Method and apparatus for controlling the weighing, mixing and packaging of powders

Abstract

The application discloses a powder weighing, mixing and packaging control method and equipment, and relates to the automatic control field of powder weighing, mixing and quantitative packaging. In the process of weighing and blending, double-speed feeding is adopted, the feeding flow is identified on line during the feeding process, the stoppage drop is corrected in real time according to the feeding flow, the total mass of the mixture is obtained according to the net feeding mass of each component, the load signal of the driving motor of the spiral ribbon mixer is collected during mixing, the ratio of the standard deviation to the average value of the load in the moving window is taken as the normalized fluctuation to on-line determine the mixing endpoint, the target of each package is dynamically redistributed according to the remaining available mass and the remaining package number during packaging, and the mixing ready signal triggers the parallel flow line operation of adjacent batches under the material isolation. The application makes the single batch metering accurate and adaptive to the material change, avoids under-mixing and over-mixing, makes the last package no longer underweight, the single package weight in line with the regulations and the material not wasted, and makes the whole line collaborative to improve the production capacity.

Description

Technical Field

[0001] This application relates to the field of automatic control of powder weighing, mixing and quantitative packaging, and in particular to a powder weighing, mixing and packaging control method and equipment. Background Technology

[0002] The production of powder materials includes weighing and batching, mixing, and quantitative packaging. In the weighing and batching stage, a dual-speed feeding method combining coarse and fine feeding is commonly used. Feeding is stopped a certain distance before the fine feeding stage to compensate for any material still en route after the valve is closed. In the mixing stage, ribbon mixers are often used to mix the various components evenly, and the degree of mixing uniformity is used to measure whether the mixing is adequate. In the quantitative packaging stage, the mixed material is divided into several packages according to the set single-package weight, and then sealed, tied, and stacked to complete the packaging.

[0003] In the weighing and batching process, existing technologies often use a fixed or one-time calibrated drop value. However, the drop actually varies with the material flowability, material level, and feeding speed. A fixed drop can cause deviations in batch measurement when the material flowability changes. A single feeding control is also difficult to simultaneously take into account both feeding speed and measurement accuracy. At the same time, when multiple components are added to the same weighing hopper in sequence, if the total weight in the hopper is added directly, the reading of the later added component will include the earlier added component, resulting in double counting and distortion of the batch reference.

[0004] In the mixing process, existing technologies either employ fixed mixing times, offline sampling and testing, or the addition of online spectroscopic detection devices. Fixed mixing times are difficult to bring to a perfectly uniform state when material and batch sizes change, easily leading to under-mixing or over-mixing; offline sampling and testing are subject to lag and disrupt continuous production; online spectroscopic detection devices are expensive and require windows to be opened in the mixing chamber, making them less adaptable to dusty environments and less maintainable.

[0005] In the quantitative packaging process, existing technologies mostly use the total mass of the mixed material as the benchmark and adopt a fixed single-bag target, with the last bag taking over the remaining material. However, the material is subject to wall adhesion, dust removal, and residual loss during its transport from the mixer to the packaging section. Using the total mass of the mixed material as the benchmark will overestimate the available material, leading to deviations in each bag. A fixed single-bag target will cause the filling errors of the preceding bags to accumulate in the last bag, resulting in insufficient quantity in the last bag. Alternatively, in order to prevent insufficient quantity in the last bag, the entire batch may be overfilled, resulting in material waste. It is also difficult for the weight of a single bag to consistently fall within the compliant range.

[0006] In addition, in existing technologies, weighing, batching, mixing and packaging are mostly operated independently according to their own rhythm, lacking coordination triggered by whether the mixing is ready and material isolation between adjacent batches. This causes each link to wait for each other, limits the overall production capacity, and poses a risk of material mixing between adjacent batches.

[0007] Therefore, it is necessary to provide a method and equipment for controlling the weighing, mixing and packaging of powder materials that can maintain accurate measurement, proper mixing and achieve production line coordination when materials and operating conditions change. Summary of the Invention

[0008] This application provides a method for controlling the weighing, mixing, and packaging of powder materials, including the following steps:

[0009] S1 Weighing and Batching: The feeding mechanism feeds the weighing unit at two speeds, coarse feeding and fine feeding. During the feeding process, the stop drop is corrected based on the fine feeding flow rate and the final batch error before feeding is stopped. The total mass of the mixture is obtained by summing the net feeding mass of each component.

[0010] S2 Mixing and Mixing Endpoint Determination: The metered components are fed into a ribbon mixer for mixing. Load signals characterizing the mixing load are collected in real time. The ratio of the standard deviation to the mean of the load signal within the moving window is calculated as the normalized fluctuation. When the normalized fluctuation meets the preset endpoint condition, the mixing is determined to be complete and a mixing ready signal is output.

[0011] S3 Quantitative Filling Packaging: Triggered by the mixing ready signal, the available packaging mass of this batch is determined and fed forward to the packaging section. The filling target weight of each bag is dynamically redistributed according to the remaining available packaging mass and the remaining number of bags. The dual-speed feeding and stop drop correction are used to quantitatively fill each bag. Then, the bags are sealed, tied, and stacked in sequence.

[0012] S4 cycle coordination: After determining that the mixing is complete and the mixed material of this batch has been discharged and isolated from the ribbon mixer, the weighing and batching of the next batch of material is started, so that the weighing, batching, mixing and quantitative filling and packaging are carried out in parallel flow.

[0013] Furthermore, the dual-speed feeding and shutdown drop correction in S1 include:

[0014] S11. During the feeding process, the real-time weight sequence of the weighing unit is fitted with a least-squares straight line within a sliding window, and the slope of the fitted straight line is taken as the instantaneous feeding flow rate:

[0015] ;

[0016] In the formula, Instantaneous feed rate; For subscript From 1 to Summation; This represents the number of sampling points in the sliding window. The sampling point number within the sliding window; For the first The time for each sampling point; This is the average time of each sampling point within the sliding window; For the first Real-time weight of each sampling point; This represents the average weight of each sampling point within the sliding window;

[0017] S12. The stopping drop is calculated in real time using the following formula:

[0018] ;

[0019] In the formula, The drop during shutdown; For the first Feeding station No. The batch flow rate is independent of the residual component; This refers to the feeding station number; This refers to the batch number of the feed material. This is the system's equivalent time lag. The instantaneous feed flow rate obtained in step S11;

[0020] S13. First, feed at a coarse feeding speed. When the fed net weight reaches the difference between the net target weight and the switching margin, switch to a fine feeding speed. The switching margin is the larger of a preset switching allowance and the stop drop. Fine feeding stops when the fed net weight reaches the difference between the net target weight and the stop drop. The fed net weight is the difference between the real-time weight and the initial weight at the start of this feeding.

[0021] Switching conditions: ;

[0022] Shutdown conditions: ;

[0023] In the formula, The real-time weight of the weighing unit; For time; This is the initial weight of the weighing unit at the start of this feeding process; This is the net target weight for this feed; To calculate the larger value; This is the preset switching margin; The aforementioned drop during shutdown;

[0024] This step enables the precision feeding stop point to be adaptively advanced according to the feeding flow rate, and ensures that switching precedes stopping, thus obtaining high-precision metering results for a single batch.

[0025] Furthermore, S1 also includes:

[0026] S14. After each batch of feeding is completed, the difference between the net feeding mass and the net target weight of that batch is taken as the final error of that batch. When the absolute value of the final error of that batch is not greater than the allowable error threshold, the flow-independent residual component of the feeding station is recursively updated and limited according to the product of the learning rate and the final error of that batch; otherwise, it remains unchanged. Each feeding station independently maintains its own flow-independent residual component. The flow-independent residual component of the first batch is determined by the preset value, the calibration value, or the previous production record.

[0027] ;

[0028] when hour, ;otherwise ;

[0029] In the formula, For the first Feeding station No. The final error of the batch; This refers to the feeding station number; This refers to the batch number of the feed material. For the first Feeding station No. The net feed mass of the batch is the difference between the weight of the weighing unit at the end of the batch and the weight at the beginning of the batch. This is the net target weight for this feed; For absolute value operations; This is the allowable threshold for the final error; For the first Feeding station No. The batch flow rate is independent of the residual component; To calculate the smaller value; To calculate the larger value; For the first Feeding station No. The batch flow rate is independent of the residual component; The learning rate; This is the lower bound for the flow-independent residual component; This represents the upper limit of the flow-independent residual component;

[0030] This step enables the shutdown drop to adaptively converge with changes in material flowability, resulting in a steady-state measurement error that approaches zero across batches.

[0031] S15. Add up the net feed masses of each component to obtain the total mass of the mixture:

[0032] ;

[0033] In the formula, The total mass of the mixture; For subscript From 1 to Summation; Group number; Component serial number; For the first The net feed mass of a component is the difference between the weight of the weighing unit at the end of feeding and the weight at the beginning of feeding.

