Mineral processing production management method and system based on digitization

Abstract

The invention provides a beneficiation production management method and system based on digitization, is applied to the technical field of beneficiation production management, and aims to calculate physical balance parameters of key procedures based on real-time process data, predict future states of the key procedures, deduce influence trends of raw material property changes on process parameters or product indexes, and improve the beneficiation production management efficiency. According to the method and the system, the influence trend of the property change of the raw ore on the subsequent process can be predicted in a prospective manner, and a globally optimized material scheduling or production parameter adjustment instruction is generated, so that the production stability is improved, the fluctuation and the cost are reduced, and the active management capability is enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of beneficiation production management, and particularly relates to a beneficiation production management method and system based on digitalization. BACKGROUND

[0002] In large-scale metal mine beneficiation plants, the beneficiation process involves multiple continuous processes, such as crushing, grinding, flotation and filtration, etc. The existing digital system collects real-time process data, lagging material test data, equipment state data and energy consumption data through sensors and information systems, realizes the monitoring and basic analysis of the production process. However, the beneficiation production faces the challenge of inherent fluctuation of raw ore properties. Different batches or ore bins of raw ore have significant differences in physical properties such as hardness, particle size and chemical properties such as mineral content, floatability. These differences will have a chain effect on subsequent processes, for example, the decrease of grinding efficiency may lead to the coarsening of grinding product particle size, which in turn affects the flotation effect, and finally leads to the deviation of key indicators such as concentrate grade or recovery rate from the planned target.

[0003] Although the existing system can detect real-time parameter deviation and lagging test results, it lacks deep correlation analysis mechanism and cannot predict the response trend of process parameters such as grinding load, pulp concentration, flotation liquid level and reagent flow in the future short time based on historical and real-time data, and the change trend of final product indicators such as concentrate grade, recovery rate and unit energy consumption. The lack of such prediction ability leads to the fact that the system can only provide post or temporary adjustment suggestions, and the operator relies on experience to make local parameter modifications, such as increasing reagent dosage to deal with the decrease of concentrate grade, but such adjustment is often inefficient, which may exacerbate production fluctuations, increase costs or cause suboptimal control problems. At the same time, the system is difficult to dynamically optimize the proportioning and grinding order of different ore bins of raw ore, as well as the production plan execution parameters, in order to smooth the impact of raw ore fluctuation.

[0004] Therefore, in the complex multi-stage continuous production environment, the existing technology cannot prospectively infer the future state and generate globally optimized material scheduling and production parameter adjustment instructions, which limits the production stability and active management capability.

[0005] In view of the above problems, the existing technology needs to be improved. SUMMARY

[0006] In view of the above deficiencies of the prior art, the present application provides a beneficiation production management method and system based on digitalization, which can prospectively predict the influence trend of raw ore property change on subsequent processes, and generate globally optimized material scheduling or production parameter adjustment instructions, thereby improving production stability, reducing fluctuations and costs, and enhancing active management capability.

[0007] In a first aspect, a digitalized beneficiation production management method is provided, which comprises: S1: calculating a physical balance parameter of at least one key process based on real-time process data of a beneficiation production process; S2: determining a variation trend of the physical balance parameter within a preset time window, and inferring a predicted state of the physical balance parameter in a future time period according to the variation trend; S3: inferring an influence trend of changes in raw material properties on process parameters of subsequent processes or final product indicators based on the predicted state and in combination with a preset beneficiation process model; S4: generating a material scheduling instruction for adjusting the proportion of raw materials from different sources, or a production parameter adjustment instruction for adjusting process parameters of the beneficiation production process, according to the influence trend.

[0008] Further, step S1 comprises: S11: obtaining periodically generated lagging material attribute data with time information in the beneficiation production process, and historical process parameters corresponding to the time information; S12: establishing a mapping relationship between the historical process parameters and the lagging material attribute data based on the lagging material attribute data and the historical process parameters; S13: calculating and generating a real-time material attribute proxy value based on the mapping relationship and the real-time process data; S14: calculating the physical balance parameter of the key process based on the material attribute proxy value and the real-time process data.

