Intelligent load regulation system and method for ore dressing plant stage crushing-dry separation linkage

Abstract

This invention belongs to the field of industrial control technology. It discloses an intelligent load control system and method for stage crushing-dry separation linkage in a mineral processing plant. The system includes: collecting equipment operating status information and performing data cleaning; a status identification module that performs real-time change analysis, simultaneously executing fluctuation statistics and numerical difference balance calculations to classify operating states; performing trend extension calculations and outputting material time delays; predicting load fluctuations to obtain predictive control intent commands; performing rhythm comparisons and outputting optimized control intent commands; adjusting sorting logic to obtain control beats; executing time offsets to generate a synchronous control execution list; dynamically adjusting equipment operating status; receiving feedback information and transmitting it back to a preset mineral processing control terminal; and improving the intelligent perception capability of the entire load control process and the flexibility of multi-stage collaborative control.

Description

Technical Field

[0001] This invention relates to the field of industrial control technology, and more specifically, to an intelligent load control system and method for stage crushing-dry separation linkage in mineral processing plants. Background Technology

[0002] Crushing and dry separation of minerals are important stages and steps in modern mineral processing plant production lines. Load regulation of this process is used to coordinate the operation of equipment in multiple stages, ensuring balanced production capacity and optimized energy consumption. Therefore, load regulation methods are widely used and of great importance in mineral processing production lines. However, traditional load regulation methods still face many technical bottlenecks when dealing with more complex mineral processing production line operation scenarios.

[0003] In actual mineral processing production line operation, buffer silos exist between the crushing and dry separation sections. The material level in these silos often undergoes dynamic changes. For example, the same material level may correspond to two completely different scenarios: a slow descent or violent fluctuations. However, traditional load control methods rely solely on the current material level, neglecting second-order dynamic information such as the rate of change and fluctuation frequency. This makes it difficult to accurately determine the stability of the material flow, easily leading to feed mismatch in the crusher or uneven load on the dry separation equipment. Furthermore, both the crusher and dry separation equipment may experience various abnormal states, such as long-term stable low-load operation and intermittent under-feed operation. However, traditional load control methods often employ a uniform low-load response strategy, failing to accurately identify multiple states, frequently resulting in misjudgments that lead to the crusher operating under no-load conditions or wasting power in the dry separation equipment. Simultaneously, material flow occurs during the material transfer process from the coarse, medium, and fine crushers to the dry separation section. The traditional load control methods, which rely heavily on real-time feedback data for decision-making, lack the ability to predict future load trends. This can lead to issues such as untimely feed replenishment in crushing or load shifts due to future overload in the dry separation section. Furthermore, mineral processing production lines exhibit periodic load variations, and traditional load control methods, overly dependent on real-time decisions, lack historical learning capabilities. They cannot optimize the crushing frequency and dry separation parameters for typical load patterns, resulting in a passive response between crushing and dry separation. Moreover, during the coordinated control of equipment in the crushing and dry separation stages, communication channels between different devices exhibit varying response delays. Traditional load control methods use a uniform start-up time configuration for multiple stages without compensating for time shifts in the actual response differences of each device. This leads to misaligned control rhythms during coordinated control, impacting the overall operating efficiency of the mineral processing production line.

[0004] In view of this, the present invention proposes an intelligent load control system and method for stage crushing-dry separation in a mineral processing plant to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a smart load control method for stage crushing-dry separation in a mineral processing plant, comprising:

[0006] S1. Real-time acquisition of equipment operating status information in the crushing section, dry separation section, and intermediate processes, and data cleaning of the equipment operating status information to obtain equipment operating dataset;

[0007] S2. Extract dynamic indicators from the equipment operation dataset and perform real-time change analysis. At the same time, perform fluctuation statistics and numerical difference balance calculation on the equipment operation dataset, analyze the load correlation and operation rhythm matching degree between the crushing section and the dry separation section, classify the operation status based on the analysis results and statistical calculation results, and output the status classification results.

[0008] S3. Perform trend extension calculation on the state classification results and output the material time delay to the dry separation section; identify the load trend of the dry separation section, combine the material time delay to predict the load fluctuation, and obtain the predictive control intention command;

[0009] S4. Acquire historical production operation data and identify load pattern characteristics. Based on the load pattern characteristics, perform rhythm comparison on the equipment operation dataset and output optimization control intention commands.

[0010] S5. Combine the optimized control intent instructions and the predicted control intent instructions to adjust the sorting logic and obtain the control cycle; offset the execution time of the control cycle and generate a synchronous control execution list;

[0011] S6. Extract control instructions from the synchronous control execution list and distribute them one by one, dynamically adjust the equipment operating status to achieve load coordination balance between the stage crushing section and the dry separation section, receive feedback information and transmit the feedback information back to the preset mineral processing control terminal.

[0012] Furthermore, the methods for performing real-time change analysis include:

[0013] The raw height sequence is formed by continuously collecting material level height data from the buffer silo between the crushing and dry separation stages in the equipment operation data during the preset sampling period, which is the dynamic indicator.

[0014] Calculate the material level height difference between adjacent moments in the original height sequence, and calculate the ratio of this material level height difference to the unit time interval to obtain the material level change rate. Calculate the average material level change rate for each moment in the corresponding time segment of the original height sequence to obtain the average change rate.

[0015] The sampling points where the direction of material level change reverses in the original height sequence are counted, and the frequency of direction change is output. At the same time, the original height sequence is linearly fitted, the linear fitting curve is plotted, and the slope of the linear fitting curve is calculated.

[0016] The standard deviation and extreme value difference of the material level change rate at each moment in the original height sequence are calculated and combined with the average change rate, the frequency of directional change, and the slope of the linear fitting curve to form a disturbance degree index. Each component of the disturbance degree index is compared with the preset control rules, and the disturbance level of the corresponding component is output. The disturbance levels of all components are weighted and summed to obtain the stability score of the corresponding buffer silo.

[0017] Furthermore, the method for performing fluctuation statistics and numerical difference balance calculations includes:

[0018] Extract the continuously recorded equipment load values ​​from the equipment operation data set within the preset sampling period of the crushing section, screening section and dry separation section respectively, and construct the load value sequence of the corresponding stage respectively;

[0019] Calculate the load standard deviation of the load value sequence for each stage, and simultaneously calculate the dispersion coefficient of the crushed particle size distribution and the fluctuation coefficient of the dry separation grade. Sort the load standard deviations according to the preset sampling period to obtain the load fluctuation sequence.

[0020] Identify the load standard deviations belonging to the same device number within adjacent preset sampling periods in the load fluctuation sequence and calculate the difference to obtain the fluctuation change amplitude, and construct the fluctuation change amplitude sequence for the corresponding stage.

[0021] A sliding window is constructed to traverse the sequence of fluctuation amplitudes, and the first derivative of each sliding window is calculated. The time intervals corresponding to consecutive sliding windows with the same first derivative sign are selected, and the number of monotonically rising segments and monotonically falling segments and their average duration are counted and combined into a fluctuation trend indicator.

[0022] Extract time segments within a sliding window whose fluctuation amplitude is lower than a preset lower limit and whose average duration is higher than a preset low amplitude duration threshold, and mark them as stable load decline segments.

[0023] The abrupt change percentage is obtained by calculating the ratio of the time period during which the fluctuation amplitude exceeds the preset upper limit to the time period during which the stable load decreases.

[0024] By integrating fluctuation trend indicators, abrupt change ratios, parameters corresponding to stable load decline segments, as well as dispersion coefficients and fluctuation coefficients, and performing normalization processing, a fluctuation feature vector is constructed.