[0034] Furthermore, the calculation of the normalized fluctuation in S2 includes:

[0035] S21. Acquire a load signal characterizing the mixed load, wherein the load signal is obtained by any of the following methods: reading the torque estimate output by the driver of the drive motor of the ribbon mixer; or reading the quadrature-axis current of the drive motor under vector control conditions and converting it into torque according to the motor torque constant; or acquiring the operating current of the drive motor and obtaining the load characterizing signal after correction by no-load current and load calibration; wherein the conversion relationship under vector control conditions is:

[0036] ;

[0037] In the formula, For the first The load signal at each sampling point; The sampling sequence number; The torque constant of the motor; For the first The quadrature-axis current of the drive motor at each sampling point;

[0038] This step obtains the load signal characterizing the load on the mixed bed without adding additional sensors;

[0039] S22. Within a moving window of a set number of sampling points, calculate the mean and standard deviation of the load signal, and take the ratio of the standard deviation to the mean as the normalized fluctuation; when the mean is less than a preset lower limit of the mean, keep the normalized fluctuation at the value of the previous sampling point.

[0040] ;

[0041] when hour, ;

[0042] In the formula, For the first The mean value of the load signal within the moving window at each sampling point; This represents the number of sampling points in the moving window; For subscript from to Summation; The sampling point number within the moving window; For the first The load signal at each sampling point; The sampling sequence number; For the first The standard deviation of the load signal within the moving window at each sampling point; This is the lower limit of the mean. For the first Normalized fluctuation at each sampling point;

[0043] This step generates a normalized fluctuation that characterizes the mixing uniformity and avoids division-by-zero amplification.

[0044] Furthermore, S2 also includes:

[0045] S23. When the normalized fluctuation is not greater than the fluctuation threshold, and the ratio of the absolute value of the difference between the normalized fluctuations of adjacent sampling points to the sampling period is not greater than the rate of change threshold, and this condition is maintained for at least the holding time, the mixing is determined to be complete and the mixing ready signal is output.

[0046] The judgment condition is and And the duration of the above conditions simultaneously is not less than ;

[0047] In the formula, For the first Normalized fluctuation at each sampling point; The threshold for fluctuation; For the first Normalized fluctuation at each sampling point; For absolute value operations; The sampling period; The rate of change threshold; To maintain the duration; The sampling sequence number;

[0048] S24. In the material system to be processed, the normalized fluctuation amount is recorded synchronously at different mixing times and samples are taken in advance. The mixing uniformity is calculated according to the coefficient of variation, where the coefficient of variation is the ratio of the sample standard deviation to the sample mean. The maximum normalized fluctuation amount that makes the coefficient of variation no greater than the target coefficient of variation is selected as the fluctuation amount threshold.

[0049] S25. When the mixing time reaches the preset maximum mixing time and the judgment condition is still not met, an abnormal signal is output and the material discharge is suspended or the process is transferred to manual confirmation.

[0050] This step ensures uniform mixing while avoiding under-mixing, over-mixing, and abnormal retention.

[0051] Furthermore, the dynamic redistribution and quantitative filling in S3 include:

[0052] S31. Determine the usable packaging weight for this batch. The usable packaging weight is obtained by weighing the material in this batch from the transition buffer silo, or by subtracting the calibrated residual conveying loss from the total mixed weight. Determine the total number of packages within a preset per-package allowable weight range, ensuring that the average target filling weight obtained by dividing the usable packaging weight by the total number of packages falls within the per-package allowable weight range. When multiple total packages meet the conditions, select the one that makes the average target filling weight closest to the nominal per-package weight. When no total packages meet the conditions, merge this batch of material into the next batch or recycle it and output an alarm.

[0053] ;

[0054] In the formula, This represents the lower limit of the permissible weight range for a single package. For packaging quality; This represents the total number of packages; This represents the upper limit of the permissible weight range for a single package;

[0055] S32. Before filling each package, the current remaining usable package mass is obtained by subtracting the sum of the measured net filling masses of all previous packages from the usable package mass:

[0056] ;

[0057] In the formula, To fill the first The remaining packaging material is of usable quality before packaging; For packaging quality; For subscript From 1 to Summation; The current package number; The serial number of the already filled package; For the first The measured net fill mass of the package;

[0058] S33. Divide the remaining usable mass of the packaging by the number of remaining packages to obtain the target filling weight of the current package:

[0059] ;

[0060] In the formula, For the first The target weight for filling the package; To fill the first The remaining packaging material is of usable quality before packaging; This represents the total number of packages; The current package number;

[0061] This step distributes the filling error of each bag evenly across the entire batch, constrains the filling deviation of the last bag, and ensures that the weight of a single bag falls within the allowable range.

[0062] S34. Using the target filling weight of the current package as the net target weight of the feed for this package, quantitatively fill the current package according to the dual-speed feeding and stop drop correction. After filling, seal, tie and stack in sequence.

[0063] Further, S4 includes:

[0064] S41. After determining that the mixing is complete and the mixed material of this batch has been discharged and isolated from the ribbon mixer, start the weighing, batching and mixing of the next batch of material, and feed the discharged material to the packaging section for quantitative filling and packaging, so that the weighing, batching, mixing and quantitative filling and packaging of adjacent batches can be carried out in parallel flow.

[0065] The present invention also provides a powder weighing, mixing and packaging control device for realizing a powder weighing, mixing and packaging control method, including a weighing and batching unit, a mixing unit, a transition buffer silo, a packaging execution unit, a feeding control module, a mixing endpoint determination module, a collaborative control module and a collaborative controller;

[0066] The weighing and batching unit includes a weighing hopper, a weighing sensor mounted on the weighing hopper, and a feeding mechanism for feeding materials into the weighing hopper.

[0067] The mixing unit includes a ribbon mixer, a drive motor for driving the ribbon mixer, a driver for supplying power to the drive motor, and a discharge valve disposed at the discharge port of the ribbon mixer.

[0068] The transition buffer silo is connected to the ribbon mixer via the discharge valve. The transition buffer silo is equipped with a weighing device or a material level detection device to buffer and isolate the current batch of materials and to provide the packaging usable quality for the current batch.

[0069] The packaging execution unit includes a packaging scale supplied by the transition buffer silo, a filling and feeding mechanism that supplies material to the packaging scale, and a sealing machine, a sealing machine, and a palletizing robot arranged sequentially downstream of the packaging scale along the material conveying direction.

[0070] The input terminal of the feeding control module is connected to the weighing sensor and the packaging scale, and the control output terminal of the feeding control module is connected to the feeding mechanism and the filling feeding mechanism respectively, for executing the dual-speed feeding and stop drop correction;

[0071] The input terminal of the mixing endpoint determination module is connected to the torque output terminal or current output terminal of the driver, and the mixing ready signal output terminal of the mixing endpoint determination module is connected to the cooperative control module, which is used to calculate the normalized fluctuation amount online and determine the mixing completion.

[0072] The collaborative control module is connected to the transition buffer hopper, the mixing endpoint determination module and the feeding control module respectively, and is used to receive the packaging available quality and the mixing ready signal, and to perform dynamic reallocation of packaging available quality feedforward and single package filling target;

[0073] The feeding control module, the mixing endpoint determination module, and the collaborative control module are integrated into the collaborative controller.

[0074] In a preferred embodiment, the feeding control module includes a weighing acquisition subunit, a flow identification subunit, a drop prediction subunit, and a batch learning subunit connected in sequence. The control output terminal of the drop prediction subunit is connected to the feeding mechanism and the filling feeding mechanism for controlling the switching between coarse feeding and fine feeding, as well as the stopping of fine feeding. The mixing endpoint determination module includes a load acquisition subunit, a fluctuation calculation subunit, and an endpoint criterion subunit connected in sequence. The mixing ready signal output terminal of the endpoint criterion subunit is connected to the collaborative control module.

[0075] In a preferred embodiment, the collaborative control module includes a total mass feedforward subunit and a single-package target redistribution subunit. The total mass feedforward subunit is connected to the weighing device of the transition buffer silo to receive the available package mass. The output of the single-package target redistribution subunit is connected to the feeding control module to provide a single-package filling target. A batch isolation control valve is provided between the discharge port of the ribbon mixer and the transition buffer silo, and / or between the transition buffer silo and the packaging scale. The batch isolation control valve is connected to the collaborative controller. The collaborative controller includes a programmable logic controller (PLC), or includes a PLC and a host computer. The collaborative controller has a cycle scheduling unit. The triggering end of the cycle scheduling unit is connected to the mixing ready signal output of the mixing endpoint determination module, and is used to trigger the weighing and batching of the next batch of materials after mixing is completed and the current batch of materials is isolated.