[0009] Further, step S12 comprises: S121: for each data pair composed of the lagging material attribute data and the historical process parameters, determining a weight value that decays over time according to the time information corresponding to the data pair; S122: establishing the mapping relationship based on the data pairs and the weight values corresponding to each of the data pairs.

[0010] Further, step S2 comprises: S21: obtaining an event time of a planned operation change occurring within the preset time window; S22: locating a data sub-window that is not divided by the event time from data containing the physical balance parameter within the preset time window according to the event time; S23: determining the variation trend based on the physical balance parameter in the data sub-window; S24: According to the determined change trend, the predicted state of the physical balance parameter in a future time period is calculated.

[0011] Further, step S22 comprises: S221: The event type of the planned operation change is acquired; S222: According to the event type, the corresponding process transition time is determined; S223: From the data containing the physical balance parameter in the preset time window, the data covered by the event time and the time range determined by the process transition time is excluded to obtain the remaining data; S224: In the remaining data, the data sub-window not divided by the event time is located.

[0012] Further, step S3 comprises: S31: For the subsequent process, a plurality of key processes that affect the subsequent process are determined; S32: The predicted state of each of the plurality of key processes is acquired to form a predicted state combination; S33: In the preset beneficiation process model, a preset influence relationship matching the predicted state combination is searched for; S34: According to the searched preset influence relationship, the net influence trend of the subsequent process is determined, and the net influence trend is taken as the influence trend of the change of the raw material properties on the subsequent process.

[0013] Further, step S33 comprises: S331: The similarity between the predicted state combination and each preset state combination stored in the beneficiation process model is calculated to obtain the similarity result of each preset state combination; S332: According to the similarity result, the preset state combination with the highest similarity to the predicted state combination is determined; S333: The preset influence relationship corresponding to the preset state combination with the highest similarity is taken as the preset influence relationship matching the predicted state combination searched for.

[0014] Further, step S4 comprises: S41: The influence trend is analyzed to determine its quantitative characteristic attribute; S42: A first selection condition corresponding to the material scheduling instruction and a second selection condition corresponding to the production parameter adjustment instruction are preset, and the first selection condition and the second selection condition are both associated with a preset threshold value of the quantitative characteristic attribute; S43: comparing the quantified characteristic attribute with the first selection condition and the second selection condition to determine a type of instruction to be generated; S44: generating a corresponding material scheduling instruction for adjusting the proportion of raw materials from different sources or a production parameter adjustment instruction for adjusting process parameters of the beneficiation production process according to the type of instruction.

[0015] Further, step S42 includes: S421: identifying a historical influence trend in the historical data, an executed adjustment instruction corresponding to the historical influence trend, and a control effect of the executed adjustment instruction to form a historical data record; S422: statistically analyzing the quantified characteristic attribute of the historical influence trend, the type of the executed adjustment instruction, and the control effect based on the historical data record to establish a correlation between the numerical value of the quantified characteristic attribute and the control effect corresponding to different types of adjustment instructions; S423: determining the preset threshold of the quantified characteristic attribute according to the correlation, and establishing the first selection condition and the second selection condition based on the preset threshold.

[0016] In a second aspect, a digital beneficiation production management system is used to implement any of the above methods, and the system includes: a calculation module configured to calculate a physical balance parameter of at least one key process based on real-time process data of the beneficiation production process; a first prediction module configured to determine a change trend of the physical balance parameter within a preset time window based on the physical balance parameter, and predict a predicted state of the physical balance parameter in a future time period according to the change trend; a second prediction module configured to infer an influence trend of changes in raw material properties on process parameters of subsequent processes or final product indicators based on the predicted state and in combination with a preset beneficiation process model; a scheduling module configured to generate a material scheduling instruction for adjusting the proportion of raw materials from different sources or a production parameter adjustment instruction for adjusting process parameters of the beneficiation production process according to the influence trend.

