake discharge point. The moisture meter calculates the percentage moisture content by measuring the absorption characteristics of the filter cake to infrared light of a specific wavelength. The detection results are transmitted to the process control system in real time.

[0024] Particle size parameters, concentration and load parameters, sorting operation parameters, and dewatering effect parameters are combined into real-time process data. The combination process includes data timestamp alignment, data validity verification, and data format standardization. First, the data collected from each process step is time-synchronized. Using the process control system's clock as a reference, data from different sampling times are unified to the same time point using interpolation methods; for example, the time synchronization interval is set to 5 seconds. Then, the validity of each parameter is checked, including the reasonableness of the numerical range, the continuity of data changes, and sensor status indicators. Data exceeding reasonable ranges or exhibiting abnormal changes are discarded. Finally, the validated data is combined according to a standardized format to form a standard data structure containing timestamps, process identifiers, parameter types, and parameter values, which is stored in a real-time database for subsequent analysis.

[0025] Real-time process data is acquired through sensors deployed at each stage and the process control system. The sensor network includes a laser particle size analyzer and machine vision inspection system for the crushing stage; gamma ray concentration meters, power sensors, and vibration sensors for the grinding stage; electromagnetic flow meters and ultrasonic level gauges for the sorting stage; and nuclear concentration meters and infrared moisture meters for the dewatering stage. These sensors are connected to the programmable logic controllers (PLCs) at each stage via fieldbus. The PLCs preprocess and digitize the sensor signals before uploading the data to the process control system via industrial Ethernet. The process control system employs a distributed architecture, including a data acquisition server, a real-time database, and an application server. The data acquisition server receives and caches data uploaded from each stage, the real-time database stores and manages real-time process data, and the application server provides data query and analysis services. The data acquisition cycle is configured based on parameter characteristics; for example, particle size and level parameters use a 10-second acquisition cycle, concentration and load parameters use a 5-second acquisition cycle, and moisture parameters use a 20-second acquisition cycle to ensure data real-time performance and integrity.

[0026] S2. Based on real-time process data, by constructing a state-space recursive graph and analyzing its complex network characteristics, determine whether there are key transmission hubs in the entire process. The specific implementation is as follows: A state-space recursion graph is constructed based on real-time process data, and the density distribution of recursive points in the graph is calculated. The state-space recursion graph is constructed using a phase space reconstruction method, which converts real-time process data into a multi-dimensional state vector. Specifically, key process parameter time series are selected from the real-time process data, such as feed particle size time series and discharge particle size time series. Phase space reconstruction is performed using the delayed coordinate method. The delay time is determined using the mutual information method; for example, the mutual information function of the time series is calculated, and the delay time corresponding to the first minimum point is selected. The embedding dimension is determined using the spurious nearest neighbor method; for example, the proportion of spurious nearest neighbors under different embedding dimensions is calculated, and the smallest embedding dimension with a proportion lower than a set threshold is selected. The threshold is determined based on the trend of the spurious nearest neighbor proportion as the embedding dimension increases; for example, by observing the curve of the spurious nearest neighbor proportion gradually decreasing from a high value, when the decrease in proportion is less than a specified value (e.g., less than 2% decrease for each additional embedding dimension), this proportion is selected as the threshold; for example, the threshold is set to 5%.

[0027] After converting real-time process data into multi-dimensional state vectors, a state-space recursion graph is constructed. This graph is a two-dimensional matrix where each element represents the recursive relationship between two state vectors in the state space. The recursive relationship is determined by calculating the Euclidean distance between the state vectors and setting a recursion threshold, for example, the 10th quantile of the state vector distance distribution. The density distribution of recursive points in the state-space recursion graph is calculated using a grid partitioning method. The state space is divided into uniform grid cells, for example, each dimension is divided into 10 equally spaced intervals. The frequency of recursive points within each grid cell is counted; this frequency refers to the total number of times all state vectors within that grid cell are identified as recursive points. The density distribution of recursive points is then calculated using a kernel density estimation method. A Gaussian kernel function is used, and the kernel bandwidth is determined using the Silverman rule, for example, the bandwidth equals the standard deviation multiplied by the negative fifth power of the sample size. The density value at the center of each grid cell is calculated; this density value represents the probability density of recursive points in that region, forming the density distribution graph of recursive points.

[0028] Complex network properties, including node centrality indices, are extracted from state-space recursive graphs. The calculation of node centrality indices is based on the complex network constructed from the state-space recursive graph. The construction of the complex network treats each state vector in the state-space recursive graph as a network node, and the connections between nodes are determined based on recursive relationships. For example, if the state vectors corresponding to two nodes have a recursive relationship in the state-space recursive graph, then a connection edge is established.