[0076] The beneficial effects of this invention are as follows:

[0077] This invention employs dual-speed feeding and online correction of stop-fall difference in weighing and batching. During feeding, the real-time weight sequence is fitted with least-squares to identify the instantaneous feed flow rate online. The stop-fall difference is expressed as the sum of the flow-independent residual component and the amount of material in transit, allowing the stop point to adaptively advance with changes in feed flow rate, maintaining consistency with the actual amount in transit even when material flowability changes. After each batch, the flow-independent residual component is recursively learned and limited based on the difference between the net feed mass and the net target weight, causing the stop-fall difference to converge with the slow drift of material flowability, thus stabilizing and reducing measurement errors across batches. Using the net feed mass, rather than the total weight in the weighing hopper, as the benchmark for measurement and judgment ensures that the measurements of each component are not superimposed when multiple components are added sequentially to the same weighing hopper, avoiding double counting caused by the reading of later components including earlier ones, resulting in an accurate and reliable total mixed mass. The switching margin is taken as the larger of the preset switching margin and the real-time stop drop, so that the switching always occurs before the stop, ensuring the feeding speed while avoiding the inertia of coarse feeding from overshooting the target.

[0078] This invention uses the load signal from the helical ribbon agitator shaft to determine the mixing endpoint online during the mixing stage. When materials separate or stratify, the varying bulk density and cohesive force of the material encountered per revolution of the ribbon cause significant fluctuations in the agitation load. As the material becomes more uniform, the load per unit rotation angle becomes more consistent, causing the load fluctuations to decrease and stabilize. Therefore, the ratio of the standard deviation to the mean of the load signal within the moving window can be used as a normalized fluctuation measure to characterize the mixing uniformity online. The signal is taken from the existing torque or current of the driver, eliminating the need for windows in the mixing chamber or specialized sampling or spectral detection devices. Online determination is achieved without adding additional sensors and exhibits good adaptability to dusty environments. Using the ratio of the standard deviation to the mean, rather than the standard deviation itself, ensures that the criterion is unaffected by the material loading amount and the absolute level of the load, providing a consistent discrimination scale for different batches and different materials. The endpoint is determined by the fluctuation amount and its rate of change, and a holding time and abnormal exit are set to ensure that the mixing will not stop too early before it is uniform, nor will it continue to mix ineffectively after it is uniform. At the same time, it avoids under-mixing, over-mixing and abnormal retention, and makes the mixing uniformity of each batch tend to be consistent.

[0079] This invention dynamically redistributes the target amount for each package during the quantitative filling and packaging stage, using the total usable mass of the entire batch as a benchmark. Since powder materials experience losses due to wall adhesion, dust removal, and residual material during transport from the mixer to the packaging section, the amount of material available for packaging may be less than the total mass of the mixed material. This invention uses the usable mass of the package after deducting these losses as the allocation benchmark, avoiding package-by-package deviations caused by overestimating usable material. The total number of packages is determined within a preset allowable weight range for each package, ensuring that the average filling target for each package falls within the compliant range. This guarantees that the weight of each package meets the requirements for quantitative packaging and avoids situations where small batches cannot be packaged. Before filling each package, the filling target for the current package is refreshed by dividing the latest remaining usable mass by the number of remaining packages. This distributes the filling errors of previous packages evenly among the unfilled packages, constraining the deviation of the last package and avoiding underfilling of the last package or material waste caused by overfilling the entire batch to prevent underfilling. The packaging filling and weighing batching use the same feeding control logic, ensuring consistency between the filling accuracy of each package and the batching measurement.

[0080] This invention triggers the streamlined operation of the production line with a mixing-ready signal and uses material isolation as a connection condition. The next batch of weighing, batching, and mixing is only initiated after mixing is determined to be complete and the current batch of material has been discharged and isolated from the ribbon mixer. This ensures that adjacent batches are isolated from each other in terms of material, enabling parallel operations of weighing, batching, mixing, and quantitative filling and packaging while avoiding waiting between adjacent batches or cross-contamination. This allows for seamless line operation and increased production capacity. The above method is implemented collaboratively by a specially designed weighing and batching unit, mixing unit, transition buffer silo, packaging execution unit, as well as a feeding control module, mixing endpoint determination module, collaborative control module, and collaborative controller. Each structure, module, and unit cooperates according to predetermined signal and control relationships, ensuring stable and coordinated operation of the metering, mixing, and packaging processes under unified scheduling. Attached Figure Description

[0081] Figure 1 The above are comparison graphs of the mixing endpoints of Examples 1, 2, and 3 and Comparative Example 2, where (a) is a line graph showing the change of normalized fluctuation over mixing time, and (b) is a line graph showing the change of coefficient of variation over mixing time.

[0082] Figure 2 The charts show the comparison of core indicators between Examples 1, 2, and 3 and Comparative Examples 1, 2, and 3. Among them, (a) is a bar chart of the standard deviation of single-batch measurement error, (b) is a bar chart of the coefficient of variation during shutdown, (c) is a bar chart of the deviation of net content of the last package, and (d) is a bar chart of the range of single package weight within the batch.

[0083] Figure 3 This is a comparison chart of the net content deviation of the final package in Examples 1, 2, and 3 and Comparative Example 3, and a box plot of the distribution of the final package deviation.

[0084] Figure 4This is a comparison chart of single-batch measurement errors of Examples 1, 2, and 3 with Comparative Example 1, and a scatter plot showing the variation of single-batch measurement error with batch.

[0085] Figure 5 This is a schematic diagram of the composition structure of the powder weighing, mixing, and packaging control device of the present invention. Detailed Implementation

[0086] The specific embodiments of the present invention will be described in detail below.

[0087] This embodiment provides a method for controlling the weighing, mixing, and packaging of powder materials, as well as equipment for implementing the method. The method sequentially includes a weighing and batching stage S1, a mixing and mixing endpoint determination stage S2, a quantitative filling and packaging stage S3, and a cycle time coordination stage S4, all uniformly scheduled by a coordination controller. The equipment includes a weighing and batching unit, a mixing unit, a transition buffer silo, a packaging execution unit, a feeding control module, a mixing endpoint determination module, a coordination control module, and a coordination controller. For ease of understanding, the method is described first, followed by the equipment.

[0088] In the weighing and batching stage S1, the feeding mechanism feeds the weighing unit at two speeds: coarse feeding and fine feeding. The powder feeding has inertia; even after the valve closing command is issued, some material remains in transit and continues to fall into the weighing unit. If the machine is stopped with a fixed lead time, variations in material flowability due to humidity, particle size, and batch size will cause single-batch measurement deviations. Therefore, this embodiment corrects the stop drop in real time during the feeding process. The stop drop refers to the weight lead time used to compensate for the material still in transit when feeding is stopped early. S1 includes an instantaneous flow rate identification step S11, a real-time stop drop calculation step S12, a dual-speed feeding switching and stop step S13, a drop batch learning step S14, and a total mixed mass accumulation step S15.

[0089] In step S11, during the feeding process, the real-time weight sequence of the weighing unit is fitted with a least-squares straight line within a sliding window, and the slope of the fitted line is taken as the instantaneous feed flow rate. According to the least-squares principle, a straight line is fitted to the scattered points of the real-time weight over time that minimizes the sum of the squares of the deviations between the ordinates of each point and the fitted value. The resulting slope is the average rate of change of weight over time during that period, which is the instantaneous feed flow rate. The formula for calculating the instantaneous feed flow rate is:

[0090] ;

[0091] In the formula, Instantaneous feed rate (kg / s); For subscript From 1 to Summation; The number of sampling points in the sliding window (dimensionless); The sampling point number within the sliding window; For the first Time (s) for each sampling point; This represents the average time (in seconds) of each sampling point within the sliding window. For the first Real-time weight (kg) of each sampling point; This represents the average weight (kg) of each sampling point within the sliding window. The optimal number of sampling points in the sliding window is chosen to filter out random noise in the weighing signal without significantly lagging behind the actual flow rate change. The flow rate is obtained by fitting the slope of the sliding window, which is less affected by single-point noise compared to differentiating between adjacent points. The resulting instantaneous feed flow rate provides a basis for real-time correction of subsequent shutdown drop.

[0092] In step S12, the shutdown drop is calculated in real time as the sum of the flow-independent residual component and the amount of material in transit. The amount of material in transit is approximately equal to the product of the feed flow rate and the system equivalent lag time. The system equivalent lag time refers to the total time delay from the issuance of the valve closing command to the complete cessation of material falling into the weighing unit, which is determined by the closing action time of the feeding mechanism and the material falling time. The flow-independent residual component refers to the post-valve closing residue and residual material in the mechanism that are independent of the feed flow rate. The formula for calculating the shutdown drop is:

[0093] ;

[0094] In the formula, The drop at which the machine stops (kg); For the first Feeding station No. Batch flow rate is independent of residual weight (kg); This refers to the feeding station number; This refers to the batch number of the feed material. The system equivalent lag time (s); The instantaneous feed flow rate (kg / s) obtained in step S11. The system equivalent lag time can be determined by a single step calibration, i.e., suddenly closing the valve during stable feeding, recording the time delay from the valve closing command to the weighing stabilization and the amount of material falling during that delay, and obtaining the equivalent lag time from the ratio of the amount of material falling to the flow rate at that time. Since the amount of material in transit changes with the feed flow rate, it is expressed as the product of the flow rate and the lag time, so that the shutdown drop still matches the actual amount in transit when the flow rate fluctuates. Combined with the compensation of the flow rate-independent residual component for the subsequent residue after valve closure, the prediction of the shutdown point is closer to the actual value.