[0017] Beneficial effects: The digital beneficiation production management method and system can predict the influence trend of changes in raw material properties on subsequent processes in a forward-looking manner, and generate globally optimized material scheduling or production parameter adjustment instructions, thereby improving production stability, reducing fluctuations and costs, and enhancing proactive management capabilities. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 A flow chart of a digital-based beneficiation production management method proposed in the present application.

[0019] Figure 2 A structure diagram of a digital-based beneficiation production management system proposed in the present application.

[0020] Figure 3 A framework diagram of a digital-based beneficiation production management system proposed in the present application.

[0021] Label explanation: 201, calculation module; 202, first prediction module; 203, second prediction module; 204, scheduling module. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. The components of the embodiments of the present application described and indicated in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0023] It should be noted that: similar labels and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first, second" and the like are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0024] In a first aspect, a digital-based beneficiation production management method, the method comprising: S1: calculating a physical balance parameter of at least one key process based on real-time process data of a beneficiation production process; S2: determining a change trend of the physical balance parameter within a preset time window, and inferring a predicted state of the physical balance parameter in a future time period according to the change trend; S3: inferring an influence trend of changes in raw material properties on process parameters of subsequent processes or final product indicators based on the predicted state and in combination with a preset beneficiation process model; S4: generating a material scheduling instruction for adjusting the proportion of raw materials from different sources, or a production parameter adjustment instruction for adjusting process parameters of the beneficiation production process, according to the influence trend.

[0025] wherein the physical balance parameter refers to a quantitative index reflecting the material or energy balance of a process. The preset time window refers to a time period for analyzing data changes. The predicted state refers to an estimated value of the physical balance parameter at a future time point. The beneficiation process model refers to a mathematical model simulating the beneficiation process. The influence trend refers to the predicted influence direction of the raw material property change on subsequent parameters or indexes. The material scheduling instruction refers to a command for adjusting the raw material ratio. The production parameter adjustment instruction refers to an instruction for modifying the process parameters, which can be implemented by a process control system command, for stabilizing the process operation.

[0026] Specifically, the method solves the raw ore fluctuation problem through a closed-loop mechanism. The physical balance parameter of a key process is calculated based on real-time process data of the beneficiation production process. Real-time data provide immediate state capture, ensuring parameter accuracy. The change trend of the physical balance parameter within a preset time window is determined, and the future predicted state is calculated. Trend analysis reveals dynamic laws, supporting forward-looking prediction. Based on the predicted state and the preset beneficiation process model, the influence trend of the raw material property change is inferred. The model simulates the process chain response and infers the changes of process parameters or product indexes. According to the influence trend, material scheduling instructions or production parameter adjustment instructions are generated. The instructions are directed to the influence direction, realizing dynamic adjustment to optimize production.

[0027] Further, step S1 includes: S11: obtaining periodically generated lagging material attribute data with time information in the beneficiation production process, and historical process parameters corresponding to the time information; S12: based on the lagging material attribute data and the historical process parameters, establishing a mapping relationship between the historical process parameters and the lagging material attribute data; S13: based on the mapping relationship and real-time process data, calculating and generating real-time material attribute proxy values; S14: based on the material attribute proxy values and real-time process data, calculating the physical balance parameter of a key process.

[0028] wherein the lagging material attribute data refers to periodically generated material attribute data.

[0029] The historical process parameters refer to historical process parameters corresponding to the time information, which can be implemented by real-time process data collected by sensors, for mapping with lagging data.

[0030] The mapping relationship refers to the association model between the historical process parameters and the lagging material attribute data, which is used to predict real-time material attributes.

[0031] The real-time process data refers to the currently collected production process data.

[0032] The material property proxy value refers to a real-time estimated material property value.

[0033] The physical balance parameter refers to a balance state parameter of a key process, which can be obtained by calculating a material balance equation and is used to evaluate a current process state.

[0034] Specifically, time-related information is collected by obtaining lagging material property data with time information and corresponding historical process parameters, laying a data foundation for establishing an accurate mapping; a pattern in historical data is learned based on the mapping relationship between the lagging material property data and the historical process parameters, and a delay problem caused by relying on lagging data is avoided; real-time material property proxy values are calculated and generated by using the mapping relationship and real-time process data, the periodicity limitation of lagging data is overcome, and approximate real-time material property estimation is provided; finally, physical balance parameters are calculated based on the material property proxy values and real-time process data, real-time and estimated data are integrated, and immediate evaluation of a key process state is realized, supporting subsequent trend prediction and adjustment decisions.