[0029] Node centrality metrics include betweenness centrality and proximity centrality. Betweenness centrality is calculated by statistically analyzing the proportion of shortest paths passing through a given node. Specifically, this involves calculating the shortest paths between all pairs of nodes in the network, using Dijkstra's algorithm. For example, the edge weights are calculated using the reciprocal of the distance between nodes, and each node is counted by dividing the number of shortest paths by the total number of shortest paths. Proximity centrality is calculated by taking the reciprocal of the sum of the shortest path distances from a node to all other nodes in the network. This is done by calculating the sum of the shortest path distances from each node to all other nodes, for example, using the Floyd-Warshall algorithm to calculate the shortest path distances between all pairs of nodes, and then taking the reciprocal of the sum as the proximity centrality value. Node centrality metrics are expressed in a standardized form; for example, both betweenness centrality and proximity centrality are normalized to the range of 0 to 1.

[0030] Based on the numerical distribution of node centrality indices, process nodes that play a key transmission role in the entire process are identified. The numerical distribution of node centrality indices is analyzed using statistical descriptive methods, such as calculating the mean, standard deviation, and quantiles of the node centrality indices. Identification of process nodes with key transmission roles is based on threshold settings for node centrality indices. These thresholds are set using statistical methods; for example, the betweenness centrality threshold is set to the median of all node betweenness centrality values ​​plus one standard deviation, and the closeness centrality threshold is set to the upper quartile of all node closeness centrality values. A process node is identified as a key transmission role process node when its node centrality indices simultaneously exceed the set thresholds. For example, a process node is identified as a key transmission role process node when its betweenness centrality is greater than 0.6 and its closeness centrality is greater than 0.7. Key transmission role process nodes correspond to specific processes in the entire mineral processing production process, such as a control node in the crushing, grinding, sorting, or dewatering processes.

[0031] Based on the number and distribution characteristics of process nodes with key transmission functions, it is determined whether there are key transmission hubs in the entire process. The number of process nodes with key transmission functions is obtained through counting methods, such as counting the total number of all process nodes identified as key transmission functions. The distribution characteristics are determined by analyzing the topological positional relationships of the process nodes with key transmission functions in the network, including the connection relationships between nodes and path importance. The determination process includes setting a centrality threshold for node centrality indicators. The centrality threshold is dynamically adjusted based on the numerical distribution of node centrality indicators, for example, the centrality threshold is set to the 75th percentile of all node centrality indicator values. The number of process nodes exceeding the centrality threshold is counted. Process nodes exceeding the centrality threshold are those whose node centrality indicator values ​​are greater than the centrality threshold.

[0032] The analysis focuses on the topological relationships of process nodes within the overall process network. This network is a directed graph where nodes represent processes and edges represent material or information flows. Topological relationships are evaluated by calculating the degree centrality and eigenvector centrality of nodes. For example, degree centrality represents the number of connections a node has, and eigenvector centrality represents the degree of connection between a node and its height-central node. When process nodes exceeding a certain centrality form a connected subgraph and are located on the critical path, a critical transmission hub is identified. A connected subgraph is defined as a network of nodes connected by paths that do not contain other non-critical nodes. The critical path is the longest path or the most important material flow path from the start to the end point in the overall process network. The critical path is determined using the critical path method, which involves calculating the duration or resource consumption of all paths and selecting the most critical one. The determination criterion is that the number of nodes in the connected subgraph exceeds a set lower limit; for example, at least three process nodes with critical transmission functions form a connected subgraph, and this connected subgraph overlaps with the critical path.

[0033] S3. When a key transmission hub exists, analyze the impact of the dynamic interaction effect between processes centered on the key transmission hub on the transmission of process target indicators. Specifically, this is implemented as follows: When a critical transmission hub exists, the impact of the dynamic interaction effect between processes centered on the critical transmission hub on the transmission of process indicators is analyzed. Based on the process nodes with key transmission functions contained in the critical transmission hub, the dynamic interaction effect between process nodes is calculated. The dynamic interaction effect is quantified by the mutual information or transmission entropy of the changes in process parameters between process nodes. The mutual information quantification method is achieved by calculating the difference between the joint probability distribution and the marginal probability distribution of the process parameter time series of two process nodes. Specifically, the process parameter time series data corresponding to the process nodes with key transmission functions are collected, such as the feed particle size time series and the discharge particle size time series. The time series are discretized into multiple intervals, and the frequency of occurrence of each interval is counted. The joint probability distribution and the marginal probability distribution are calculated. The mutual information value is equal to the expected value of the logarithm of the ratio of the product of the joint probability distribution and the marginal probability distribution. The transfer entropy quantification method achieves this by calculating the predictive information gain of the transmission of process parameter time series from one process node to the process index target of another process node. Specifically, it involves constructing an embedding vector of the process parameter time series, which includes current and historical values. The conditional probability distribution is then calculated. The transfer entropy value is equal to the expected value of the logarithm of the ratio of the conditional probability of the target node's current value given historical values ​​to the conditional probability of the target node's current value given source node and historical values. The quantification results of the dynamic interaction effect are expressed numerically; for example, the mutual information value ranges from 0 to 1, and the transfer entropy value ranges from 0 to positive infinity.