[0095] In step S13, material is first fed at a coarse feeding rate to quickly approach the target, then switched to a fine feeding rate to ensure accuracy. To avoid sudden changes in material flow causing a larger stop drop and exceeding the stop point, the switching margin is taken as the larger of the preset switching margin and the real-time stop drop. The switching condition is:

[0096] ;

[0097] The shutdown conditions are:

[0098] ;

[0099] In the formula, The real-time weight (kg) of the weighing unit; Time (s); This is the initial weight (kg) of the weighing unit at the start of this feeding. This is the net target weight (kg) for this feed. To calculate the larger value; Preset switching margin (kg); The stop drop (kg) obtained in step S12. The net mass input is the difference between the real-time weight and the initial weight. Net mass, rather than the total weight in the hopper, is used as the criterion to ensure that the metering of each component does not overlap when multiple components are added sequentially to the same weighing hopper. The larger of the preset switching margin and the stop drop is taken from the switching margin. Switching always occurs before stopping, thus preserving the fine feeding stage and preventing the inertia of coarse feeding from directly overshooting the target. A balance is achieved between speed and accuracy in single-batch metering.

[0100] In step S14, after each batch of feeding is completed and the weighing reading stabilizes, the difference between the net feed mass and the net target weight of that batch is taken as the final error, and the flow-independent residual component of that feeding station is updated batch by batch accordingly. The final error and the recursive update are as follows:

[0101] ;

[0102] when hour, ;

[0103] otherwise ;

[0104] In the formula, For the first Feeding station No. Final batch error (kg); This refers to the feeding station number; This refers to the batch number of the feed material. For the first Feeding station No. The net feed mass of the batch is the difference (kg) between the weight of the weighing unit at the end of the batch and the weight at the beginning of the batch. This is the net target weight (kg) for this feed. For absolute value operations; The allowable threshold (kg) for the final error; For the first Feeding station No. Batch flow rate is independent of residual weight (kg); To calculate the smaller value; To calculate the larger value; For the first Feeding station No. Batch flow rate is independent of residual weight (kg); The learning rate (dimensionless); The lower limit (kg) for flow-independent residual components; This represents the upper limit (kg) of the flow-independent residual component. This update is equivalent to applying an integral correction to the flow-independent residual component. When the measured net feed mass is too high, the residual component is increased to cause the next batch to stop earlier; conversely, it is decreased. The flow-independent residual component gradually approaches the direction that ultimately results in zero error. The learning rate is preferably between 0 and 1. A larger value results in faster convergence but poorer noise immunity, while a smaller value provides stability but slower convergence. Allowable thresholds and upper and lower limits are set to eliminate abnormal batches such as bridging and empty batches and to prevent over-adjustment. Each feeding station maintains its own flow-independent residual component. The batching station and the packaging / filling station do not share the same component. The flow-independent residual component for the first batch is given by a preset value, a calibration value, or the previous production record, ensuring a reasonable stop-loss for the first batch. As batches accumulate, the stop-loss adaptively follows the slow drift of material flow, and the average measurement error across batches tends to zero.

[0105] In step S15, the net feed mass of each component is summed to obtain the total mass of the mixture, calculated as follows:

[0106] ;

[0107] In the formula, Total mass of the mixture (kg); For subscript From 1 to Summation; Components (dimensionless); Component serial number; For the first The net feed mass of a component is the difference (kg) between the weight of the weighing unit at the end and the beginning of the feeding of that component. By adding the net feed masses of each component, duplicate counting caused by the reading of the later component including the previous component when the same weighing hopper is used for cumulative feeding is avoided. The resulting total mass of the mixture provides an accurate batch reference for subsequent mixing and packaging.

[0108] In the mixing and mixing endpoint determination stage S2, the metered components are fed into a ribbon mixer for mixing. The ribbon mixer consists of a U-shaped cylinder and coaxial forward and reverse ribbon blades. The outer ribbon pushes the material from both ends to the middle, while the inner ribbon pushes the material from the middle to both ends, forming convective mixing. Whether the mixing is in place is measured by the uniformity of the mixture. Traditional methods either use a fixed mixing time, offline sampling and testing, or online spectrometers such as near-infrared or Raman spectrometers. The former is prone to under-mixing or over-mixing when the material and batch size change, while the latter is expensive and has high requirements for dust environment maintenance. This embodiment utilizes the characteristic that the load on the ribbon stirring shaft tends to be uniform and stable with the material distribution, and judges the mixing endpoint online by the degree of fluctuation of the load signal. When the material stratifies or segregates, the bulk density and cohesive force of the material encountered by the ribbon per revolution are different, and the load on the stirring shaft fluctuates greatly; as mixing progresses, the material distribution tends to be uniform, the load per unit rotation angle tends to be consistent, and the load fluctuation decreases and tends to be stable. S2 includes the load signal acquisition step S21, the normalized fluctuation calculation step S22, the mixed endpoint determination step S23, the fluctuation threshold calibration step S24, and the abnormal exit step S25.

[0109] In step S21, a load signal characterizing the hybrid load is acquired. Considering the different relationships between current and torque under different driving methods, the load signal is obtained through one of three methods. The first method is to directly read the torque estimate output by the driver of the ribbon mixer's drive motor. The second method is to read the quadrature-axis current of the drive motor under vector control conditions and convert it into torque. Since the electromagnetic torque of the permanent magnet synchronous motor is approximately proportional to the quadrature-axis current, and the proportionality coefficient is the motor torque constant, the conversion formula is:

[0110] ;

[0111] In the formula, For the first Load signal (N·m) at each sampling point; The sampling sequence number; is the motor torque constant (N·m / A); For the first The first method involves sampling the quadrature-axis current (A) of the drive motor. The second method involves collecting the operating current of the drive motor and correcting it with no-load current and load calibration to obtain a signal characterizing the load. This method is suitable for situations where the line current and torque of a common asynchronous frequency converter motor are not in a simple proportional relationship. All three methods utilize the existing signals of the driver, eliminating the need to open windows in the mixing chamber or install special sensors, thus providing good adaptability to dusty environments.

[0112] In step S22, the mean and standard deviation of the load signal are calculated within a moving window of a set number of sampling points, and the ratio of the standard deviation to the mean is taken as the normalized fluctuation. The moving block standard deviation is a known method in process analysis technology for determining the endpoint of mixing. Its principle is that as mixing progresses, the standard deviation of the signal decreases and tends to stabilize within the sliding data block. This embodiment uses this method to characterize the load signal of a mixed load. The load mean, standard deviation, and normalized fluctuation within the window are as follows:

[0113] ;

[0114] when hour, ;

[0115] In the formula, For the first The mean value of the load signal within the moving window at each sampling point (N·m); The number of sampling points in the moving window (dimensionless); For subscript from to Summation; The sampling point number within the moving window; For the first Load signal (N·m) at each sampling point; The sampling sequence number; For the first The standard deviation (N·m) of the load signal within the moving window at each sampling point; The lower limit of the mean (N·m); For the first The normalized fluctuation (dimensionless) of each sampling point is calculated as the ratio of the standard deviation to the mean, rather than the standard deviation itself. This ensures the criterion is unaffected by the material loading and absolute load level, providing a consistent dimension and discrimination scale for different batches and materials. The number of sampling points in the moving window is preferably chosen to cover an integer number of rotation cycles of the stirring shaft, thus reducing the interference of periodic load fluctuations within a single rotation. When the mean load within the window is below the lower limit of the mean, the normalized fluctuation is maintained at the value of the previous sampling point to avoid the denominator approaching zero at the end of the discharge phase or near no-load conditions, which would amplify the ratio.

[0116] In step S23, when the normalized fluctuation is not greater than the fluctuation threshold, and the ratio of the absolute value of the difference between the normalized fluctuations of adjacent sampling points to the sampling period is not greater than the rate of change threshold, and this condition is maintained for at least the holding time, the mixing is determined to be complete and a mixing ready signal is output. The determination condition is:

[0117] ;

[0118] ;

[0119] And the duration of the above conditions simultaneously is not less than ;

[0120] In the formula, For the first Normalized fluctuation (dimensionless) at each sampling point; The fluctuation threshold (dimensionless). For the first Normalized fluctuation (dimensionless) at each sampling point; For absolute value operations; The sampling period is (s). The rate of change threshold (1 / s); To maintain duration (s); This is the sampling sequence number. Simultaneously, the fluctuation amount must be below a threshold and its rate of change must be sufficiently small to confirm that the mixing has entered a plateau rather than a brief dip. Then, the holding time is used to filter out occasional disturbances, ensuring stable and reliable endpoint determination. The sampling period is preferably chosen to be on the order of magnitude that can follow load changes without introducing excessive computational burden, typically between several milliseconds and tens of milliseconds.