[0025] Furthermore, the methods for classifying operating states include:

[0026] The judgment conditions are set as follows: if the stability score is higher than the stability score threshold, it is recorded as the material level status is stable; if the mutation rate is lower than the mutation ratio threshold, it is recorded as the disturbance level is low; if the duration of the stable load decrease section is higher than the preset low amplitude duration threshold, it is recorded as the continuous trend is stable; if the number of monotonic change sections is less than the number threshold, it is recorded as the change trend is stable.

[0027] If any three of the judgment conditions are met, the operating status is determined to be a stable low-load state; if the stability score is lower than the stability score threshold, the mutation rate is higher than the mutation ratio threshold, and the number of monotonic change segments is higher than the quantity threshold, the operating status is determined to be an intermittent material shortage state.

[0028] If the dispersion coefficient is greater than the particle size dispersion threshold but the fluctuation coefficient is less than the grade fluctuation threshold, it is determined that the crushing particle size control is inappropriate; if the dispersion coefficient is less than the particle size dispersion threshold but the fluctuation coefficient is greater than the grade fluctuation threshold, it is determined that the dry separation parameters are mismatched.

[0029] Furthermore, the method for performing trend extension calculation includes:

[0030] Extract the status classification labels corresponding to the status classification results, obtain the corresponding buffer silo material level height, and associate them with the time-series status material record;

[0031] A sliding window for material level changes is constructed to segment the sequence of material level heights in the buffer silo. The material level descent rate within each sliding window is calculated, and a descent rate sequence is constructed by sorting the segments.

[0032] The descent rate sequence is fitted with a trend. Based on the fitting results, it is determined whether there is a continuous downward trend in the material level. The corresponding downward trend segment is extracted, and the ratio of the remaining material level height in the silo to the slope of the corresponding downward trend segment is calculated to obtain the first prediction condition.

[0033] The system obtains the real-time operating speed, conveying distance, and real-time material conveying length of the belt in the upstream feeding area of ​​the conveying path, and calculates the conveying arrival time of the fed material to the dry separation section.

[0034] Based on the first prediction condition and the delivery arrival time, the material delay when the material reaches the dry selection section is calculated. It is then determined whether the material delay is higher than the preset lag judgment threshold. If it is higher, a material supply disconnection label is added.

[0035] Furthermore, the methods for predicting load fluctuations include:

[0036] Extract the equipment load data of the latest preset sampling period and the previous preset sampling period of the dry separation section, and at the same time extract the concentrate grade and tailings grade data of the corresponding time period, calculate the dry separation operation load difference between adjacent preset sampling periods, and determine the direction of load change.

[0037] The dry separation load change rate is calculated based on the dry separation operating load difference. The frequency of continuous increase and continuous decrease of the dry separation load change rate in all preset sampling periods up to the latest preset sampling period is counted as the trend jump frequency.

[0038] If the trend jump frequency is higher than the preset frequency threshold, but there is a material delay and a feeding disconnection label, then add an unstable trend prediction label to the dry selection section.

[0039] The rate of change of concentrate grade and the rate of change of particle size of crushed products are calculated based on the concentrate grade and tailings grade data. At the same time, the correlation value between the rate of change of concentrate grade and the rate of change of particle size of crushed products is calculated. If the correlation value is lower than the preset correlation threshold, a parameter mismatch label is added.

[0040] By combining the trend prediction instability label, parameter mismatch label and state classification results, and matching them with the preset instruction template, the predicted control intention instruction is output.

[0041] Furthermore, the method for performing rhythm comparison includes:

[0042] Extract the operating load values, timestamps, control command records, and actual response records from the historical production operation data belonging to the crushing stage, dry separation stage, and intermediate stage to form a historical dataset;

[0043] The historical dataset is grouped according to the existing operating cycle to obtain the periodic operating data group; the load change trend is calculated for each periodic operating data group, and the load fluctuation characteristics are extracted based on the calculation results;

[0044] The characteristics of load pattern fluctuations are statistically analyzed, and a load pattern feature library is constructed based on all load pattern fluctuation characteristics and actual response records;

[0045] Real-time acquisition of relevant operational data belonging to the latest production stage from the equipment operation dataset, and construction of existing load rhythm feature vectors based on this relevant operational data;

[0046] Calculate the pattern similarity between the existing load rhythm feature vector and any load pattern feature in the load pattern feature library. If the pattern similarity is lower than the preset rhythm similarity threshold, output the optimized control intention command.

[0047] Furthermore, the method for adjusting the sorting logic includes:

[0048] The optimized control intent commands and the predictive control intent commands are initially classified according to the command generation time and the production stage of the target equipment.

[0049] Identify the urgency and priority of each type of control intent instruction, and use the preset execution conflict rules as an auxiliary judgment basis to sort each type of control intent instruction and determine the execution priority sequence;

[0050] The execution priority sequence is matched with the actual scheduling specification, and the control tick is output.

[0051] Furthermore, the execution time offset includes the following methods:

[0052] Identify the target execution device number and type, communication channel type, and response time of each control intent instruction in the control cycle, and combine them into an instruction execution mapping subset;

[0053] The target execution device with the longest response time among the target execution devices corresponding to the subset of identified instruction execution mapping is used as the synchronous triggering reference device;

[0054] Calculate the response time difference between the response time of the synchronous triggering reference device and the response time of the other target execution devices. At the same time, calculate the instruction response time difference between the response time of all control instructions of the other target execution devices and the response time of the corresponding control instruction of the synchronous triggering reference device. Construct a response time offset matrix based on the response time difference and the instruction response time difference.

[0055] The time difference of each item in the response time offset matrix is ​​rounded down to calculate the delay, and a synchronous trigger time base unit is constructed. Based on the synchronous trigger time base unit, the execution time of all control commands is offset and adjusted to generate a standardized synchronous trigger time axis.

[0056] The standardized synchronous trigger timeline is bound to the instruction execution mapping subset and the corresponding target execution device number to form a synchronous control execution list.

[0057] A mineral processing plant stage crushing-dry separation integrated intelligent load control system, which is used to realize the intelligent load control method for stage crushing-dry separation in a mineral processing plant, including:

[0058] The data acquisition module collects real-time equipment operating status information from the crushing section, dry separation section, and intermediate processes, and performs data cleaning on the equipment operating status information to obtain the equipment operating dataset.

[0059] The status identification module extracts dynamic indicators from the equipment operation dataset and performs real-time change analysis. At the same time, it performs fluctuation statistics and numerical difference balance calculation on the equipment operation dataset, analyzes the load correlation and operating rhythm matching degree between the crushing section and the dry separation section, classifies the operating status based on the analysis results and statistical calculation results, and outputs the status classification results.

[0060] The trend prediction module performs trend extension calculations on the state classification results and outputs the material arrival time delay of the dry separation section; it identifies the load trend of the dry separation section, combines the material arrival time delay to predict load fluctuations, and obtains predictive control intention commands.

[0061] The rhythm optimization module acquires historical production operation data and identifies load pattern characteristics. Based on the load pattern characteristics, it performs rhythm comparison on the equipment operation dataset and outputs optimization control intention commands.

[0062] The instruction synchronization module combines optimized control intent instructions and predicted control intent instructions to adjust the sorting logic and obtain the control cycle; it also offsets the execution time of the control cycle and generates a synchronized control execution list.

[0063] The linkage execution module extracts control instructions from the synchronous control execution list and distributes them one by one. It dynamically adjusts the equipment operating status to achieve load coordination and balance between the stage crushing section and the dry separation section. It receives feedback information and transmits the feedback information back to the preset mineral processing control terminal. The modules are connected to each other via wired and / or wireless means.