[0034] The analysis focuses on the correlation strength between key process nodes and adjacent process nodes. Correlation strength is determined by calculating the correlation coefficient or covariance of process parameter changes between process nodes. The correlation coefficient is calculated using the Pearson correlation coefficient method, which calculates the product of the covariance of the time series of process parameters for two process nodes divided by their respective standard deviations. The covariance is calculated as the average of the products of the deviations of the two time series from their respective means. Correlation strength can also be assessed using the Granger causality test for process parameter changes between process nodes. The Granger causality test determines the significance of causality by comparing the prediction errors of an autoregressive model that includes historical values ​​with that that does not. The numerical values ​​of correlation strength are expressed in standardized form; for example, the correlation coefficient ranges from -1 to +1, and the F-statistic of the Granger causality test ranges from 0 to positive infinity. The correlation strength analysis is based on real-time process data time series, and the time series length must meet statistical analysis requirements, such as including at least 100 sampling points to ensure statistical significance.

[0035] The influence of dynamic interaction effects on the transmission of process target indicators is determined based on the correlation strength. A mapping relationship between correlation strength and the transmission of process target indicators is established using regression analysis methods, such as linear or logistic regression. Correlation strength is used as the independent variable, and the transmission of process target indicators is used as the dependent variable. The transmission of process target indicators refers to the efficiency or consistency of process parameters transmission between processes, such as the effect of the discharge particle size in the crushing process on the feed concentration in the grinding process. The correlation strength is divided into level intervals based on the numerical distribution of the correlation strength. For example, the correlation strength is divided into three levels: low, medium, and high. Low level corresponds to a correlation coefficient less than 0.3, medium level corresponds to a correlation coefficient between 0.3 and 0.7, and high level corresponds to a correlation coefficient greater than 0.7. The influence of each level interval is determined through statistical analysis using analysis of variance (ANOVA). ANOVA calculates the ratio of between-group variance to within-group variance of the process target transmission values ​​for different level intervals to determine the significant influence of each level interval on the transmission of process target indicators. The final impact is calculated using a weighted average method. The weights of the weighted average are set based on the sample size or significance level of each level interval. For example, the weight is equal to the proportion of the sample size of each level interval to the total sample size. The final impact value is the weighted sum of the impact values ​​of each level interval.

[0036] Establish the correspondence between key transmission function process nodes and process indicator target transmission. This correspondence is achieved through association rule mining methods, employing either the Apriori algorithm or the FP-growth algorithm. The Apriori algorithm discovers association rules between key transmission function process nodes and process indicator target transmission through frequent itemset generation and rule extraction, while the FP-growth algorithm efficiently mines association rules by constructing a frequent pattern tree. Correspondence can also be established using clustering analysis methods, employing K-means clustering or hierarchical clustering. K-means clustering divides the key transmission function process nodes and process indicator target transmission data into multiple clusters, where nodes and transmission values ​​within each cluster share similar characteristics. Hierarchical clustering identifies correspondences by constructing a dendrogram based on the distances between data points. The establishment of correspondences is based on both historical and real-time data. Historical data is used to train the model, while real-time data is used to validate and update the correspondences. The correspondence is stored in the form of a rule set or a mapping table. For example, the rule set contains the process indicator target transmission of discharge particle size consistency if the process node of the key transmission function is the crushing process. The mapping table contains key-value pairs of process node identifiers of key transmission functions and process indicator target transmission values.

[0037] S4. Based on the impact of dynamic interaction effects on the transmission of process target indicators, the key transmission path that causes nonlinear distortion of upstream process optimization instructions after execution in downstream processes is traced, and the specific implementation is as follows: Based on the impact of dynamic interaction effects on the transmission of process target indicators, this study traces the key transmission path that leads to nonlinear distortion of upstream process optimization instructions after execution in downstream processes. It identifies the deviation between upstream process optimization instructions and downstream process execution results based on the dynamic interaction effect on the transmission of process target indicators. Specifically, this involves collecting time-series data of upstream process optimization instructions and downstream process execution results. Upstream process optimization instructions include, for example, the feed particle size setpoint for crushing, while downstream process execution results include, for example, the actual discharge particle size for grinding. Deviation is identified by calculating the difference between the two time series, for example, using the root mean square error (RMSE) method. The RMS error is equal to the square root of the average of the sum of squares of the differences between upstream process optimization instructions and downstream process execution results. Deviation identification is based on the influence of dynamic interaction effects. The dynamic interaction effect is adjusted by using the mutual information or transit entropy values ​​calculated in the aforementioned steps to adjust the deviation calculation weights. The weights are set as the normalized result of the dynamic interaction effect value, for example, the mutual information value divided by the maximum possible mutual information value. Data preprocessing includes timestamp alignment and outlier handling. Timestamp alignment is based on the process control system clock and uses linear interpolation to unify the data to the same time point. Outlier handling adopts the three-standard-deviation principle to remove data points that exceed three standard deviations of the mean.