[0121] In step S24, the fluctuation threshold is calibrated. Normalized fluctuations are recorded synchronously at different mixing times within the material system to be processed, and samples are taken. The mixing uniformity is calculated using the coefficient of variation (COP), which is the ratio of the sample standard deviation to the sample mean and is a commonly used indicator in evaluating mixing uniformity. The maximum normalized fluctuation that ensures the COP is no greater than the target COP is selected as the fluctuation threshold. This calibration establishes a correspondence between the fluctuation derived from the load signal and the uniformity obtained from sampling and analysis, making the fluctuation threshold correspond to the target uniformity. The calibration can be reused after the first calibration; recalibration is required when changing the material system.

[0122] In step S25, if the mixing time reaches the preset maximum mixing time but still fails to meet the judgment condition of step S23, an abnormal signal is output and material discharge is paused or manual confirmation is initiated. When the powder state is abnormal or a certain component is missing, the normalized fluctuation may not decrease for a long time. Setting a maximum mixing time serves as a safety net to prevent ineffective mixing from continuously occupying the production line and prompting manual investigation. The maximum mixing time is preferably selected as several times the time required for normal mixing to reach the endpoint, so as to detect abnormalities in a timely manner without accidental triggering.

[0123] In the quantitative filling and packaging stage S3, triggered by a mixing-ready signal, the batch of material is divided into several bags. The powder is transported from the mixer to the packaging section, resulting in wall adhesion, dust removal, and residual losses; therefore, the mass of usable material for packaging is not equal to the total mass of the mixed material. Simultaneously, the weight of each bag must fall within the allowable net content range for quantitative packaging. S3 includes steps S31 (determining usable packaging mass and total number of bags), S32 (updating remaining usable packaging mass), S33 (dynamic reallocation of target mass per bag), and S34 (quantitative filling of a single bag).

[0124] In step S31, the usable packaging weight for this batch is first determined by weighing the material in the transition buffer silo, or by subtracting the calibrated residual loss during conveying from the total mixed weight. Then, the total number of packages is determined within a preset per-package allowable weight range, ensuring that the average target filling weight obtained by dividing the usable packaging weight by the total number of packages falls within this range. The constraint is:

[0125] ;

[0126] In the formula, This is the lower limit (kg) of the permissible weight range for a single package. Packaging usable weight (kg); Total number of packages (dimensionless); This represents the upper limit (kg) of the permissible weight range for a single package. When multiple packages meet the constraints, the one that makes the average target filling weight closest to the nominal single package weight is selected. When no packages meet the constraints, the current batch of material is either incorporated into the next batch or recalled, and an alarm is triggered. Determining the total number of packages based on the average target filling weight falling within the permissible range, rather than simply dividing the total mass of the mixture by the nominal single package weight and rounding down, ensures that the weight of each package is within the compliant range and avoids the total number of packages being zero when the batch size is too small. The permissible weight range for a single package is given according to the applicable quantitative packaging net content measurement rules or the company's internal control limits.

[0127] In step S32, before filling each package, the remaining usable package mass is obtained by subtracting the sum of the measured net filling masses of all previous packages from the usable package mass. The calculation formula is as follows:

[0128] ;

[0129] In the formula, To fill the first The remaining usable weight of the packaging before packaging (kg); Packaging usable weight (kg); For subscript From 1 to Summation; The current package number; The serial number of the already filled package; For the first The measured net fill weight (kg) of each bag. The remaining amount is deducted from the measured net fill weight of each bag on a bag-by-bag basis, rather than estimated based on the nominal value, so that the remaining baseline before each bag always matches the actual amount.

[0130] In step S33, the remaining usable mass of the packaging is divided by the number of remaining packages to obtain the target filling weight of the current package. The calculation formula is:

[0131] ;

[0132] In the formula, For the first Target filling weight of the package (kg); To fill the first The remaining usable weight of the packaging before packaging (kg); Total number of packages (dimensionless); This is the current package sequence number. The number of bags that have not yet been filled is the remaining number of bags. Since the target is redefined by dividing the latest remaining amount by the number of remaining bags after each bag is filled, the filling error of the previous bags is distributed to the unfilled bags. The last bag no longer simply carries the surplus, and the situation of insufficient amount in the last bag or excessive amount in the whole batch to prevent insufficient amount is suppressed, and the overall feed surplus of the batch decreases accordingly.

[0133] In step S34, the target filling weight of the current package is used as the net target weight for feeding this package. The packaging scale is filled according to the dual-speed feeding and stop-drop correction methods described in steps S11 to S13. Then, in step S14, the flow-independent residual component of the packaging and filling station is learned and updated package by package. After filling, the package is sealed by a sealing machine, tied by a sealing machine, and palletized by a palletizing robot. The metering and feeding of the packaging and filling process uses the same control logic, and the filling accuracy of a single package is consistent with that of the feeding process.

[0134] In the cycle time coordination stage S4, the mixing ready signal triggers the current batch's discharge and packaging while simultaneously driving the production line's streamlined operation. In step S41, after determining that mixing is complete and that the mixed material of this batch has been discharged and isolated from the ribbon mixer, the weighing, batching, and mixing of the next batch of material are initiated. The discharged material is then fed to the packaging section for quantitative filling and packaging, allowing the weighing, batching, mixing, and quantitative filling and packaging of adjacent batches to proceed in parallel. Using the discharge and isolation of material as a prerequisite for initiating the next batch avoids adjacent batches waiting for each other or material mixing in the mixer or packaging section, compressing the waiting time of each section and ensuring seamless connection of the entire production line's cycle time.

[0135] The following describes the equipment used to implement the above method, such as... Figure 5 As shown, a powder weighing, mixing and packaging control device of the present invention includes a weighing and batching unit, a mixing unit, a transition buffer silo, a packaging execution unit, a feeding control module, a mixing endpoint determination module, a collaborative control module and a collaborative controller.

[0136] The weighing and batching unit includes a weighing hopper, a weighing sensor, and a feeding mechanism. The weighing sensor is mounted on the support of the weighing hopper, which receives material fed by the feeding mechanism. The feeding mechanism includes a screw conveyor and a vacuum conveyor, respectively adaptable to materials with different flowability and feeding methods. The output of the weighing sensor is connected to the input of the feeding control module, sending the real-time weight of the weighing hopper to the feeding control module as feedback for feeding control, thus providing a closed-loop basis for batching and metering.

[0137] The mixing unit includes a ribbon mixer, a drive motor, a driver, and a discharge valve. The drive motor drives the stirring shaft of the ribbon mixer, the driver supplies power to the drive motor and outputs torque or current signals, and the discharge valve is located at the discharge port of the ribbon mixer. The ribbon mixer undertakes the convective mixing of materials, the torque or current signal output by the driver reflects the load on the stirring shaft, providing a signal source for online determination of the mixing endpoint, and the discharge valve controls the timing of the discharge of the current batch of materials.

[0138] The transition buffer silo is connected to the ribbon mixer via a discharge valve and is located between the ribbon mixer and the packaging scale. The transition buffer silo is equipped with a weighing device or a level detection device. The transition buffer silo receives the material discharged from the discharge valve, separating this batch from the next batch in terms of material, ensuring that the mixing of the next batch and the packaging of the current batch do not interfere with each other. Its weighing device or level detection device is used to obtain the mass of the material transferred to the packaging section, providing an accurate material baseline for the allocation of single-package targets.

[0139] The packaging execution unit includes a packaging scale, a filling and feeding mechanism, a sealing machine, a sealing machine, and a palletizing robot. The packaging scale is supplied with material from a transition buffer hopper, and the filling and feeding mechanism feeds material to the packaging scale. The sealing machine, sealing machine, and palletizing robot are sequentially positioned downstream of the packaging scale along the material conveying direction. The packaging scale handles single-package measurement, the filling and feeding mechanism performs single-package filling, the sealing machine seals the packages, the sealing machine seals the packages, and the palletizing robot palletizes the packages, forming a complete packaging process from measurement to stacking.

[0140] The input terminal of the feeding control module is connected to the weighing sensor and the packaging scale, while the control output terminal is connected to the feeding mechanism and the filling feeding mechanism, respectively. The feeding control module is used to achieve dual-speed feeding for both batching and packaging / filling, as well as stop drop correction. The feeding control module includes a weighing acquisition subunit, a flow identification subunit, a drop prediction subunit, and an inter-batch learning subunit, connected in sequence. The weighing acquisition subunit acquires real-time weight, the flow identification subunit obtains the instantaneous feeding flow rate, the drop prediction subunit obtains the stop drop based on the flow rate and residual component, and the inter-batch learning subunit corrects the residual component after each batch. The control output terminal of the drop prediction subunit is connected to the feeding mechanism and the filling feeding mechanism to control the switching between coarse and fine feeding, as well as the stopping of fine feeding. The batching and packaging / filling processes share the same feeding control module, and their metering logic is consistent, with the same single-pack filling accuracy as the batching process.