[0035] Further, step S12 comprises: S121: For each data pair composed of lagging material property data and historical process parameters, a weight value decaying over time is determined according to time information corresponding to the data pair; S122: A mapping relationship is established based on the data pair and the weight value corresponding to each data pair.

[0036] Specifically, in the mapping establishment process, each data pair is given a decaying weight value according to its time information, which ensures that newer data obtains a higher weight, reflecting its closer proximity to real-time production conditions; further, the data pair with the weight value is used to construct a mapping model, so that the model automatically prioritizes high-timeliness data, thereby strengthening the mapping's ability to capture dynamic production changes and solving the deviation problem caused by the difference in data timeliness.

[0037] The weight value decaying over time is an exponential weight factor that decreases over time, which can be calculated using an exponential decay function, for example, weight value = e^{-k(t_current - t_data)}, where k can be a decay coefficient, t_current is the current time, and t_data is the data time, which gives newer data more influence and avoids outdated data interfering with mapping accuracy.

[0038] Further, step S2 comprises: S21: Obtain an event time of a planned operation change occurring within a preset time window; S22: According to the event time, locate a data sub-window that is not divided by the event time from data containing the physical balance parameter within the preset time window. S23: determining a change trend based on the physical balance parameter within the data sub-window; S24: predicting a state of the physical balance parameter in a future time period according to the determined change trend.

[0039] The event type of the planned operation change refers to a pre-arranged production adjustment event category, which can be implemented by an event classification system, such as a maintenance event or a parameter adjustment event, for distinguishing the disturbance characteristics of different changes on the production process and avoiding processing bias.

[0040] The process transition time refers to a time period required for the production to recover to stability after the operation change, which can be determined according to historical data statistical analysis, such as setting different time lengths for different event types, for dynamically adapting the disturbance period and preventing analysis errors caused by fixed thresholds.

[0041] The preset time window refers to a time interval for analyzing the change trend of the physical balance parameter, which can be a user-defined time period, such as the past one hour, for providing a basic data source.

[0042] The data of the physical balance parameter refers to a parameter value sequence reflecting the production balance state, which can be a grinding load or a pulp concentration, for subsequent trend calculation.

[0043] The event time refers to a specific time when the planned operation change occurs, which can be obtained from system logs for locating the disturbance starting point. The time range refers to a data exclusion interval determined by the event time and the process transition time, which can be a time period, such as a few minutes before and after the event time, for removing the disturbed data segment.

[0044] The remaining data refers to a physical balance parameter data set retained after the disturbance, which is a continuous data point sequence, for ensuring analysis reliability.

[0045] The data sub-window not divided by the event time refers to a continuous data segment in the remaining data that is not interrupted by the event, which can be multiple time segments, such as data blocks during stable production, for providing a complete data basis and facilitating the extraction of information representing stable state.

[0046] Specifically, the scheme identifies the characteristics of different operation changes by obtaining the event type, which is used for targeted processing; further, the process transition time is determined according to the event type, and the interference period is dynamically adapted; thus, the data covered by the event time and the time range determined by the process transition time is excluded from the preset time window, and the disturbed part is directly removed; the data sub-window not segmented by the event time is located in the remaining data, ensuring data continuity and integrity; the whole process is based on stable state data for change trend analysis, avoiding the transition period interference of planned operation changes, and improving the accuracy of trend prediction.

[0047] Further, step S22 comprises: S221: obtaining the event type of the planned operation change; S222: determining the corresponding process transition time according to the event type; S223: excluding the data covered by the event time and the time range determined by the process transition time from the data containing the physical balance parameter in the preset time window, to obtain the remaining data; S224: locating the data sub-window not segmented by the event time in the remaining data.