[0038] The degree of nonlinear distortion of upstream process optimization instructions during downstream process execution is calculated. This nonlinear distortion is determined by establishing a transfer function model between the upstream process optimization instructions and the downstream process execution results. The transfer function model is established using a system identification method, such as an autoregressive exogenous model. The autoregressive exogenous model estimates its parameters using the least squares method. The model input is the time series of the upstream process optimization instructions, and the output is the time series of the downstream process execution results. The model order is determined using the Akaike Information Criterion, for example, selecting the model order that minimizes the Akaike Information Criterion value.

[0039] The gain variation of the transfer function under different operating conditions is calculated. Gain variation is determined by calculating the amplitude ratio of the transfer function at a specific frequency, such as the DC gain at 0 frequency. Different operating conditions include different process parameters, such as changes in feed concentration or mill load. Operating condition classification is based on cluster analysis of process parameter values, for example, using K-means clustering to divide operating conditions into multiple categories. The nonlinear characteristics of the gain variation are analyzed. Nonlinearity manifests as the fluctuation of the gain value with changes in operating conditions. The standard deviation and coefficient of variation of the gain variation are used to quantify the degree of nonlinear distortion. The standard deviation calculates the dispersion of the gain value sequence, and the coefficient of variation equals the standard deviation divided by the mean. The degree of nonlinear distortion is the weighted sum of the standard deviation and the coefficient of variation, with the weights set based on the significance of the operating condition variation, for example, the weight is the ratio of the variance between operating condition categories to the total variance.

[0040] Process connections whose nonlinear distortion exceeds a preset distortion threshold are selected based on the degree of nonlinear distortion. The preset distortion threshold is determined based on the statistical distribution of historical nonlinear distortion data. By analyzing the distribution characteristics of nonlinear distortion values ​​in historical data, an appropriate percentile is selected as the threshold; for example, the preset distortion threshold is set to the 90th percentile of historical nonlinear distortion values. This threshold is chosen based on the frequency of historical anomalies and adjusted by optimizing the false positive and false negative rates. For example, the classification performance corresponding to different percentiles is tested on the validation set, and a percentile value that balances precision and recall is selected. A process connection refers to a directed edge between an upstream process node and a downstream process node. The degree of nonlinear distortion of the process connection is obtained through the aforementioned calculation. The selection process compares the nonlinear distortion value of each process connection with the preset distortion threshold, retaining process connections whose nonlinear distortion value is greater than the preset distortion threshold. The selected process connections form a candidate set, which is used for subsequent path analysis. The candidate set is validated using cross-validation methods, such as dividing historical data into training and test sets, and verifying the stability of the selection results on the test set.

[0041] This analysis examines the path characteristics of process parameter transmission in process connections exceeding a preset distortion threshold. Path characteristics include path length, node degree centrality, transmission delay, and path redundancy. Path length refers to the number of hops from upstream to downstream processes, calculated using a breadth-first search algorithm that traverses the network from the upstream process node to the downstream process node. Node degree centrality represents the number of connections between nodes on the path, calculated by counting the number of adjacent edges between nodes. Transmission delay is determined by calculating the time difference between the upstream and downstream process parameter signals, analyzed using a cross-correlation function. The cross-correlation function calculates the correlation between two time series at different time offsets, and the transmission delay is taken as the time offset corresponding to the peak value of the cross-correlation function. Path redundancy indicates whether there are alternative paths on the path, determined by calculating the number of edge-independent paths, calculated using a maximum flow algorithm. The path characteristic analysis is based on the network topology of the entire mineral processing production process, constructed using process nodes and material flow edges.

[0042] The key transmission paths leading to nonlinear distortion are identified based on path characteristics. The identification of key transmission paths utilizes a multi-feature comprehensive evaluation method, such as a weighted scoring model. This model normalizes path feature values ​​and then sums them with weights based on feature importance. Feature importance is determined through principal component analysis or expert scoring; for example, principal component analysis calculates the contribution of each path feature to the overall variance, assigning higher weights to features with higher contributions. Path features include path length, node degree centrality, transmission delay, and path redundancy. The normalization method uses min-max normalization, mapping feature values ​​to a range of 0 to 1. The weighted scoring model calculates a comprehensive score for each path, and paths with high comprehensive scores are identified as key transmission paths. Key transmission paths may involve multiple process nodes, such as a path from crushing to grinding to sorting. The process connections along the path exhibit high degrees of nonlinear distortion and significant path characteristics. Validation of key transmission paths involves historical data backtesting, such as checking the frequency of the path's occurrence in historical anomalies; paths with high frequencies are identified as key transmission paths.

[0043] S5. Based on the key transmission path, assess the degree of game-theoretic conflict between the control objectives of each process. The specific implementation is as follows: Based on the critical transmission path, assess the degree of game-theoretic conflict between the control objectives of each process. Identify the process control objectives corresponding to the process nodes on the critical transmission path. The critical transmission path originates from the critical transmission path that leads to nonlinear distortion, as determined in the preceding steps. A process node refers to a specific process unit in the entire mineral processing production process, such as a crushing process node, a grinding process node, a separation process node, or a dewatering process node. A process control objective refers to the target values ​​of process parameters that each process needs to maintain or achieve, such as the target values ​​of feed particle size and discharge particle size in the crushing process, the target values ​​of feed concentration and mill load in the grinding process, the target values ​​of reagent addition and flotation cell level in the separation process, and the target values ​​of feed concentration and filter cake moisture in the dewatering process.