[0141] The input terminal of the mixing endpoint determination module is connected to the torque output terminal or current output terminal of the driver, and the mixing ready signal output terminal of the mixing endpoint determination module is connected to the cooperative control module. The mixing endpoint determination module is used to determine online whether mixing is complete based on the load signal. The mixing endpoint determination module includes a load acquisition subunit, a fluctuation calculation subunit, and an endpoint criterion subunit connected in sequence. The load acquisition subunit acquires the load signal characterizing the mixing load, the fluctuation calculation subunit obtains the normalized fluctuation value based on this signal, and the endpoint criterion subunit determines the completion of mixing based on the fluctuation value and its changes, and outputs a mixing ready signal. The mixing endpoint determination module operates using the existing signals of the driver, eliminating the need for a dedicated sensor on the mixing chamber, and exhibits good adaptability to dusty environments.

[0142] The collaborative control module is connected to the transition buffer silo, the mixing endpoint determination module, and the feeding control module. The collaborative control module includes a total mass feedforward subunit and a single-package target redistribution subunit. The total mass feedforward subunit is connected to the weighing device of the transition buffer silo to receive the available packaging mass. The output of the single-package target redistribution subunit is connected to the feeding control module to provide the single-package filling target. Based on the available packaging mass and filling status of the current batch, the collaborative control module allocates filling targets to each package, ensuring that the filling error of each package is evenly distributed and the deviation of the last package is constrained.

[0143] The batch isolation control valve is located between the discharge port of the ribbon mixer and the transition buffer hopper, and / or between the transition buffer hopper and the packaging scale. The batch isolation control valve is connected to the co-controller. Under the control of the co-controller, the batch isolation control valve opens and closes to keep adjacent batches isolated in terms of material, providing material isolation for parallel flow operations of weighing, mixing, and packaging.

[0144] The collaborative controller includes a programmable logic controller (PLC), or a PLC and a host computer. The feeding control module, mixing endpoint determination module, and collaborative control module are integrated within the collaborative controller. The collaborative controller has a cycle time scheduling unit, whose trigger terminal is connected to the mixing ready signal output terminal of the mixing endpoint determination module. After mixing is complete and the current batch of materials is isolated, the system triggers the weighing and batching of the next batch. The collaborative controller provides unified scheduling for all units and modules, ensuring that batching, mixing, and packaging are sequentially connected and material-isolated, maintaining stability throughout the entire production line under parallel operation.

[0145] The following specific embodiments, comparative examples, and experimental examples verify the implementation results of the powder weighing, mixing, and packaging control method and equipment of the present invention. In the following description:

[0146] EPS refers to expandable polystyrene. The coefficient of variation (CV) is the ratio of the standard deviation to the mean of a set of sampled measurements, expressed as a percentage. It measures the uniformity of mixing; a smaller value indicates more homogeneous material. Normalized fluctuation (NFD) is the ratio of the standard deviation to the mean of the load signal within a moving window. It is used to characterize mixing uniformity online; its value decreases as the material becomes more homogeneous. Stoppage margin refers to the weight advance required to compensate for material already in transit. Packaging usable mass refers to the actual mass of the batch of material available for packaging after deducting conveyor residues and dust removal losses. Moving block standard deviation is a known method in process analysis technology for calculating the standard deviation within a sliding data block to determine the mixing endpoint. The permissible shortage quantity refers to the maximum amount by which the actual net content of a single item is allowed to be less than the labeled net content. It is determined according to the labeled net content tiers: 1.5% for labeled net content between 1000g and 10000g, 150g for 10000g and 15000g, and 1.0% for 15000g and 50000g. Therefore, a 5kg item is approximately 0.075kg, a 25kg item is approximately 0.25kg, and a 50kg item is approximately 0.50kg.

[0147] Example 1 illustrates the implementation process of this invention using 25kg quantitative packaging of EPS powder as an example. In the weighing and batching step S1, a screw feeder and a vacuum feeder feed material to the weighing hopper at dual speeds. In S11, the real-time weight sequence of the symmetrical weighing hopper is fitted with a least-squares straight line within a sliding window, and the slope of the fitted line is taken as the instantaneous feeding flow rate. In S12, the product of the instantaneous feeding flow rate and the system's equivalent lag time is added to the flow rate-independent residual component to obtain the real-time shutdown drop. In S13, the coarse feeding speed is first used to approach the target, then switched to a fine feeding speed, and the machine stops when the fed net mass reaches the difference between the net target weight and the shutdown drop. The switching margin is taken as the larger of the preset switching margin and the shutdown drop to ensure that the machine switches before stopping. In S14, after each batch, the difference between the net fed mass and the net target weight is taken as the final error of the batch, and the flow rate-independent residual component of the station is recursively updated and limited according to the learning rate. In S15, the total mass of each component's net feed is accumulated to obtain the total mass of the mixture. In the mixing and mixing endpoint determination step S2, the material enters the ribbon mixer. In S21, the load signal is obtained by reading the torque or current from the driver. In S22, the normalized fluctuation is obtained by calculating the ratio of the standard deviation to the mean of the load signal within the moving window. In S23, mixing is determined to be complete when the normalized fluctuation is not greater than the threshold and its rate of change is not greater than the set value and remains at the set time. In S24, the fluctuation threshold is pre-calibrated using the coefficient of variation method. In S25, the maximum mixing time is set to exit abnormally. In the quantitative filling and packaging step S3, S31 weighs the usable packaging mass from the transition buffer silo to determine the total number of packages within the allowable weight range. S32 updates the remaining usable mass of each package with the measured net filling mass. S33 divides the remaining usable mass by the remaining number of packages to obtain the current package target. S34 fills according to S11 to S14, then seals, ties, and stacks the packages. In the cycle coordination step S4, after the current batch of materials is discharged and isolated, the next batch is started, and the three steps are carried out in parallel flow. In this embodiment, a medium learning rate is used, and the moving window covers an integer number of rotation cycles of the stirring shaft.

[0148] Example 2 uses a 50kg quantitative packaging of coating powder as an example. The coating powder has poor flowability and mixes slowly, so the learning rate is set to a smaller value, while the moving window and maximum mixing time are set to a larger value. The total number of packages is less than in Example 1, and the remaining steps are the same as in Example 1.

[0149] Example 3 uses a 5kg quantitative packaging of chemical powder as an example. Since the chemical powder has good flowability and mixes quickly, a relatively large learning rate is used, while the sampling period and moving window are smaller. The total number of packages is more than in Example 1, and the remaining steps are the same as in Example 1.

[0150] Comparative Example 1: The material used in this comparative example is the same as that in Example 1, which is 25 kg of EPS powder. The batching and packaging filling adopts a fixed drop dual-speed feeding method. The online flow identification in S11 is not performed. The stop drop in S12 is a fixed value and does not include the flow item. The batch learning in S14 is not performed. The rest is the same as in Example 1.

[0151] Comparative Example 2: The material used in this comparative example is the same as that in Example 1, which is 25 kg of EPS powder. The mixing process uses a fixed mixing time, and the online load fluctuation determination in S21 to S25 of this invention is not performed. The machine stops and the material is discharged once the preset fixed time is reached. The rest is the same as in Example 1.

[0152] Comparative Example 3 uses the same material as Example 1, namely 25 kg of EPS powder. Its packaging is based on the total mass of the mixture, using a fixed target weight per package and taking the remaining weight from the last package. The usable weight and allowable weight range for packaging as described in S31 of this invention are not implemented, and the dynamic redistribution in S32 and S33 is not performed. The rest is the same as in Example 1.

[0153] Experimental Example 1 uses Examples 1, 2, and 3, along with Comparative Example 2, to verify the determination of the mixing endpoint in S21 to S25 of the present invention. During the mixing process, the coefficient of variation of each material decreases approximately according to a first-order law with mixing time, and the normalized fluctuation decreases synchronously with the coefficient of variation. Examples 1, 2, and 3 all end when the normalized fluctuation decreases to the threshold and enters a plateau; the coefficient of variation at shutdown is approximately 4.5%, and the shutdown times are approximately 8.1 min, 13.4 min, and 5.4 min, respectively. Comparative Example 2 uses a fixed mixing time of 6 min and stops using the same EPS powder as in Example 1; the coefficient of variation at shutdown is approximately 6.6%.