[0064] The technical effects and advantages of the intelligent load control system and method for stage crushing-dry separation in mineral processing plants according to this invention are as follows:

[0065] By collecting real-time equipment operation datasets, and subsequently performing state identification, trend prediction, rhythm optimization, command synchronization, and coordinated execution based on these datasets, a smart load control system and method for stage crushing-dry separation in a mineral processing plant was realized. Compared with existing technologies, by extracting real-time dynamic information and constructing disturbance degree indicators, the accuracy of identifying the stability of material flow state between crushing and dry separation was improved. By setting multi-dimensional judgment conditions to distinguish between stable low-load states and intermittent material shortage states, damage to dry separation equipment and power waste caused by misjudgment were avoided. Through cross-judgment of crushing particle size and dry separation grade fluctuation coefficient, a system for intelligent load control of crushing-dry separation in a mineral processing plant was realized. It accurately pinpoints two types of process anomalies: improper crushing particle size control and dry separation parameter mismatch. By calculating material time delay and combining it with correlation analysis, it enhances the predictive ability of load shift and the diagnostic ability of parameter mismatch. By comparing rhythms, it realizes the identification and scheduling optimization of load patterns. By identifying the response delay differences of different communication channels, it eliminates the control rhythm misalignment caused by communication delays in multi-stage equipment. It improves the intelligent perception capability of the entire process of load regulation of crushing-dry separation linkage in the concentrator and the flexibility of multi-stage collaborative regulation, while ensuring that the load regulation has the accuracy of state identification and the timeliness of anomaly response. Attached Figure Description

[0066] Figure 1 This is a schematic diagram of the intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to the present invention.

[0067] Figure 2 This is a schematic diagram of the intelligent load control system for stage crushing and dry separation in a mineral processing plant according to the present invention. Detailed Implementation

[0068] 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.

[0069] Example 1

[0070] Please see Figure 1 As shown in this embodiment, the intelligent load control method for stage crushing-dry separation in a mineral processing plant includes:

[0071] S1. Real-time acquisition of equipment operating status information in the crushing section, dry separation section, and intermediate processes, and data cleaning of the equipment operating status information to obtain equipment operating dataset;

[0072] S2. Extract dynamic indicators from the equipment operation dataset and perform real-time change analysis. At the same time, perform fluctuation statistics and numerical difference balance calculation on the equipment operation dataset, analyze the load correlation and operation rhythm matching degree between the crushing section and the dry separation section, classify the operation status based on the analysis results and statistical calculation results, and output the status classification results.

[0073] S3. Perform trend extension calculation on the state classification results and output the material time delay to the dry separation section; identify the load trend of the dry separation section, combine the material time delay to predict the load fluctuation, and obtain the predictive control intention command;

[0074] S4. Acquire historical production operation data and identify load pattern characteristics. Based on the load pattern characteristics, perform rhythm comparison on the equipment operation dataset and output optimization control intention commands.

[0075] S5. Combine the optimized control intent instructions and the predicted control intent instructions to adjust the sorting logic and obtain the control cycle; offset the execution time of the control cycle and generate a synchronous control execution list;

[0076] S6. Extract control instructions from the synchronous control execution list and distribute them one by one, dynamically adjust the equipment operating status to achieve load coordination balance between the stage crushing section and the dry separation section, receive feedback information and transmit the feedback information back to the preset mineral processing control terminal.

[0077] In this embodiment, the equipment operating status information refers to the operating parameters of relevant equipment collected in real time during the production process of the mineral processing plant, including the material crushing stage, the dry separation stage, and the transmission stage between the two stages. These parameters include data that reflect the operating conditions of the relevant equipment, such as motor voltage, current, equipment start-up and shutdown status, material level height, particle size distribution of crushed products, magnetic field strength of dry separation equipment, load status, and communication status. The crushing stage includes multiple crushing stages such as coarse crushing, medium crushing, and fine crushing, which are set according to the particle size progression. The dry separation stage includes methods such as magnetic separation or air separation. Data cleaning is achieved by performing noise filtering and outlier removal to obtain a more accurate equipment operating dataset, thereby improving the accuracy of the original data.

[0078] Methods for performing real-time change analysis include:

[0079] The raw height sequence, which is the dynamic indicator, is formed by continuously collecting material level height data from the buffer silo between the crushing stage and the dry separation stage in the equipment operation data during the preset sampling period.

[0080] The process involves extracting the material level height data from the buffer silo located between the crushing and dry separation stages from the real-time collected equipment operation data and sorting it according to the timestamp to obtain the original height sequence. The preset sampling period refers to the period time set based on the production plan specifications of the concentrator. This original height sequence reflects the increase and decrease process of material inside the silo and serves as the data basis for subsequent operations.

[0081] Calculate the height difference between adjacent moments in the original height sequence, and calculate the ratio of this height difference to the unit time interval to obtain the material level change rate. Calculate the average rate of change of material level for each moment in the corresponding time segment of the original height sequence to obtain the average change rate.

[0082] The process involves calculating the difference between consecutive timestamp sample points in the original height sequence and comparing it with the unit time of a preset sampling period to output the material level change rate. Then, the average change rate is obtained by calculating the average value. This average change rate is used to reflect the overall change in material level height.

[0083] The sampling points where the direction of material level change reverses in the original height sequence are counted, and the frequency of direction change is output.

[0084] The reversal of the material level change direction refers to identifying the rise or fall of the material level at each moment as the direction of change. When the direction of change at a certain moment is opposite to that at the previous moment, it is recorded as a reversal. The frequency of the direction change is obtained by counting the number of times this reversal occurs. The larger the value, the more unstable the material flow.

[0085] Simultaneously, a linear fit is performed on the original height sequence, a linear fit curve is plotted, and the slope of the linear fit curve is calculated.

[0086] Since the original height sequence is a set of continuous points with time as the X-axis and material level as the Y-axis, the least squares method is used to linearly fit the sequence in this embodiment to obtain a fitted straight line representing the overall trend of change. The slope of the fitted curve is used to characterize the directionality of the overall increase or decrease of the material level. A positive slope indicates continuous feeding, a negative slope indicates continuous discharge, and a very small slope indicates that the material level is close to stable. This is used to identify the changing trend of the material supply and demand status.

[0087] The standard deviation and extreme value difference of the material level change rate at each moment in the original height sequence are calculated, and combined with the average change rate, the frequency of directional change, and the slope of the linear fitting curve, they form an index of the degree of disturbance.

[0088] The standard deviation of the rate of change of material level is used to measure the degree of material level fluctuation, and the extreme difference of the rate of change of material level is used to identify strong short-term disturbances. These two values ​​are combined with the previously obtained average rate of change, frequency of directional change, and slope reflecting the overall trend to form a set of disturbance degree indicators, which are used to reflect the degree of fluctuation activity of the silo in the corresponding time period.

[0089] Each component of the disturbance level index is compared with the preset control rules, and the disturbance level of the corresponding component is output. The disturbance levels of all components are weighted and summed to obtain the stability score of the corresponding buffer silo.

[0090] The preset control rules refer to the threshold ranges for various parameters related to the operation of the buffer silo, which are set based on the known production specifications of the buffer silo. Each component index in the disturbance degree index is matched with the preset control rules to determine the disturbance level of each component index under the production specifications of the buffer silo, output the value corresponding to the disturbance level, and then the disturbance level of each component index is weighted and summed to obtain the stability score of the buffer silo.

[0091] The weight of each component index in the weighting process is set based on the importance of each component in the production conditions, and those skilled in the art can adjust it based on the specific disclosure; the higher the stability score, the more stable the operation of the buffer equipment in the crushing and dry separation stage, and the lower the score, the more significant and frequent the material fluctuations. This value is the analysis result.

[0092] Methods for performing fluctuation statistics and numerical difference balance calculations include:

[0093] Extract the continuously recorded equipment load values ​​from the equipment operation data set within the preset sampling period of the crushing section, screening section and dry separation section respectively, and construct the load value sequence of the corresponding stage.