[0044] The identification process is achieved by querying the process control target database. This database stores the mapping relationship between each process node and its corresponding process control target. This mapping relationship is established based on the mineral processing flow design document and historical operational data. For example, crushing process nodes are mapped to feed particle size target values ​​and discharge particle size target values. The identification result is a set containing the process control targets corresponding to all process nodes along the critical transmission path. Data preprocessing includes verifying the validity and consistency of the process control target values. Validity checks ensure that the target values ​​are within a reasonable range; for example, the feed particle size target value ranges from 0 mm to 300 mm. Consistency checks ensure that the target values ​​do not contradict other process parameters; for example, the feed concentration target value for the grinding process is consistent with the discharge particle size target value for the crushing process.

[0045] The conflict index values ​​between process control objectives are calculated using an analysis of variance (ANOVA) of the process control objectives. The ANOVA employs a one-way ANOVA method, treating the process control objectives as factors and the values ​​of different process control objectives as treatment groups. Between-group variance and within-group variance are calculated. Between-group variance represents the differences between different process control objectives, while within-group variance represents the fluctuation of the same process control objective at different time points. The specific calculation process involves collecting time-series data of the control targets for each process. This time-series data comes from a real-time process database, and the data length must meet statistical analysis requirements, such as including at least 30 sampling points to ensure statistical significance. The mean of the control target values ​​for each process is calculated, along with the overall mean of all process control target values. The between-group variance is equal to the sum of the squares of the differences between the means of each process control target and the overall mean, divided by the between-group degrees of freedom (between the number of groups minus one). The within-group variance is equal to the sum of the squares of the differences between the control target values ​​of each process and their mean, divided by the within-group degrees of freedom (within the number of groups minus the total number of samples). The conflict index is the ratio of the between-group variance to the within-group variance, i.e., the F-statistic. A larger conflict index value indicates a higher degree of conflict between process control targets. For example, an F-statistic value greater than the critical value indicates significant conflict between process control targets. The critical value is determined using the F-distribution table based on the significance level, for example, a significance level set to 0.05.

[0046] The degree of game conflict between process control objectives is determined based on the conflict index value, and this degree of conflict is quantified using the Nash equilibrium distance. Calculating the Nash equilibrium distance requires constructing a game model of the process control objectives. This model treats each process as a player, the process control objective as the player's strategy, and the payoff function is defined based on the deviation between the actual value and the target value of the process control objective. For example, the payoff function equals the negative sum of squared deviations, where the deviation is the difference between the actual process parameter value and the process control objective value. Nash equilibrium refers to a state where no player can improve their payoff by individually changing their strategy. The Nash equilibrium distance is quantified by calculating the difference between the current strategy combination and the Nash equilibrium point. This difference is calculated using Euclidean distance, which is equal to the square root of the sum of the squares of the differences between the current values ​​of each process control objective and the Nash equilibrium value.

[0047] The Nash equilibrium is found through iterative algorithms, such as the optimal response dynamic algorithm. Starting from the current strategy, each player takes turns choosing the strategy that maximizes their own reward until the strategy combination stabilizes. The stability condition is that the strategy change is less than a set tolerance, for example, 0.001. The game conflict level is equal to the Nash equilibrium distance; a larger distance indicates a higher level of game conflict. For example, the distance ranges from 0 to positive infinity, with larger values ​​indicating more severe conflict. The quantified result of the game conflict level is used for subsequent optimization decisions; for example, an adjustment mechanism is triggered when the game conflict level exceeds a set threshold.

[0048] Establish a correspondence between key transmission paths and the degree of game conflict. This correspondence is stored as key-value pairs, where the key is the identifier of the key transmission path, generated based on the sequence of process nodes along the path (e.g., a string concatenation of crushing, grinding, and sorting processes), and the value is the degree of game conflict corresponding to that key transmission path. Alternatively, the correspondence can be implemented using a relational database table containing two fields: a key transmission path number and a game conflict degree value. The key transmission path number is consistent with the path identifier mentioned earlier. The correspondence is established based on real-time calculations, and the table is updated after each evaluation to ensure data timeliness. The correspondence is used for subsequent optimization decisions; for example, when the game conflict degree value exceeds a set threshold, an adjustment mechanism is triggered. This threshold is set based on historical conflict data, such as the 80th percentile of historical game conflict degree values. Verification of the correspondence is performed through backtracking analysis, such as checking the consistency of the correspondence in historical events; correspondences with high consistency are considered valid. Maintenance of the correspondence includes regular checks and updates to ensure its accuracy and usability; for example, the game conflict degree is recalculated monthly and the correspondence is adjusted.