[0154] The results are as follows Figure 1 As shown. Figure 1 In (a), the detection endpoints for Examples 1, 2, and 3 are marked with circles, squares, and triangles, respectively. From Figure 1 As can be seen in (a), the normalized fluctuation of all three embodiments decreases with mixing time, and each drops to the same threshold at its respective detection endpoint. The fixed mixing time of Comparative Example 2 is marked with a vertical line. At this time, the normalized fluctuation of Example 1 and Example 2 (which mixes more slowly) is still above the threshold and is undermixed, while the normalized fluctuation of Example 3 (which mixes more quickly) is below the threshold and has passed its mixing endpoint. Figure 1 As can be seen in (b), the coefficient of variation of the three embodiments at the time of shutdown is consistent at about 4.5% and not greater than the target of 5%. The coefficient of variation of Comparative Example 2 at a fixed 6 min is about 6.6%, greater than 5%, and is undermixed. At the same time, the shutdown time of the three embodiments is significantly different, indicating that no fixed mixing time can be adapted to the three materials at the same time.

[0155] The present invention employs a method S22 that uses the ratio of the standard deviation of the load to the mean, and S23 that uses the fluctuation amount and its rate of change to determine the endpoint. This method can stably stop at the same target uniformity on different materials, indicating that the ribbon agitator load tends to stabilize after the material becomes uniform, and the normalized fluctuation amount decreases to the threshold. Stopping the machine at this point corresponds to the target uniformity. The mechanism is that the increased mixing uniformity reduces the dispersion of the unit rotation load, thus indirectly reflecting the uniformity through load fluctuation. This method can determine the uniformity online without sampling or a spectrometer, thereby avoiding the undermixing and overmixing caused by the fixed time on different materials in Comparative Example 2.

[0156] Experimental Example 2 uses Examples 1, 2, and 3, and Comparative Examples 1, 2, and 3, respectively, to verify the beneficial effects of S11 to S14, S21 to S25, and S31 to S34 of the present invention. The relevant indicators were obtained from statistical analysis of multiple batches of production in each example. Figure 2 As can be seen in (a), the standard deviations of single-batch measurement errors for Examples 1, 2, and 3 are approximately 0.04 kg, 0.06 kg, and 0.02 kg, respectively, while that for Comparative Example 1 is approximately 0.20 kg. The errors in Examples 1 and 3 are significantly smaller than those in Comparative Example 1. Figure 2 As can be seen in (b), the coefficients of variation at shutdown for Examples 1, 2, and 3 were all approximately 4.5% and not greater than the 5% target line, while for Comparative Example 2 it was approximately 6.6% and greater than 5%. From... Figure 2 As can be seen from (c), the average deviation of the final package net content in Examples 1, 2, and 3 is close to 0, while the final package shortage in Comparative Example 3 is approximately 0.40 kg, and the allowable shortage for specifications exceeding 25 kg is 0.25 kg. From Figure 2 As can be seen from (d), the intra-batch single-package weight range for Examples 1, 2, and 3 is approximately 0.13 kg, 0.15 kg, and 0.09 kg, respectively, while that for Comparative Example 3 is approximately 0.45 kg.

[0157] This invention outperforms the corresponding comparative example in four indicators: measurement error, shutdown uniformity, last package deviation, and single package range. S11 identifies the flow rate online and corrects the shutdown drop through S13 and cross-batch learning through S14, thereby reducing the measurement error of a single batch. S22 and S23 determine the endpoint based on load fluctuation, ensuring that the shutdown uniformity meets the standard. S31 uses the usable packaging quality as a benchmark and dynamically redistributes it through S32 and S33, thereby reducing the last package deviation and single package range.

[0158] Experimental Example 3 uses Examples 1, 2, and 3, along with Comparative Example 3, to verify the packaging control of S31 to S34 of the present invention. Deviations in the final net content of each package were statistically analyzed across multiple production batches. The results are as follows: Figure 3 As shown. From Figure 3It can be seen that the final package deviation boxes of Examples 1, 2, and 3 are concentrated around 0, with very small dispersion. They fall within the allowable shortage amounts of 0.25kg, 0.50kg, and 0.075kg for the 25kg, 50kg, and 5kg specifications, respectively. The center of the final package deviation box of Comparative Example 3 is approximately -0.40kg, that is, the final package shortage is about 0.40kg. The shortage amount of most batches exceeds the allowable shortage amount of 0.25kg for the 25kg specification and is judged as defective due to shortage.

[0159] In this invention, S32 updates the remaining usable mass based on the measured net filling mass, and S33 refreshes the current package target by dividing the remaining usable mass by the number of remaining packages. This can distribute the filling error evenly to each unfilled package, and use the usable package mass after deducting losses in S31 as the benchmark, so that the last package no longer bears the loss of the entire batch alone. In contrast, Comparative Example 3 uses the total mass of the mixture as the benchmark and the last package has a surplus. The residue and dust removal loss during the transportation process are concentrated in the last package, causing a systemic shortage. However, this invention determines the target for each package with the latest remaining quantity and the number of remaining packages, so that each package only bears a small and uniform loss component, thereby avoiding the last package shortage exceeding the standard.

[0160] Experimental Example 4 uses Examples 1, 2, and 3, along with Comparative Example 1, to verify the feeding control of S11 to S14 of the present invention. A 30th batch of material replacement is included to simulate a sudden change in material flowability. The results are as follows: Figure 4 As shown. From Figure 4 It can be seen that the measurement error scatter points of Examples 1, 2 and 3 are always densely distributed around 0. After the material is changed, they converge again after only a few batches. The measurement error scatter points of Comparative Example 1 are more dispersed, and after the 30th batch of material is changed, the overall shift is about 0.30 kg and does not fall back.

[0161] This invention minimizes single-batch metering errors and adapts to material changes, whereas Comparative Example 1 suffers from large metering errors and an unavoidable bias after material changes. In S11, the instantaneous flow rate is identified online during feeding. In S13, the shutdown drop is corrected in real-time based on the flow rate and the system's equivalent lag time. In S14, after each batch, the residual component unrelated to the flow rate is recursively learned based on the final error, allowing the shutdown drop to follow changes in material flowability. In contrast, Comparative Example 1 has a fixed shutdown drop and no inter-batch learning; after material changes, the drop does not match the actual value, resulting in a persistent bias.

[0162] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for controlling the weighing, mixing, and packaging of powder materials, characterized in that, Includes the following steps: S1 Weighing and Batching: The feeding mechanism feeds the weighing unit at two speeds, coarse feeding and fine feeding. During the feeding process, the stop drop is corrected based on the fine feeding flow rate and the final batch error before feeding is stopped. The total mass of the mixture is obtained by summing the net feeding mass of each component. S2 Mixing and Mixing Endpoint Determination: The metered components are fed into a ribbon mixer for mixing. Load signals characterizing the mixing load are collected in real time. The ratio of the standard deviation to the mean of the load signal within the moving window is calculated as the normalized fluctuation. When the normalized fluctuation meets the preset endpoint condition, the mixing is determined to be complete and a mixing ready signal is output. S3 Quantitative Filling Packaging: Triggered by the mixing ready signal, the available packaging mass of this batch is determined and fed forward to the packaging section. The filling target weight of each bag is dynamically redistributed according to the remaining available packaging mass and the remaining number of bags. The dual-speed feeding and stop drop correction are used to quantitatively fill each bag. Then, the bags are sealed, tied, and stacked in sequence. S4 cycle coordination: After determining that the mixing is complete and the mixed material of this batch has been discharged and isolated from the ribbon mixer, the weighing and batching of the next batch of material is started, so that the weighing, batching, mixing and quantitative filling and packaging are carried out in parallel flow.

2. The method for controlling the weighing, mixing, and packaging of powders according to claim 1, characterized in that, The dual-speed feeding and shutdown drop correction in S1 include: S11. During the feeding process, the real-time weight sequence of the weighing unit is fitted with a least-squares straight line within a sliding window, and the slope of the fitted straight line is taken as the instantaneous feeding flow rate: ； In the formula, Instantaneous feed rate; For subscript From 1 to Summation; This represents the number of sampling points in the sliding window. The sampling point number within the sliding window; For the first The time for each sampling point; This is the average time of each sampling point within the sliding window; For the first Real-time weight of each sampling point; This represents the average weight of each sampling point within the sliding window; S12. The stopping drop is calculated in real time using the following formula: ； In the formula, The drop during shutdown; For the first Feeding station No. The batch flow rate is independent of the residual component; This refers to the feeding station number; This refers to the batch number of the feed material. This is the system's equivalent time lag. The instantaneous feed flow rate obtained in step S11; S13. First, feed at a coarse feeding speed. When the fed net weight reaches the difference between the net target weight and the switching margin, switch to a fine feeding speed. The switching margin is the larger of a preset switching allowance and the stop drop. Fine feeding stops when the fed net weight reaches the difference between the net target weight and the stop drop. The fed net weight is the difference between the real-time weight and the initial weight at the start of this feeding. Switching conditions: ; Shutdown conditions: ; In the formula, The real-time weight of the weighing unit; For time; This is the initial weight of the weighing unit at the start of this feeding process; This is the net target weight for this feed; To calculate the larger value; This is the preset switching margin; The aforementioned drop during shutdown; This step enables the precision feeding stop point to be adaptively advanced according to the feeding flow rate, and ensures that switching precedes stopping, thus obtaining high-precision metering results for a single batch.