[0094] This involves collecting real-time load data from relevant equipment in the crushing, screening, and dry separation stages in chronological order, and serializing the data according to a preset sampling period to obtain independent operating load value sequences belonging to the three process sections of crushing, screening, and dry separation.

[0095] The standard deviation of the load value sequence for each stage is calculated, along with the dispersion coefficient of the crushed particle size distribution and the fluctuation coefficient of the dry separation grade. The standard deviation of the load is then sorted according to the preset sampling period to obtain the load fluctuation sequence.

[0096] The standard deviation of the load value sequence for each process stage is calculated to obtain the load standard deviation, which is used to measure the activity of load changes in that process stage. These load standard deviations are sorted according to the time sequence of the preset sampling period to obtain the load fluctuation sequence, which identifies the degree of fluctuation of each process stage in production operation and serves as the data basis for subsequent operations. The dispersion coefficient of crushed particle size distribution and the fluctuation coefficient of dry separation grade are parameters extracted based on the process requirements of the production line, reflecting the results of the crushing stage and the process of the dry separation stage.

[0097] The load standard deviations belonging to the same equipment number within adjacent preset sampling periods in the load fluctuation sequence are identified and the difference is calculated to obtain the fluctuation change amplitude. The fluctuation change amplitude sequence of the corresponding stage is constructed. In this case, by analyzing the fluctuation change of a single equipment in two consecutive preset sampling periods, the sudden change information of the equipment's operation fluctuation is obtained based on the difference of the load standard deviation. All fluctuation change amplitudes are integrated to construct the fluctuation change amplitude sequence of the equipment in the corresponding process stage.

[0098] A sliding window is constructed to traverse the fluctuation amplitude sequence, and the first derivative of each sliding window is calculated. The time intervals corresponding to consecutive sliding windows with the same first derivative sign are selected, and the number and average duration of monotonically rising and monotonically falling segments are counted and combined into a fluctuation trend indicator.

[0099] This involves constructing a fixed-length sliding window to traverse the fluctuation amplitude sequence, dividing the fluctuation amplitude sequence into window segments, and calculating the first derivative of the corresponding data segment within each sliding window. The specific length of the sliding window is set according to the length of the entire fluctuation amplitude sequence to ensure that the fluctuation amplitude sequence can be completely divided into several window segments.

[0100] The method uses the sign of the first derivative to identify the direction of the change trend, selects several sampling points with the same sign of the first derivative in the time interval formed by the continuous sliding window, counts the number of segments with monotonic change trends, and calculates the average duration of the monotonic change trend segments in each direction to form a fluctuation trend index, which is used to reflect the speed of load operation status evolution.

[0101] Extract time segments from sliding windows whose fluctuation amplitude is below the preset lower limit and whose average duration is above the preset low amplitude duration threshold, and mark them as stable load decline segments.

[0102] The system sets a preset lower limit for amplitude and a preset threshold for the duration of low amplitude based on historical experience. These are used to determine whether the amplitude of fluctuation is lower than the theoretical lower limit and whether the average duration is higher than the theoretical duration. If the amplitude of fluctuation in a certain time period is always less than the preset lower limit for amplitude and the average duration of that time period is higher than the preset threshold for the duration of low amplitude, then this time period can be considered to be a continuous and stable decreasing operating state, corresponding to a typical operating mode of long-term stable low load.

[0103] The abrupt change percentage is obtained by calculating the ratio of the time period during which the fluctuation amplitude exceeds the preset upper limit to the time period during which the stable load decreases.

[0104] The upper limit of the preset amplitude is set based on historical experience. This threshold is used to determine whether the fluctuation amplitude is higher than the theoretical upper limit. The ratio of the cumulative time length of the time segment where the fluctuation amplitude is higher than the preset upper limit to the cumulative time length of the stable load decrease segment is used to obtain the mutation ratio. The larger the value, the more unstable and mutated fluctuations there are in the corresponding process stage.

[0105] By integrating fluctuation trend indicators, mutation ratios, parameters corresponding to stable load decline segments, as well as the coefficient of variation and fluctuation coefficient, and performing normalization processing, a fluctuation feature vector is constructed. The parameters corresponding to stable load decline segments refer to the duration, quantity, and load value of the corresponding time segment. After normalization processing, the parameters in the fluctuation trend indicators, mutation ratios, and parameters corresponding to stable load decline segments are combined with the coefficient of variation and fluctuation coefficient to form a vector form, which is the statistical calculation result.

[0106] The ways to classify operating status include:

[0107] The judgment conditions are set as follows: if the stability score is higher than the stability score threshold, it is recorded as the material level status is stable; if the mutation rate is lower than the mutation ratio threshold, it is recorded as the disturbance level is low; if the duration of the stable load decrease section is higher than the preset low amplitude duration threshold, it is recorded as the continuous trend is stable; if the number of monotonic change sections is less than the number threshold, it is recorded as the change trend is stable.

[0108] Four judgment conditions were set, and corresponding thresholds were set for each condition based on historical judgment experience. These thresholds are used to represent the values ​​that the parameters of each dimension should meet in theory. If the stability score is higher than the stability frequency division threshold, it indicates that the material transmission is continuous and the fluctuation is small. If the mutation ratio is lower than the mutation proportion threshold, it indicates that the material level change is consistent and there is no obvious supply interruption. The longer the duration of the stable load decline section, the more stable and predictable the equipment in the corresponding process stage is in a long-term low rate of change trend. If the number of monotonically increasing or decreasing monotonically changing segments is small, it indicates that the load change of the equipment in the corresponding process stage is repeated less often and is relatively stable.

[0109] If any three of the judgment conditions are met, the operating state will be determined as a stable low-load state.

[0110] If three or more of the above judgment conditions are met, it indicates that the equipment in the corresponding process stage is in a state of overall stable operation, good material supply continuity and small local disturbances, and a clear trend; that is, it is in a low-load but stable operation mode, such as a period of phased production interruption or energy consumption optimization period with reduced load. Therefore, this type of state is named stable low-load state.

[0111] If the stability score is lower than the stability score threshold, the mutation rate is higher than the mutation ratio threshold, and the number of monotonic segments is higher than the quantity threshold, then the operating status will be judged as intermittent material shortage status.

[0112] A low stability score indicates unstable material level fluctuations, a high percentage of sudden changes indicates frequent material jumps and an imbalance between inflow and outflow. A large number of monotonic change segments indicate that the load changes repeatedly and is in a discontinuous state. Therefore, this type of state is named the intermittent material shortage state.

[0113] If the dispersion coefficient is greater than the particle size dispersion threshold but the fluctuation coefficient is less than the grade fluctuation threshold, it is determined that the crushing particle size control is inappropriate.

[0114] The particle size dispersion threshold is set based on the relevant industrial production requirements of the crushing stage, and the grade fluctuation threshold is set based on the relevant industrial production requirements of the dry separation stage. If the dispersion coefficient is large but the fluctuation coefficient is small, it indicates that the crushing is not uniform enough, but the concentrate grade is relatively stable. This means that the dry separation equipment can adapt to poor particle size. In this case, the relevant parameters of the crushing stage need to be adjusted first, so a label for improper crushing particle size control is added.

[0115] If the dispersion coefficient is less than the particle size dispersion threshold but the fluctuation coefficient is greater than the grade fluctuation threshold, it is determined to be a dry separation parameter mismatch.

[0116] If the dispersion coefficient is small, it indicates that the particle size distribution is concentrated and the crushing stage is successful. However, if the fluctuation coefficient is large, it indicates that the concentrate grade fluctuates in the dry separation stage. The relevant parameters in the dry separation stage may not be suitable for the current particle size. This could be due to a mismatch between parameters such as magnetic field strength or cylinder rotation speed and the particle size. The dry separation parameters need to be reset according to the particle size of the crushed product. Therefore, a dry separation parameter mismatch label is added.