[0049] S6. Based on the degree of game conflict, the local control objectives of each process are adjusted through the industrial control system to generate optimization decisions for the entire mineral processing production process indicators. The specific implementation is as follows: Based on the degree of game conflict, the local control objectives of each process are adjusted through the industrial control system to generate optimization decisions for the entire mineral processing production process. The adjustment weights of the local control objectives for each process are determined based on the degree of game conflict. These local control objective adjustment weights are positively correlated with the degree of game conflict, and this positive correlation is achieved through a linear function. For example, the linear function is that the adjustment weight of the local control objective equals the game conflict degree value of that process divided by the sum of the game conflict degree values ​​of all processes. The game conflict degree value originates from the Nash equilibrium distance value calculated in the preceding steps. The setting of the local control objective adjustment weights is based on the historical data distribution of the game conflict degree. For example, the weight coefficients are adjusted by analyzing the mean and standard deviation of historical game conflict degree values ​​to ensure that the weights reflect the relative conflict degree of each process. Historical data comes from long-term operating records of the entire mineral processing production process, covering different production batches and operating conditions. Data preprocessing includes verifying the validity and reasonableness of the game conflict degree values, such as checking whether the game conflict degree values ​​are within the range of 0 to positive infinity and removing outliers. Outlier removal adopts the three-standard-deviation principle, that is, removing data points exceeding three standard deviations of the mean.

[0050] The adjustment amount of the local control target for each process is calculated based on the weights of the local control targets. This adjustment amount is obtained by multiplying the weights by the optimization target deviation, which is the difference between the current process control target value and the optimization target value. The optimization target value is determined based on the mineral processing design specifications and historical best data. For example, the control target value that optimizes the process indicators is selected by querying a historical database. The historical database stores the process parameter values ​​of each process at its best performance. The process of obtaining the optimization target value includes data filtering and average calculation. Data filtering selects records from historical data where the process indicators reach their optimal values, such as the control target values ​​of each process when the mineral processing recovery rate is highest. Average calculation is performed on these records... The arithmetic mean of the target values ​​is used to calculate the local control target adjustment amount. The local control target adjustment amount is equal to the local control target adjustment weight multiplied by the optimization target deviation. The optimization target deviation is equal to the optimization target value minus the current process control target value. The current process control target value comes from the real-time process database. The calculation of the local control target adjustment amount takes into account the constraints between processes, such as material balance constraints or equipment capacity constraints. The constraints are expressed by inequality equations. For example, the feed particle size adjustment amount is limited by the crusher's processing capacity. After the adjustment amount is calculated, a range check needs to be performed to ensure that the adjustment amount is within the allowable range. For example, the adjustment amount does not exceed 10% of the current value. The range check is based on equipment specifications and process limitations, such as the maximum processing capacity of the crusher or the minimum load of the mill.

[0051] The industrial control system executes local control target adjustments through its control algorithms. This local control target adjustment is achieved by a programmable logic controller (PLC) and a distributed control system (DCS). The PLC executes the underlying control logic, while the DCS coordinates the control actions of each process. The control algorithm uses a proportional-integral-derivative (PID) controller. The PLD calculates the output signal based on the local control target adjustment and sends it to the actuator, such as a control valve or motor driver. Execution is achieved through industrial communication protocols, such as PROFIBUS or Modbus, to transmit the adjustment data. Control algorithm parameters are tuned according to process characteristics; for example, the proportional gain, integral time, and derivative time are determined using the Ziegler-Nichols method. The Ziegler-Nichols method obtains the critical gain and period through step response testing, calculating the proportional gain to be 0.6 times the critical gain, the integral time to be half the period, and the derivative time to be one-eighth of the period. Real-time monitoring is performed during the execution of the local control target adjustment to ensure a smooth adjustment process and avoid drastic fluctuations. Real-time monitoring is achieved through sensor feedback data, such as monitoring the rate of change of process parameters; if the rate of change exceeds a set threshold, the adjustment is paused.

[0052] The synergistic effect between the local control objectives of each process after adjustment is verified. The synergistic effect is evaluated by calculating the consistency index and efficiency index of the process parameters of each process. The consistency index is calculated using the analysis of variance method. The analysis of variance calculates the within-group variance and between-group variance of the process parameter values ​​of each process after adjustment. The within-group variance represents the fluctuation of the same process parameter at different time points, and the between-group variance represents the difference between different process parameters. The synergistic effect value is the ratio of the within-group variance to the between-group variance. The smaller the ratio, the better the synergistic effect. The efficiency index is determined by calculating the degree of improvement in the transmission of process indicator objectives. The degree of improvement is equal to the difference between the transmitted process indicator objective value after adjustment and the transmitted process indicator objective value before adjustment, divided by the transmitted process indicator objective value before adjustment. The transmitted process indicator objective value comes from the correspondence established in the previous steps. The verification process uses real-time process data, and the data acquisition frequency is synchronized with the adjustment cycle. For example, data is collected and verified every five minutes. The verification of synergistic effect also includes checking whether the conflict between processes has decreased. For example, by comparing the game conflict degree value before and after adjustment, a decrease in the game conflict degree value indicates a good synergistic effect.