[0048] Specifically, the scheme identifies the characteristics of different operation changes by obtaining the event type, which is used for targeted processing; further, the process transition time is determined according to the event type, and the interference period is dynamically adapted; thus, the data covered by the event time and the time range determined by the process transition time is excluded from the preset time window, and the disturbed part is directly removed; the data sub-window not segmented by the event time is located in the remaining data, ensuring data continuity and integrity; the whole process is based on stable state data for change trend analysis, avoiding the transition period interference of planned operation changes, and improving the accuracy of trend prediction.

[0049] Further, step S3 comprises: S31: determining a plurality of key processes that have an impact on the subsequent process for the subsequent process; S32: obtaining the predicted state of each of the plurality of key processes to form a predicted state combination; S33: searching for a preset influence relationship matching the predicted state combination in a preset beneficiation process model; S34: determining the net influence trend of the subsequent process according to the found preset influence relationship, and taking the net influence trend as the influence trend of the change of raw material properties on the subsequent process.

[0050] Specifically, a plurality of key processes are determined for subsequent processes in the beneficiation process, avoiding the limitation of a single process perspective; the respective predicted states of the key processes are obtained to form a predicted state combination, and the synergistic effect of multi-source state information is utilized; in the preset beneficiation process model, a preset influence relationship matching the predicted state combination is searched, and the rules library preset by the model is used to ensure that the inference is based on the associated logic verified by history; according to the preset influence relationship found, the net influence trend of the subsequent process is determined, and by integrating multiple influence sources, interference factors are excluded, and more reliable trend results are output.

[0051] Further, step S33 comprises: S331: Calculate the similarity between the predicted state combination and each preset state combination stored in the beneficiation process model, to obtain the similarity results of each preset state combination; S332: According to the similarity results, determine the preset state combination with the highest similarity to the predicted state combination; S333: The preset influence relationship corresponding to the preset state combination with the highest similarity is taken as the preset influence relationship found matching the predicted state combination.

[0052] Wherein, the predicted state combination refers to the set formed by the predicted states of the plurality of key processes.

[0053] The preset state combination refers to a known state combination stored in the beneficiation process model in advance.

[0054] The similarity refers to the proximity quantitative index between the predicted state combination and the preset state combination, which can be realized by using Euclidean distance, cosine similarity or Manhattan distance algorithm. The introduction of this index makes the matching process have objective quantitative basis, and eliminates the subjectivity and randomness of artificial experience judgment.

[0055] The preset influence relationship refers to the causal mapping between the preset state combination and the change trend of the subsequent process parameters or product indicators, which can be established by data mining or process mechanism modeling and stored in the beneficiation process model.

[0056] Specifically, when the beneficiation process model performs a matching operation, firstly, similarity calculation is performed on the predicted state combination and each preset state combination, wherein the similarity algorithm can be Euclidean distance method, and quantitative evaluation is realized by calculating the sum of squares of differences of each dimension parameter. Thus, a similarity result set of all preset combinations is generated. Further, through sorting or maximum value retrieval operation, the preset state combination with the highest similarity to the predicted state combination is identified from the result set. Finally, the preset influence relationship associated with the preset state combination is directly called as the basis for inferring the influence trend of the change of raw material properties on subsequent processes. The whole process replaces subjective selection with objective calculation, ensuring that the matching result is consistent with the actual response characteristics of the process.

[0057] Further, step S4 includes: S41: analyzing the influence trend to determine its quantitative characteristic attribute; S42: presetting a first selection condition corresponding to the material scheduling instruction and a second selection condition corresponding to the production parameter adjustment instruction, both the first selection condition and the second selection condition being associated with a preset threshold value of the quantitative characteristic attribute; S43: comparing the quantitative characteristic attribute with the first selection condition and the second selection condition to determine the instruction type of the to-be-generated instruction; S44: generating a corresponding material scheduling instruction for adjusting the proportion of raw materials of different sources or a production parameter adjustment instruction for adjusting the process parameters of the beneficiation production process according to the instruction type.