[0117] Methods for performing trend extension calculations include:

[0118] Extract the status classification labels corresponding to the status classification results, obtain the corresponding buffer silo material level height, and associate them with the time-series status material records.

[0119] The status classification labels refer to stable low load status, improper crushing particle size control, dry separation parameter mismatch, and intermittent material shortage status. The obtained buffer silo level height refers to the latest obtained buffer silo level height data within the corresponding preset sampling period. The status classification labels are combined with the corresponding level height data to form a time series record with labels.

[0120] A sliding window for material level changes is constructed to segment the sequence of material level heights in the buffer silo. The material level descent rate within each sliding window is calculated, and a descent rate sequence is constructed by sorting the segments.

[0121] The window size of the sliding window for material level changes is set based on the length of the sequence consisting of material level height data, ensuring that the sequence can be completely segmented into windows. Within each window, the average material level drop rate for that local time period is calculated, and the drop rate sequence is obtained by sorting the windows in order.

[0122] The descent rate sequence is fitted with a trend. Based on the fitting results, it is determined whether there is a continuous downward trend in the material level. The corresponding downward trend segment is extracted, and the ratio of the remaining material level height in the silo to the slope of the corresponding downward trend segment is calculated to obtain the first prediction condition.

[0123] In this embodiment, a fitting algorithm is used to perform time series fitting on the descent rate sequence to identify whether there is a continuous negative slope trend and extract the corresponding descent trend segment. The first prediction condition is obtained by calculating the ratio of the collected remaining material level height in the silo to the absolute value of the change slope of the corresponding descent trend segment, which is used to indicate how long the silo will run out if the descent trend remains unchanged.

[0124] The system obtains the real-time operating speed, conveying distance, and real-time material conveying length of the belt in the upstream feeding area of ​​the conveying path, and calculates the conveying arrival time of the fed material to the dry separation section.

[0125] The system obtains the real-time running speed of the belt corresponding to the upstream area where the feeding operation is performed in the conveying path between the crushing section and the dry separation section, the completed conveying distance, and the conveying length of the material flow still in the conveying process. The delivery time is obtained by calculating the ratio of the real-time material conveying length to the real-time running speed of the belt.

[0126] The material delay at which the material reaches the dry selection section is calculated based on the first prediction condition and the delivery arrival time. It is then determined whether the material delay is higher than the preset lag judgment threshold. If it is higher, a material supply disconnection label is added. The time difference obtained by subtracting the first prediction condition and the delivery arrival time is the material delay. A corresponding preset lag judgment threshold is set based on the production line operation specifications. If the material delay is higher than the threshold, the material delay is determined.

[0127] Methods for predicting load fluctuations include:

[0128] Extract the equipment load data of the latest preset sampling period and the previous preset sampling period of the dry separation section. At the same time, extract the concentrate grade and tailings grade data of the corresponding time period, calculate the dry separation operation load difference between adjacent preset sampling periods, and determine the direction of load change.

[0129] The dry separation operation load difference is obtained by real-time acquisition of the load parameters in the latest preset sampling period and the load parameters in the previous period in the dry separation section, and the load difference is calculated to identify the load change amplitude and load change direction in the current state. At the same time, the concentrate grade data and tailings grade data obtained in the corresponding period are also acquired as real-time status control references.

[0130] The dry separation load change rate is calculated based on the dry separation operating load difference. The frequency of continuous increase and continuous decrease of the dry separation load change rate in all preset sampling periods up to the latest preset sampling period is counted as the trend jump frequency.

[0131] The dry separation load change rate is obtained by calculating the ratio of the dry separation load difference between adjacent preset sampling periods to the unit time interval. The number of times the direction sign of all dry separation load change rates alternates in all preset sampling periods up to the latest preset sampling period is counted. In this embodiment, "continuous" means at least three consecutive preset sampling periods. The larger the value, the more frequent the control, which is related to the intermittent supply of material flow between the crushing section and the dry separation section.

[0132] If the frequency of trend jumps is higher than the preset frequency threshold, but there is a material supply disconnection label due to material delay, then add an unstable trend prediction label to the dry selection section.

[0133] The system sets a preset frequency threshold based on the conventional theoretical conditions recorded in historical records. If the frequency of trend jumps is higher than this threshold, and the material delay at this time carries a feeding disconnection tag, then the current operating status trend of the equipment in the dry separation section is judged to be unstable. The control terminal needs to select a more stable and conservative operating strategy according to the tag.

[0134] The rate of change of concentrate grade and the rate of change of particle size of crushed products are calculated based on the concentrate grade and tailings grade data. At the same time, the correlation value between the rate of change of concentrate grade and the rate of change of particle size of crushed products is calculated. If the correlation value is lower than the preset correlation threshold, a parameter mismatch label is added.

[0135] This method calculates the rate of change in concentrate grade and the rate of change in particle size of crushed products based on the relationship between mineral grade and particle size in mineral processing principles. In this embodiment, the Pearson coefficient between the rate of change in concentrate grade and the rate of change in particle size of crushed products is used as the correlation value. Based on theoretical knowledge about the application of the Pearson coefficient in the mining field, a correlation threshold is set. If the correlation value is lower than the corresponding threshold, it may indicate that the mineral particle size has become finer, that the crushing stage has been over-crushed, resulting in a decrease in grade, or that the particle size is stable but the grade fluctuates drastically, which may be due to an inappropriate magnetic field strength, or other abnormal situations. However, no matter what kind of abnormal situation, it reflects that there is a parameter mismatch between the crushing stage and the dry separation stage. Therefore, a parameter mismatch label is added to such abnormalities.

[0136] By combining the trend prediction instability label, parameter mismatch label and state classification results, and matching them with the preset instruction template, the predicted control intention instruction is output.

[0137] The system combines trend prediction instability labels, parameter mismatch labels, and state classification results to reflect the operating conditions of the relevant equipment between the crushing section and the dry separation section. This operating condition is then matched with a preset instruction template based on production line operating specifications and historical control experience. According to this template, a preset control intention instruction suitable for the corresponding operating condition is selected. The purpose of this type of instruction is to output control signals that can guide equipment or manual actions in advance before the actual load of the equipment between the crushing section and the dry separation section changes, thereby indirectly improving the overall system operating efficiency.

[0138] Methods for rhythm comparison include:

[0139] Extract the operating load values, timestamps, control command records, and actual response records from the historical production operation data belonging to the crushing stage, dry separation stage, and intermediate stage to form a historical dataset.

[0140] The historical production and operation data includes the operating load data and timestamps of various equipment in the crushing stage, dry separation stage and intermediate process stage of the ore dressing plant production line in multiple continuous operating cycles. It also records the execution of each control command and the equipment response. These data constitute the historical dataset, which serves as the data foundation for subsequent operations.

[0141] The historical dataset is grouped according to the existing running cycle to obtain the periodic running data group. The data in the historical dataset is grouped according to the time length of the preset sampling cycle to obtain the periodic running data group for each cycle.

[0142] For each cycle of operational data, load change trends are calculated, and load fluctuation characteristics are extracted based on the calculation results.

[0143] The calculation of load change trends refers to the calculation of the overall load increase or decrease of each device in the operating data group of the calculation period, the degree of load fluctuation, the maximum load value, and the duration of the maximum load, which reflect the operating load conditions. These values ​​are vectorized to obtain the load fluctuation characteristics, and each load fluctuation characteristic corresponds to the load change of the device in a historical record.

[0144] The characteristics of load pattern fluctuations are statistically analyzed, and a load pattern feature library is constructed based on all load pattern fluctuation characteristics and actual response records.