[0053] Based on the verification results, optimization decisions for the entire mineral processing production process are generated. This generation process is implemented through a decision rule engine. The engine triggers different decision actions based on the synergy effect verification results. For example, if the synergy effect value is lower than a set threshold but the efficiency indicator improvement is higher than the set threshold, a fixed adjustment decision is generated. This fixed adjustment decision permanently writes the local control target adjustment amount into the control system's set value. The threshold for the synergy effect value is determined based on the historical synergy effect data distribution; for example, the threshold is set to the 70th percentile of historical synergy effect values. Historical synergy effect data comes from long-term operating records, and percentile calculation is done by sorting historical data and selecting values ​​at corresponding positions. The threshold for the efficiency indicator improvement is determined based on the historical efficiency indicator improvement data distribution; for example, the threshold is set to the historical... The 50th percentile of the efficiency indicator improvement value is used. Historical efficiency indicator improvement data comes from long-term operation records. Percentile calculation is performed by sorting historical data and selecting the value at the corresponding position. If the synergistic effect value or the efficiency indicator improvement level does not meet the standard, a readjustment decision is generated. The readjustment decision triggers a new round of weight calculation and adjustment optimization. The output of the optimization decision includes control instruction sets and report documents. The control instruction sets are issued to the controllers of each process through the industrial control system. The report documents record the adjustment process, verification results, and decision basis for subsequent analysis and auditing. The generation of optimization decisions ensures the overall optimization of process indicators throughout the entire process, such as improving ore beneficiation recovery rate or reducing energy consumption. The effect of the decision is verified through long-term monitoring, such as weekly evaluation of the trend of process indicator changes after the implementation of the decision.

[0054] Example 2: Figure 2A schematic diagram of the process index optimization decision system for the entire mineral processing production process of this invention is provided. The process index optimization decision system for the entire mineral processing production process includes: The data acquisition module is used to acquire real-time process data for multiple processes throughout the entire mineral processing production process; The hub identification module is used to determine whether there are key transmission hubs in the entire process based on real-time process data by constructing a state-space recursive graph and analyzing its complex network characteristics. The impact analysis module is used to analyze the impact of the dynamic interaction effect between processes centered on the key transmission hub on the transmission of process indicators when a key transmission hub exists. The path tracing module is used to trace the key transmission paths that cause nonlinear distortions in the execution of upstream process optimization instructions after downstream processes, based on the impact of dynamic interaction effects on the transmission of process indicator targets. The game theory evaluation module is used to assess the degree of game-theoretic conflict between the control objectives of each process based on key transmission paths. The decision generation module is used to generate optimization decisions for the entire mineral processing production process by adjusting the local control objectives of each process through the industrial control system based on the degree of game conflict.

[0055] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0056] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0057] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. Computer-readable storage media can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0058] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0059] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0060] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0061] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0062] If a function is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0063] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0064] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing process indicators throughout the entire mineral processing production process, characterized in that, include: S1. Obtain real-time process data for multiple processes throughout the entire mineral processing production process; S2. Based on real-time process data, by constructing a state-space recursive graph and analyzing its complex network characteristics, determine whether there are key transmission hubs in the entire process. S3. When a key transmission hub exists, analyze the impact of the dynamic interaction effect between processes centered on the key transmission hub on the transmission of process indicators. S4. Based on the impact of dynamic interaction effects on the transmission of process indicators, trace the key transmission path that causes nonlinear distortion of upstream process optimization instructions after execution in downstream processes. S5. Based on the key transmission path, assess the degree of game-theoretic conflict between the control objectives of each process. S6. Based on the degree of game conflict, adjust the local control objectives of each process through the industrial control system to generate optimization decisions for the entire mineral processing production process indicators.

2. The method for optimizing process indicators throughout the mineral processing production process according to claim 1, characterized in that, Acquire real-time process data for multiple steps in the entire mineral processing production process, including: Collect particle size parameters for the crushing process, concentration and load parameters for the grinding process, separation operation parameters for the sorting process, and dewatering effect parameters for the dewatering process. Particle size parameters, concentration and load parameters, sorting operation parameters, and dewatering effect parameters are combined into real-time process data; real-time process data is collected by sensors and process control systems deployed in each process step.

3. The method for optimizing process indicators throughout the entire mineral processing production process according to claim 1, characterized in that, Based on real-time process data, by constructing a state-space recursive graph and analyzing its complex network characteristics, we can determine whether there are key transmission hubs in the entire process, including: A state-space recursion graph is constructed based on real-time process data, and the density distribution of recursion points in the state-space recursion graph is calculated: real-time process data is converted into multi-dimensional state vectors through phase space reconstruction, grid partitioning is used to statistically analyze the distribution frequency of recursion points in the state space, and the kernel density estimation method is used to calculate the density distribution of recursion points. Complex network characteristics are extracted from the state-space recursive graph, and node centrality indices are calculated, including betweenness centrality and proximity centrality. Based on the numerical distribution of node centrality indicators, identify the process nodes that play a key transmission role in the entire process; Based on the number and distribution characteristics of process nodes with key transmission functions, determine whether there are key transmission hubs in the entire process.