[0058] The quantitative characteristic attribute refers to the process of converting the influence trend into a measurable numerical attribute, which can be specifically realized by calculating the trend amplitude or duration, which helps to avoid misjudgment caused by subjective experience and ensures the objectivity of decision-making.

[0059] The first selection condition refers to a preset condition for triggering the material scheduling instruction, which can be specifically a threshold value related to the quantitative characteristic attribute, such as a trend change rate exceeding a certain value, which provides a basis for selecting the material scheduling and solves the problem of arbitrary condition setting.

[0060] The second selection condition refers to a preset condition for triggering the production parameter adjustment instruction, which can be specifically another threshold value related to the quantitative characteristic attribute, such as a trend duration reaching a certain length, which provides a basis for selecting the production parameter adjustment and ensures the pertinence of the instruction.

[0061] The preset threshold value refers to a pre-set numerical limit, which can be specifically a value obtained based on historical data statistical analysis, which ensures the objectivity of decision-making and avoids the blindness of reactive adjustment.

[0062] Specifically, the impact trend is first analyzed to determine a quantitative characteristic attribute, such as by calculating the slope or range of the trend to achieve the numerical value of the abstract trend. Further, a first selection condition and a second selection condition are preset, which are associated with preset threshold values of the quantitative characteristic attribute, wherein the threshold values can be values obtained based on statistical analysis of historical data. Thus, the quantitative characteristic attribute is compared with the first selection condition and the second selection condition to determine the instruction type of the to-be-generated instruction. Finally, a specific material scheduling instruction or production parameter adjustment instruction is generated according to the instruction type. Through the coherent process of quantification, presetting, comparison, and generation, the entire mechanism realizes closed-loop optimization from trend analysis to instruction generation, and solves the problem of intelligent selection of adjustment instruction types based on impact trends.

[0063] Further, step S42 includes: S421: identifying historical impact trends in the historical data, executed adjustment instructions corresponding to the historical impact trends, and control effects of the executed adjustment instructions, to form a historical data record; S422: based on the historical data record, statistically analyzing the quantitative characteristic attributes of the historical impact trends, the types of the executed adjustment instructions, and the control effects, to establish an association between the numerical values of the quantitative characteristic attributes and the control effects corresponding to different adjustment instruction types; S423: determining preset threshold values of the quantitative characteristic attributes according to the association, and establishing the first selection condition and the second selection condition based on the preset threshold values.

[0064] Specifically, the historical impact trends in the historical data, the executed adjustment instructions, and the control effects thereof are identified and form a historical data record, focusing on actual production adjustment behaviors; based on the record, the quantitative characteristic attributes, the types of the adjustment instructions, and the control effects are statistically analyzed to establish an association between attribute values and instruction effects; according to the association, the preset threshold values are determined and used to establish the selection conditions, so that the instruction generation is based on historical actual effects, thereby solving the problem of lacking historical experience to intelligently determine the selection conditions.

[0065] In some specific embodiments, the statistical analysis can employ a linear regression model, the quantitative characteristic attribute can be a change rate of the impact trend; the preset threshold values can be numerical limits set based on optimal values of the control effects; the first selection condition can be to generate a material scheduling instruction when the change rate exceeds the threshold value; and the second selection condition can be to generate a production parameter adjustment instruction when the change rate is within the threshold value range.

[0066] Please refer to Figure 2 、 Figure 3 A digital-based beneficiation production management system for implementing any of the above methods, the system comprising: The computing module 201 calculates the physical balance parameters of at least one key process based on real-time process data of the beneficiation production process. The first prediction module 202 determines the change trend of the physical balance parameters within a preset time window based on the physical balance parameters, and calculates the predicted state of the physical balance parameters in a future time period according to the change trend. The second prediction module 203 infers the influence trend of changes in raw material properties on process parameters or final product indicators of subsequent processes based on the predicted state and a preset beneficiation process model. The scheduling module 204 generates material scheduling instructions for adjusting the proportion of raw materials from different sources, or production parameter adjustment instructions for adjusting the process parameters of the beneficiation production process, according to the influence trend.