[0145] This embodiment integrates the load pattern fluctuation characteristics from all historical records and uses a clustering algorithm to process these characteristics, forming high-frequency periodic load change patterns, such as "staggered feeding - continuous high load operation" or "operation - short stop - restart operation," which are feature labels reflecting the equipment's operating status. At the same time, the actual response records in each historical record are associated with these features to establish a behavioral pattern database that maps the operating mode to the actual implementation effect, which is the load pattern feature library.

[0146] The system acquires relevant operational data from the latest production stage in the equipment operation dataset in real time. Based on this relevant operational data, it constructs an existing load rhythm feature vector. This involves real-time acquisition of relevant operational data from each piece of equipment in the latest production stage of the crushing section, dry separation section, and intermediate stage, extracting relevant operational indicators that are consistent with the structural dimensions in the load pattern feature library, and encapsulating these indicators into vector form to obtain the load rhythm feature vector, which constitutes a real-time expression of the current load situation.

[0147] Calculate the pattern similarity between the existing load rhythm feature vector and any load pattern feature in the load pattern feature library. If the pattern similarity is lower than the preset rhythm similarity threshold, output the optimized control intention command.

[0148] The formula for calculating pattern similarity is as follows: ;in, Indicate pattern similarity; Represents the existing load rhythm feature vector; This indicates that the load pattern conforms to any load pattern feature in the pattern feature library. Based on historical rhythm comparison experience, a preset rhythm similarity threshold is set. If the current pattern similarity is lower than the threshold, it means that the current running rhythm is less similar to any known pattern, and it is considered to be in a rhythm mismatch state. Therefore, an optimization control intention instruction is output. The purpose of this instruction is to enhance the adjustment capability and execute system regulation docking in advance. For example, it can introduce a backup crushing unit or dry separation unit in advance for regulation, delay or cancel the feeding of the next cycle, and adjust the crushing frequency to smooth the rhythm imbalance point, so that the operation mode intervention is controllable and the rhythm calibration is achieved under complex conditions.

[0149] The methods for adjusting the sorting logic include:

[0150] The optimized control intent commands and the predictive control intent commands are initially classified according to the command generation time and the production stage of the target equipment.

[0151] The optimized control intent commands and the predictive control intent commands originate from the trend prediction module and the rhythm optimization module, respectively. In the current command synchronization module, they are arranged in sequence according to the timestamp of command generation and grouped and categorized according to the production stage of the target equipment, ensuring that the subsequent sorting logic can be adjusted for the control needs of equipment at different stages.

[0152] The system identifies the urgency priority of each type of control intent instruction and uses preset execution conflict rules as an auxiliary judgment basis to sort each type of control intent instruction and determine the execution priority sequence. The urgency priority of each type of control intent instruction is identified based on the production stage and target equipment type of its effect. The urgency priority of the control instruction is obtained by matching it with the known planned priority settings for various operation tasks of the production line.

[0153] It should be noted that the crushing particle size adjustment command has a higher priority than the dry separation parameter adjustment command, because particle size changes need to be transmitted to the downstream in advance, and the grade-related adjustment command has a higher priority than the output-related adjustment command.

[0154] The preset execution conflict rule refers to the judgment rule set based on the production line operation specifications to determine whether there are contradictions or resource competition among multiple instructions in the same time period. This rule is used to filter or merge multiple instructions that have logical conflicts. Through priority identification and instruction conflict resolution, the execution order of each instruction is determined to form an execution priority sequence.

[0155] The execution priority sequence is matched with the actual scheduling specification, and the control tick is output.

[0156] The actual scheduling specification refers to the basic scheduling rules for each piece of equipment in the production line, such as the minimum time interval requirement for equipment start-up and shutdown, the minimum control execution cycle, and safety operation constraints. The execution priority sequence is matched with the actual scheduling specification to ensure that the output control instructions meet the process requirements and safety constraints of the production line. The output set of instructions is the control cycle.

[0157] The methods for execution time offset include:

[0158] The target execution device number and type, communication channel type, and response time of each control intent instruction in the control cycle are identified and combined into an instruction execution mapping subset.

[0159] This involves extracting control instructions one by one from the control cycle, identifying the target execution device number and device type corresponding to the instruction, such as crusher No. 1 or dry separator No. 1; communication channel type refers to communication channels such as PLC control bus or industrial Ethernet; response time refers to the delay time from receiving the instruction to completing the task required by the corresponding target execution device. These three types of indicators are combined into an instruction execution mapping subset, and each instruction execution mapping subset corresponds to a control intention instruction.

[0160] The target execution device with the longest response time among the target execution devices corresponding to the instruction execution mapping subset is used as the synchronization trigger reference device. The target execution device with the longest response time is the slowest response device. The response time of this device is used as the benchmark anchor point for time synchronization to calculate the time offset by which other faster response devices need to delay sending instructions.

[0161] Calculate the response time difference between the response time of the synchronous triggering reference device and the response time of the other target execution devices. At the same time, calculate the instruction response time difference between the response time of all control instructions of the other target execution devices and the response time of the corresponding control instruction of the synchronous triggering reference device. Construct a response time offset matrix based on the response time difference and the instruction response time difference.

[0162] The response time difference of the device and the response time difference of the instruction were calculated separately. The response time difference of the device needs to be determined in conjunction with the communication channel type of the device. The two time differences are arranged in a matrix with the corresponding device number and instruction number to obtain the response time offset matrix. Each element in the matrix represents the time offset of a device relative to the synchronous triggering reference device when executing a certain instruction.

[0163] The time difference of each item in the response time offset matrix is ​​calculated by rounding down the delay to construct a synchronous trigger time base unit.

[0164] Due to the minimum latency limitations imposed by clock precision and communication cycles in actual operation (e.g., a PLC scan cycle of 10ms and a communication cycle of 20ms), the time differences of each item in the response time offset matrix need to be rounded according to the minimum latency limit to ensure that the offset is an integer multiple of the smallest executable time granularity. The formula for rounding is as follows: ;in, This represents the synchronization trigger time base unit after rounding down a certain time offset. express A specific original time offset in the corresponding response time offset matrix; This indicates the minimum latency limit for the communication type corresponding to the instruction that requires rounding.

[0165] The execution times of all control commands are offset and adjusted based on the synchronous trigger time reference unit to generate a standardized synchronous trigger time axis.

[0166] By adjusting the original execution time of each control intention command using a synchronous trigger time reference unit, the time length of the synchronous trigger time reference unit is delayed for fast-responding equipment, ensuring that the command execution time of all equipment is synchronized and executed in the set order, thus forming a standardized synchronous trigger time axis. This ensures that equipment in multiple stages, such as the crushing section and the dry separation section, achieves true synchronous response during the linkage control process, eliminating the misalignment of control rhythm caused by communication delay differences.

[0167] The standardized synchronous trigger timeline is bound to the instruction execution mapping subset and the corresponding target execution device number to form a synchronous control execution list.

[0168] Each instruction in the instruction execution mapping subset is executed in the order of the standardized synchronous trigger time axis and bound to the corresponding application device number, forming a complete list of planned instruction synchronous control execution for the equipment between the crushing section and the dry separation section.

[0169] The methods for dynamic adjustment include:

[0170] The operating parameters of the target execution device are set based on the control instructions in the synchronous control execution list. At the same time, the actual operating status parameters of the target execution device are collected in real time. The deviation between the actual operating status parameters and the set operating parameters in the control instructions is compared, and the actual operating status parameters are adjusted in real time based on the deviation.