4. The method for optimizing process indicators throughout the mineral processing production process according to claim 3, characterized in that, Based on the number and distribution characteristics of process nodes with key transmission functions, the determination of whether there is a key transmission hub in the whole process includes: setting a centrality threshold for node centrality index, counting the number of process nodes that exceed the centrality threshold, analyzing the topological position relationship of process nodes in the whole process network, and determining the existence of a key transmission hub when process nodes that exceed the centrality form a connected subgraph and are located on the critical path.

5. The method for optimizing process indicators throughout the mineral processing production process according to claim 1, characterized in that, When a key transmission hub exists, analyze the impact of the dynamic interaction effects between processes centered on the key transmission hub on the transmission of process performance targets, including: Based on the key transmission nodes containing key transmission functions, the dynamic interaction effect between process nodes is calculated. The dynamic interaction effect is quantified by the mutual information or transmission entropy of the changes in process parameters between process nodes. Analyze the correlation strength between key transmission function process nodes and adjacent process nodes; Determine the impact of dynamic interaction effects on the transmission of process indicators based on the correlation strength: Establish a mapping relationship between correlation strength and the transmission of process indicators, divide the correlation strength into level intervals, determine the impact of each level interval through statistical analysis, and calculate the final impact using a weighted average method; Establish the correspondence between key transmission function process nodes and process indicator targets.

6. The method for optimizing process indicators throughout the mineral processing production process according to claim 1, characterized in that, Based on the impact of dynamic interaction effects on the transmission of process performance targets, the key transmission paths that lead to nonlinear distortions in upstream process optimization instructions after execution in downstream processes are traced, including: Based on the impact of dynamic interaction effects on the transmission of process target indicators, the deviation between upstream process optimization instructions and downstream process execution results is identified. Calculate the degree of nonlinear distortion of the upstream process optimization instructions during the execution of the downstream process; Process connections whose nonlinear distortion exceeds a preset distortion threshold are selected based on the degree of nonlinear distortion. Analyze the path characteristics of process parameter target transmission in process connections that exceed the preset distortion threshold; The key transmission paths leading to nonlinear distortions are identified based on path characteristics.

7. The method for optimizing process indicators throughout the mineral processing production process according to claim 6, characterized in that, The calculation of the degree of nonlinear distortion includes: establishing a transfer function model of the upstream process optimization command and the downstream process execution result, calculating the gain change of the transfer function under different operating conditions, analyzing the nonlinear characteristics of the gain change, and using the standard deviation and coefficient of variation of the gain change to quantify the degree of nonlinear distortion.

8. The method for optimizing process indicators throughout the entire mineral processing production process according to claim 1, characterized in that, Based on the key transmission path, assess the degree of game-theoretic conflict between the control objectives of each process, including: Identify the process control objectives corresponding to the process nodes on the critical transmission path; Calculate the conflict index values ​​between process control objectives. The conflict index values ​​are obtained through variance analysis of process control objectives. The degree of game-related conflict between process control objectives is determined based on the conflict index value, and the degree of game-related conflict is quantified by the Nash equilibrium distance; Establish the correspondence between key transmission paths and the degree of game conflict.

9. The method for optimizing process indicators throughout the entire mineral processing production process according to claim 1, characterized in that, Based on the degree of game-theoretic conflict, the industrial control system adjusts the local control objectives of each process to generate optimization decisions for the entire mineral processing production process, including: The adjustment weights of local control objectives for each process are determined based on the degree of game conflict, and the adjustment weights of local control objectives are positively correlated with the degree of game conflict. Calculate the adjustment amount of the local control target for each process based on the adjustment weight of the local control target; The local control target adjustment is executed by the industrial control system, and the adjustment of the local control target is achieved by the control algorithm of the industrial control system. Verify the synergistic effect between the adjusted local control objectives of each process; Based on the verification results, optimization decisions are generated for the entire mineral processing production process indicators.

10. A decision-making system for optimizing process indicators throughout the mineral processing production process, used to implement the decision-making method for optimizing process indicators throughout the mineral processing production process as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire real-time process data for multiple processes throughout the entire mineral processing production process; The hub identification module is used to determine whether there are key transmission hubs in the entire process based on real-time process data by constructing a state-space recursive graph and analyzing its complex network characteristics. The impact analysis module is used to analyze the impact of the dynamic interaction effect between processes centered on the key transmission hub on the transmission of process indicators when a key transmission hub exists. The path tracing module is used to trace the key transmission paths that cause nonlinear distortions in the execution of upstream process optimization instructions after downstream processes, based on the impact of dynamic interaction effects on the transmission of process indicator targets. The game theory evaluation module is used to assess the degree of game-theoretic conflict between the control objectives of each process based on key transmission paths. The decision generation module is used to generate optimization decisions for the entire mineral processing production process by adjusting the local control objectives of each process through the industrial control system based on the degree of game conflict.

Images