[0067] The computing module 201 is a component for processing real-time collected data in the beneficiation production process, which can be implemented by a data processing unit, for example, by receiving sensor data in real time and executing a calculation algorithm, which provides a quantitative basis for the current production state and directly reflects the immediate impact of changes in raw ore properties on key processes, thereby laying a data foundation for subsequent prediction.

[0068] The first prediction module 202 is a component for analyzing short-term changes in parameters, which can be implemented by a time series analysis algorithm, such as a sliding window statistical method, to realize forward-looking prediction by identifying the dynamic change trend of the physical balance parameters, ensuring that the system can capture potential fluctuations in advance.

[0069] The second prediction module 203 is a component for simulating cascading effects, which can be implemented by a beneficiation process model, such as a simulation engine based on historical data and process rules, which uses model knowledge to convert the predicted state into a specific influence trend, solving the problem of insufficient data correlation.

[0070] The scheduling module 204 is a component for outputting adjustment instructions, which can be implemented by an instruction generation unit, such as a decision system based on a rule engine, which dynamically generates operation instructions based on the influence trend to achieve active optimization of production parameters.

[0071] Specifically, the system realizes early response to raw ore property fluctuation and active optimization of production process through modular design. The calculation module calculates the physical balance parameters of key processes based on real-time process data, and the real-time data directly reflects the immediate impact of raw ore changes on the process, solving the problem of lack of real-time correlation analysis in the background technology. The first prediction module determines the trend of changes in the physical balance parameters within a preset time window and calculates the future predicted state based on the physical balance parameters, and the preset time window can be several hours, for example, through analysis of short-term data fluctuations to realize forward-looking prediction, so that the system can identify potential fluctuations in advance. The second prediction module infers the influence trend of raw material property changes on subsequent processes based on the predicted state combined with a preset beneficiation process model, and the beneficiation process model can be a model trained based on historical data, for example, by matching the predicted state combination to infer the chain effect, solving the problem that the background technology cannot deeply correlate data. The scheduling module generates material scheduling instructions or production parameter adjustment instructions according to the influence trend, and the material scheduling instructions can be instructions for adjusting the proportion of the bin, for example, dynamically optimizing the grinding sequence according to the influence trend to realize active execution of the production plan.

[0072] Through the above technical solutions, the present application solves the uncertainty problem caused by raw ore property fluctuation, and improves the stability of production plan execution and the initiative of process management, wherein the stability is reflected in the reduction of fluctuation of key process parameters, and the initiative is reflected in that the instruction generation is based on prediction results rather than lagging data.

[0073] In this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any such actual relationship or order between such entities or actions.

[0074] The above merely illustrates the embodiments of the present application and is not used to limit the protection scope of the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A digital-based mineral processing production management method, characterized in that: The method comprises: S1: Calculate the physical balance parameters of at least one key process based on the real-time process data of the mineral processing production process; S2: Based on the physical balance parameter, determining its change trend within a preset time window, and calculating the predicted state of the physical balance parameter in a future time period based on the change trend; S3: Based on the predicted state and in combination with a preset mineral processing process model, infer the impact trend of the change in raw material properties on the process parameters of subsequent processes or the final product indicators; S4: Based on the influencing trend, a material scheduling instruction for adjusting the proportion of raw materials from different sources, or a production parameter adjustment instruction for adjusting the process parameters of the mineral processing production process is generated.

2. A digital-based mineral processing production management method according to claim 1, characterized in that: Step S1 includes: S11: Acquire lagged material attribute data with time information that is periodically generated during the mineral processing production process, and historical process parameters corresponding to the time information; S12: Based on the lagged material attribute data and the historical process parameters, establishing a mapping relationship between the historical process parameters and the lagged material attribute data; S13: Calculating and generating a real-time material attribute proxy value based on the mapping relationship and the real-time process data; S14: Calculating the physical balance parameters of the key process based on the material attribute proxy value and the real-time process data.

3. A digital-based mineral processing production management method according to claim 2, characterized in that: Step S12 includes: S121: for each data pair consisting of the lagged material attribute data and the historical process parameters, determine, based on time information corresponding to the data pair: a weight value that decays over time; S122: Establishing the mapping relationship based on the data pairs and the weight values ​​corresponding to the data pairs.