[0171] The process involves extracting control commands one by one from the synchronous control execution list according to the standardized synchronous trigger timeline, and distributing these commands to the target execution equipment in the crushing section, dry separation section, and intermediate processes through corresponding communication channels. The commands specify the required operating parameters, such as motor speed, feed rate, or vibration amplitude. Simultaneously, sensors and data acquisition units deployed on each target execution equipment collect real-time data on the actual operating status parameters. Compensation and adjustment commands are generated based on the deviation direction and magnitude of the real-time data to correct the actual operating status parameters in real time. For example, this includes increasing or decreasing the equipment operating frequency, adjusting the feed rate, and correcting the belt speed. When the crushed product particle size is too coarse, the crusher discharge opening is automatically lowered, and the magnetic field strength of the dry separation equipment is increased to compensate for the separation efficiency. When the dry separation concentrate grade decreases, the crusher parameters are adjusted in the opposite direction to increase the ore liberation degree until the actual parameters return to the allowable deviation range of the target parameters.

[0172] The execution result status of the target execution device for each instruction is extracted synchronously and bound to the original control instruction to form a feedback record of control execution. This record is then uploaded to a preset mineral processing control terminal via a wired communication network or industrial wireless signal channel to track the control effect and serve as a reference for subsequent system status updates, historical data archiving, and operational strategy optimization and adjustment.

[0173] This embodiment achieves an intelligent load control system and method for the crushing-dry separation linkage in a mineral processing plant by collecting real-time equipment operation datasets and performing state identification, trend prediction, rhythm optimization, command synchronization, and coordinated execution based on these datasets. Compared with existing technologies, it improves the accuracy of identifying the stability of material flow states between crushing and dry separation by extracting real-time dynamic information and constructing disturbance degree indicators. By setting multi-dimensional judgment conditions to distinguish between stable low-load states and intermittent material shortage states, it avoids damage to dry separation equipment and power waste caused by misjudgment. Furthermore, it achieves this by cross-judging the crushing particle size and dry separation grade fluctuation coefficient. This system enables precise identification of two types of process anomalies: improper crushing particle size control and mismatch of dry separation parameters. By calculating material time delay and combining it with correlation analysis, it enhances the predictive ability of load shifts and the diagnostic ability of parameter mismatches. By comparing rhythms, it achieves the identification and scheduling optimization of load patterns. By identifying the response delay differences of different communication channels, it eliminates the control rhythm misalignment caused by communication delays in multi-stage equipment. It improves the intelligent perception capability of the entire process of load regulation of crushing-dry separation linkage in the mineral processing plant and the flexibility of multi-stage collaborative regulation, while ensuring that the load regulation has the accuracy of state identification and the timeliness of anomaly response.

[0174] Example 2

[0175] Please see Figure 2As shown, parts not described in detail in this embodiment are described in Embodiment 1. A mineral processing plant stage crushing-dry separation linkage intelligent load control system is provided, including:

[0176] The data acquisition module collects real-time equipment operating status information from the crushing section, dry separation section, and intermediate processes, and performs data cleaning on the equipment operating status information to obtain the equipment operating dataset.

[0177] The status identification module extracts dynamic indicators from the equipment operation dataset and performs real-time change analysis. At the same time, it performs fluctuation statistics and numerical difference balance calculation on the equipment operation dataset, analyzes the load correlation and operating rhythm matching degree between the crushing section and the dry separation section, classifies the operating status based on the analysis results and statistical calculation results, and outputs the status classification results.

[0178] The trend prediction module performs trend extension calculations on the state classification results and outputs the material arrival time delay of the dry separation section; it identifies the load trend of the dry separation section, combines the material arrival time delay to predict load fluctuations, and obtains predictive control intention commands.

[0179] The rhythm optimization module acquires historical production operation data and identifies load pattern characteristics. Based on the load pattern characteristics, it performs rhythm comparison on the equipment operation dataset and outputs optimization control intention commands.

[0180] The instruction synchronization module combines optimized control intent instructions and predicted control intent instructions to adjust the sorting logic and obtain the control cycle; it also offsets the execution time of the control cycle and generates a synchronized control execution list.

[0181] The linkage execution module extracts control instructions from the synchronous control execution list and distributes them one by one. It dynamically adjusts the equipment operating status to achieve load coordination and balance between the stage crushing section and the dry separation section. It receives feedback information and transmits the feedback information back to the preset mineral processing control terminal. The modules are connected to each other via wired and / or wireless means.

[0182] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. 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.

[0183] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0184] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for intelligent load control of stage crushing-dry separation in a mineral processing plant, characterized in that, include: S1. Real-time acquisition of equipment operating status information in the crushing section, dry separation section, and intermediate processes, and data cleaning of the equipment operating status information to obtain equipment operating dataset; S2. Extract dynamic indicators from the equipment operation dataset and perform real-time change analysis. At the same time, perform fluctuation statistics and numerical difference balance calculation on the equipment operation dataset, analyze the load correlation and operation rhythm matching degree between the crushing section and the dry separation section, classify the operation status based on the analysis results and statistical calculation results, and output the status classification results. S3. Perform trend extension calculations on the state classification results and output the material delay before it reaches the dry selection section; Identify the load trend of the dry section, combine it with the material time delay to predict load fluctuations, and obtain the predictive control intention command; S4. Acquire historical production operation data and identify load pattern characteristics. Based on the load pattern characteristics, perform rhythm comparison on the equipment operation dataset and output optimization control intention commands. S5. Combine the optimized control intent instructions and the predicted control intent instructions to adjust the sorting logic and obtain the control cycle; offset the execution time of the control cycle and generate a synchronous control execution list; S6. Extract control instructions from the synchronous control execution list and distribute them one by one, dynamically adjust the equipment operating status to achieve load coordination balance between the stage crushing section and the dry separation section, receive feedback information and transmit the feedback information back to the preset mineral processing control terminal.

2. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 1, characterized in that, The methods for performing real-time change analysis include: The raw height sequence is formed by continuously collecting material level height data from the buffer silo between the crushing and dry separation stages in the equipment operation data during the preset sampling period, which is the dynamic indicator. Calculate the material level height difference between adjacent moments in the original height sequence, and calculate the ratio of this material level height difference to the unit time interval to obtain the material level change rate. Calculate the average material level change rate for each moment in the corresponding time segment of the original height sequence to obtain the average change rate. The sampling points where the direction of material level change reverses in the original height sequence are counted, and the frequency of direction change is output. At the same time, the original height sequence is linearly fitted, the linear fitting curve is plotted, and the slope of the linear fitting curve is calculated. The standard deviation and extreme value difference of the material level change rate at each moment in the original height sequence are calculated and combined with the average change rate, the frequency of directional change, and the slope of the linear fitting curve to form a disturbance degree index. Each component of the disturbance degree index is compared with the preset control rules, and the disturbance level of the corresponding component is output. The disturbance levels of all components are weighted and summed to obtain the stability score of the corresponding buffer silo.

3. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 2, characterized in that, The methods for performing fluctuation statistics and numerical difference balance calculations include: Extract the continuously recorded equipment load values ​​from the equipment operation data set within the preset sampling period of the crushing section, screening section and dry separation section respectively, and construct the load value sequence of the corresponding stage respectively; Calculate the load standard deviation of the load value sequence for each stage, and simultaneously calculate the dispersion coefficient of the crushed particle size distribution and the fluctuation coefficient of the dry separation grade. Sort the load standard deviations according to the preset sampling period to obtain the load fluctuation sequence. Identify the load standard deviations belonging to the same device number within adjacent preset sampling periods in the load fluctuation sequence and calculate the difference to obtain the fluctuation change amplitude, and construct the fluctuation change amplitude sequence for the corresponding stage. A sliding window is constructed to traverse the sequence of fluctuation amplitudes, and the first derivative of each sliding window is calculated. The time intervals corresponding to consecutive sliding windows with the same first derivative sign are selected, and the number of monotonically rising segments and monotonically falling segments and their average duration are counted and combined into a fluctuation trend indicator. Extract time segments within a sliding window whose fluctuation amplitude is lower than a preset lower limit and whose average duration is higher than a preset low amplitude duration threshold, and mark them as stable load decline segments. The abrupt change percentage is obtained by calculating the ratio of the time period during which the fluctuation amplitude exceeds the preset upper limit to the time period during which the stable load decreases. By integrating fluctuation trend indicators, abrupt change ratios, parameters corresponding to stable load decline segments, as well as dispersion coefficients and fluctuation coefficients, and performing normalization processing, a fluctuation feature vector is constructed.

4. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 3, characterized in that, The methods for classifying operating states include: The judgment conditions are set as follows: if the stability score is higher than the stability score threshold, it is recorded as the material level status is stable; if the mutation rate is lower than the mutation ratio threshold, it is recorded as the disturbance level is low; if the duration of the stable load decrease section is higher than the preset low amplitude duration threshold, it is recorded as the continuous trend is stable; if the number of monotonic change sections is less than the number threshold, it is recorded as the change trend is stable. If any three of the judgment conditions are met, the operating status is determined to be a stable low-load state; if the stability score is lower than the stability score threshold, the mutation rate is higher than the mutation ratio threshold, and the number of monotonic change segments is higher than the quantity threshold, the operating status is determined to be an intermittent material shortage state. If the dispersion coefficient is greater than the particle size dispersion threshold but the fluctuation coefficient is less than the grade fluctuation threshold, it is determined that the crushing particle size control is inappropriate; if the dispersion coefficient is less than the particle size dispersion threshold but the fluctuation coefficient is greater than the grade fluctuation threshold, it is determined that the dry separation parameters are mismatched.

5. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 4, characterized in that, The methods for performing trend extension calculations include: Extract the status classification labels corresponding to the status classification results, obtain the corresponding buffer silo material level height, and associate them with the time-series status material record; A sliding window for material level changes is constructed to segment the sequence of material level heights in the buffer silo. The material level descent rate within each sliding window is calculated, and a descent rate sequence is constructed by sorting the segments. The descent rate sequence is fitted with a trend. Based on the fitting results, it is determined whether there is a continuous downward trend in the material level. The corresponding downward trend segment is extracted, and the ratio of the remaining material level height in the silo to the slope of the corresponding downward trend segment is calculated to obtain the first prediction condition. The system obtains the real-time operating speed, conveying distance, and real-time material conveying length of the belt in the upstream feeding area of ​​the conveying path, and calculates the conveying arrival time of the fed material to the dry separation section. Based on the first prediction condition and the delivery arrival time, the material delay when the material reaches the dry selection section is calculated. It is then determined whether the material delay is higher than the preset lag judgment threshold. If it is higher, a material supply disconnection label is added.

6. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 5, characterized in that, The methods for predicting load fluctuations include: Extract the equipment load data of the latest preset sampling period and the previous preset sampling period of the dry separation section, and at the same time extract the concentrate grade and tailings grade data of the corresponding time period, calculate the dry separation operation load difference between adjacent preset sampling periods, and determine the direction of load change. The dry separation load change rate is calculated based on the dry separation operating load difference. The frequency of continuous increase and continuous decrease of the dry separation load change rate in all preset sampling periods up to the latest preset sampling period is counted as the trend jump frequency. If the trend jump frequency is higher than the preset frequency threshold, but there is a material delay and a feeding disconnection label, then add an unstable trend prediction label to the dry selection section. The rate of change of concentrate grade and the rate of change of particle size of crushed products are calculated based on the concentrate grade and tailings grade data. At the same time, the correlation value between the rate of change of concentrate grade and the rate of change of particle size of crushed products is calculated. If the correlation value is lower than the preset correlation threshold, a parameter mismatch label is added. By combining the trend prediction instability label, parameter mismatch label and state classification results, and matching them with the preset instruction template, the predicted control intention instruction is output.

7. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 6, characterized in that, The methods for performing rhythm comparison include: Extract the operating load values, timestamps, control command records, and actual response records from the historical production operation data belonging to the crushing stage, dry separation stage, and intermediate stage to form a historical dataset; The historical dataset is grouped according to the existing operating cycle to obtain the periodic operating data group; the load change trend is calculated for each periodic operating data group, and the load fluctuation characteristics are extracted based on the calculation results; The characteristics of load pattern fluctuations are statistically analyzed, and a load pattern feature library is constructed based on all load pattern fluctuation characteristics and actual response records; Real-time acquisition of relevant operational data belonging to the latest production stage from the equipment operation dataset, and construction of existing load rhythm feature vectors based on this relevant operational data; Calculate the pattern similarity between the existing load rhythm feature vector and any load pattern feature in the load pattern feature library. If the pattern similarity is lower than the preset rhythm similarity threshold, output the optimization control intention command.

8. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 7, characterized in that, The methods for adjusting the sorting logic include: The optimized control intent commands and the predictive control intent commands are initially classified according to the command generation time and the production stage of the target equipment. Identify the urgency and priority of each type of control intent instruction, and use the preset execution conflict rules as an auxiliary judgment basis to sort each type of control intent instruction and determine the execution priority sequence; The execution priority sequence is matched with the actual scheduling specification, and the control tick is output.

9. The intelligent load control method for stage crushing-dry separation linkage in a mineral processing plant according to claim 8, characterized in that, The execution time offset methods include: Identify the target execution device number and type, communication channel type, and response time of each control intent instruction in the control cycle, and combine them into an instruction execution mapping subset; The target execution device with the longest response time among the target execution devices corresponding to the subset of identified instruction execution mapping is used as the synchronous triggering reference device; Calculate the response time difference between the response time of the synchronous triggering reference device and the response time of the other target execution devices. At the same time, calculate the instruction response time difference between the response time of all control instructions of the other target execution devices and the response time of the corresponding control instruction of the synchronous triggering reference device. Construct a response time offset matrix based on the response time difference and the instruction response time difference. The time difference of each item in the response time offset matrix is ​​rounded down to calculate the delay, and a synchronous trigger time base unit is constructed. Based on the synchronous trigger time base unit, the execution time of all control commands is offset and adjusted to generate a standardized synchronous trigger time axis. The standardized synchronous trigger timeline is bound to the instruction execution mapping subset and the corresponding target execution device number to form a synchronous control execution list.

10. A mineral processing plant stage crushing-dry separation linkage intelligent load control system, used to implement the mineral processing plant stage crushing-dry separation linkage intelligent load control method as described in any one of claims 1-9, characterized in that, include: The data acquisition module collects real-time equipment operating status information from the crushing section, dry separation section, and intermediate processes, and performs data cleaning on the equipment operating status information to obtain the equipment operating dataset. The status identification module extracts dynamic indicators from the equipment operation dataset and performs real-time change analysis. At the same time, it performs fluctuation statistics and numerical difference balance calculation on the equipment operation dataset, analyzes the load correlation and operating rhythm matching degree between the crushing section and the dry separation section, classifies the operating status based on the analysis results and statistical calculation results, and outputs the status classification results. The trend prediction module performs trend extension calculations on the state classification results and outputs the material arrival time delay at the dry separation section. Identify the load trend of the dry section, combine it with the material time delay to predict load fluctuations, and obtain the predictive control intention command; The rhythm optimization module acquires historical production operation data and identifies load pattern characteristics. Based on the load pattern characteristics, it performs rhythm comparison on the equipment operation dataset and outputs optimization control intention commands. The instruction synchronization module combines optimized control intent instructions and predicted control intent instructions to adjust the sorting logic and obtain the control cycle; it also offsets the execution time of the control cycle and generates a synchronized control execution list. The linkage execution module extracts control instructions from the synchronous control execution list and distributes them one by one. It dynamically adjusts the equipment operating status to achieve load coordination and balance between the stage crushing section and the dry separation section. It receives feedback information and transmits the feedback information back to the preset mineral processing control terminal. The modules are connected to each other via wired and / or wireless means.

Images