4. A digital-based mineral processing production management method according to claim 1, characterized in that: Step S2 includes: S21: Obtaining the event time of the planned operation change occurring within the preset time window; S22: Locating, based on the event time, a data sub-window that is not divided by the event time from the data containing the physical balance parameter within the preset time window; S23: determining the change trend based on the physical balance parameter in the data sub-window; S24: Calculating a predicted state of the physical balance parameter in a future time period based on the determined change trend.

5. A digital-based mineral processing production management method according to claim 4, characterized in that: Step S22 includes: S221: Obtaining the event type of the planned operation change; S222: Determine a corresponding process transition time according to the event type; S223: Excluding data covered by the event time and the time range determined by the process transition time from the data containing the physical balance parameter within the preset time window to obtain remaining data; S224: Locate the data sub-windows not divided by the event time in the remaining data.

6. A digital-based mineral processing production management method according to claim 1, characterized in that: Step S3 includes: S31: for subsequent processes, determining multiple key processes that have an impact on the subsequent processes; S32: Obtain the predicted status of each of the plurality of key processes to form a predicted status combination; S33: searching for a preset influence relationship matching the predicted state combination in the preset mineral processing process model; S34: Determine the net impact trend of the subsequent process based on the found preset impact relationship, and use the net impact trend as the impact trend of the change in raw material properties on the subsequent process.

7. A digital-based mineral processing production management method according to claim 6, characterized in that: Step S33 includes: S331: Calculating the similarity between the predicted state combination and each preset state combination stored in the mineral processing process model to obtain a similarity result for each preset state combination; S332: Determine, based on the similarity result, a preset state combination having the highest similarity with the predicted state combination; S333: The preset influence relationship corresponding to the preset state combination with the highest similarity is used as the found preset influence relationship matching the predicted state combination.

8. A digital-based mineral processing production management method according to claim 1, characterized in that: Step S4 includes: S41: Analyze the impact trend and determine its quantitative characteristic attributes; S42: Presetting a first selection condition corresponding to the material scheduling instruction and a second selection condition corresponding to the production parameter adjustment instruction, wherein the first selection condition and the second selection condition are both associated with a preset threshold value of the quantitative feature attribute; S43: Compare the quantitative feature attribute with the first selection condition and the second selection condition to determine the instruction type of the instruction to be generated; S44: generating, based on the instruction type, corresponding material scheduling instructions for adjusting the proportion of raw materials from different sources, or production parameter adjustment instructions for adjusting process parameters of the mineral processing production process.

9. A digital-based mineral processing production management method according to claim 8, characterized in that: Step S42 includes: S421: Identify historical impact trends in historical data, executed adjustment instructions corresponding to the historical impact trends, and control effects of the executed adjustment instructions to form historical data records; S422: Based on the historical data records, statistically analyzing the quantitative characteristic attributes of the historical impact trend, the types of the executed adjustment instructions, and the control effects, to establish a correlation between the values ​​of the quantitative characteristic attributes and the control effects corresponding to different adjustment instruction types; S423: Determine the preset threshold of the quantitative feature attribute according to the association relationship, and establish the first selection condition and the second selection condition based on the preset threshold.

10. A digital-based mineral processing production management system, characterized in that: For implementing the method according to any one of claims 1 to 9, the system comprises: Calculation module: Calculates the physical balance parameters of at least one key process based on real-time process data of the mineral processing production process; A first prediction module: determining a change trend of the physical balance parameter within a preset time window based on the physical balance parameter, and calculating a predicted state of the physical balance parameter in a future time period based on the change trend; The second prediction module: based on the predicted status and combined with the preset mineral processing process model, infers the impact trend of the change in raw material properties on the process parameters of subsequent processes or the final product indicators; Scheduling module: Based on the influencing trend, it generates material scheduling instructions for adjusting the proportion of raw materials from different sources, or production parameter adjustment instructions for adjusting the process parameters of the mineral processing production process.

Images