3. The method for controlling the weighing, mixing, and packaging of powders according to claim 2, characterized in that, S1 further includes: S14. After each batch of feeding is completed, the difference between the net feeding mass and the net target weight of that batch is taken as the final error of that batch. When the absolute value of the final error of that batch is not greater than the allowable error threshold, the flow-independent residual component of the feeding station is recursively updated and limited according to the product of the learning rate and the final error of that batch; otherwise, it remains unchanged. Each feeding station independently maintains its own flow-independent residual component. The flow-independent residual component of the first batch is determined by the preset value, the calibration value, or the previous production record. ； when hour, ;otherwise In the formula, For the first Feeding station No. The final error of the batch; This refers to the feeding station number; This refers to the batch number of the feed material. For the first Feeding station No. The net feed mass of the batch is the difference between the weight of the weighing unit at the end of the batch and the weight at the beginning of the batch. This is the net target weight for this feed; For absolute value operations; This is the allowable threshold for the final error; For the first Feeding station No. The batch flow rate is independent of the residual component; To calculate the smaller value; To calculate the larger value; For the first Feeding station No. The batch flow rate is independent of the residual component; The learning rate; This is the lower bound for the flow-independent residual component; This represents the upper limit of the flow-independent residual component; This step enables the shutdown drop to adaptively converge with changes in material flowability, resulting in a steady-state measurement error that approaches zero across batches. S15. Add up the net feed masses of each component to obtain the total mass of the mixture: ； In the formula, The total mass of the mixture; For subscript From 1 to Summation; Group number; The component number; For the first The net feed mass of a component is the difference between the weight of the weighing unit at the end of feeding and the weight at the beginning of feeding.

4. The method for controlling the weighing, mixing, and packaging of powders according to claim 1, characterized in that, The calculation of normalized fluctuation in S2 includes: S21. Acquire a load signal characterizing the mixed load, wherein the load signal is obtained by any of the following methods: reading the torque estimate output by the driver of the drive motor of the ribbon mixer; or reading the quadrature-axis current of the drive motor under vector control conditions and converting it into torque according to the motor torque constant; or acquiring the operating current of the drive motor and obtaining the load characterizing signal after correction by no-load current and load calibration; wherein the conversion relationship under vector control conditions is: ； In the formula, For the first The load signal at each sampling point; The sampling sequence number; The torque constant of the motor; For the first The quadrature-axis current of the drive motor at each sampling point; This step obtains the load signal characterizing the load on the mixed bed without adding additional sensors; S22. Within a moving window of a set number of sampling points, calculate the mean and standard deviation of the load signal, and take the ratio of the standard deviation to the mean as the normalized fluctuation; when the mean is less than a preset lower limit of the mean, keep the normalized fluctuation at the value of the previous sampling point. ； when hour, In the formula, For the first The mean value of the load signal within the moving window at each sampling point; This represents the number of sampling points in the moving window; For subscript from to Summation; The sampling point number within the moving window; For the first The load signal at each sampling point; The sampling sequence number; For the first The standard deviation of the load signal within the moving window at each sampling point; This is the lower limit of the mean; For the first Normalized fluctuation at each sampling point; This step generates a normalized fluctuation that characterizes the mixing uniformity and avoids division-by-zero amplification.

5. The method for controlling the weighing, mixing, and packaging of powders according to claim 4, characterized in that, S2 further includes: S23. When the normalized fluctuation is not greater than the fluctuation threshold, and the ratio of the absolute value of the difference between the normalized fluctuations of adjacent sampling points to the sampling period is not greater than the rate of change threshold, and this condition is maintained for at least the holding time, the mixing is determined to be complete and the mixing ready signal is output. The judgment condition is and And the duration of the above conditions simultaneously is not less than In the formula, For the first Normalized fluctuation at each sampling point; The threshold for fluctuation; For the first Normalized fluctuation at each sampling point; For absolute value operations; The sampling period; The rate of change threshold; To maintain the duration; The sampling sequence number; S24. In the material system to be processed, the normalized fluctuation amount is recorded synchronously at different mixing times and samples are taken in advance. The mixing uniformity is calculated according to the coefficient of variation, where the coefficient of variation is the ratio of the sample standard deviation to the sample mean. The maximum normalized fluctuation amount that makes the coefficient of variation no greater than the target coefficient of variation is selected as the fluctuation amount threshold. S25. When the mixing time reaches the preset maximum mixing time and the judgment condition is still not met, an abnormal signal is output and the material discharge is suspended or the process is transferred to manual confirmation. This step ensures uniform mixing while avoiding under-mixing, over-mixing, and abnormal retention.

6. The method for controlling the weighing, mixing, and packaging of powders according to claim 1, characterized in that, The dynamic redistribution and quantitative filling in S3 include: S31. Determine the usable packaging weight for this batch. The usable packaging weight is obtained by weighing the material in this batch from the transition buffer silo, or by subtracting the calibrated residual conveying loss from the total mixed weight. Determine the total number of packages within a preset per-package allowable weight range, ensuring that the average target filling weight obtained by dividing the usable packaging weight by the total number of packages falls within the per-package allowable weight range. When multiple total packages meet the conditions, select the one that makes the average target filling weight closest to the nominal per-package weight. When no total packages meet the conditions, merge this batch of material into the next batch or recycle it and output an alarm. ； In the formula, This represents the lower limit of the permissible weight range for a single package. For packaging quality; This represents the total number of packages; This represents the upper limit of the permissible weight range for a single package; S32. Before filling each package, the current remaining usable package mass is obtained by subtracting the sum of the measured net filling masses of all previous packages from the usable package mass: ； In the formula, To fill the first The remaining packaging material is of usable quality before packaging; For packaging quality; For subscript From 1 to Summation; The current package number; The serial number of the already filled package; For the first The measured net fill mass of the package; S33. Divide the remaining usable mass of the packaging by the number of remaining packages to obtain the target filling weight of the current package: ； In the formula, For the first The target weight for filling the package; To fill the first The remaining packaging material is of usable quality before packaging; This represents the total number of packages; The current package number; This step distributes the filling error of each bag evenly across the entire batch, constrains the filling deviation of the last bag, and ensures that the weight of a single bag falls within the allowable range. S34. Using the target filling weight of the current package as the net target weight of the feed for this package, quantitatively fill the current package according to the dual-speed feeding and stop drop correction. After filling, seal, tie and stack in sequence.

7. The method for controlling the weighing, mixing, and packaging of powders according to claim 1, characterized in that, S4 includes: S41. After determining that the mixing is complete and the mixed material of this batch has been discharged and isolated from the ribbon mixer, start the weighing, batching and mixing of the next batch of material, and feed the discharged material to the packaging section for quantitative filling and packaging, so that the weighing, batching, mixing and quantitative filling and packaging of adjacent batches can be carried out in parallel flow.

8. A powder weighing, mixing, and packaging control device, used to implement the powder weighing, mixing, and packaging control method according to claim 1, characterized in that, It includes a weighing and batching unit, a mixing unit, a transition buffer silo, a packaging execution unit, a feeding control module, a mixing endpoint determination module, a collaborative control module, and a collaborative controller; The weighing and batching unit includes a weighing hopper, a weighing sensor mounted on the weighing hopper, and a feeding mechanism for feeding materials into the weighing hopper. The mixing unit includes a ribbon mixer, a drive motor for driving the ribbon mixer, a driver for supplying power to the drive motor, and a discharge valve disposed at the discharge port of the ribbon mixer. The transition buffer silo is connected to the ribbon mixer via the discharge valve. The transition buffer silo is equipped with a weighing device or a material level detection device to buffer and isolate the current batch of materials and to provide the packaging usable quality for the current batch. The packaging execution unit includes a packaging scale supplied by the transition buffer silo, a filling and feeding mechanism that supplies material to the packaging scale, and a sealing machine, a sealing machine, and a palletizing robot arranged sequentially downstream of the packaging scale along the material conveying direction. The input terminal of the feeding control module is connected to the weighing sensor and the packaging scale, and the control output terminal of the feeding control module is connected to the feeding mechanism and the filling feeding mechanism respectively, for executing the dual-speed feeding and stop drop correction; The input terminal of the mixing endpoint determination module is connected to the torque output terminal or current output terminal of the driver, and the mixing ready signal output terminal of the mixing endpoint determination module is connected to the cooperative control module, which is used to calculate the normalized fluctuation amount online and determine the mixing completion. The collaborative control module is connected to the transition buffer hopper, the mixing endpoint determination module and the feeding control module respectively, and is used to receive the packaging available quality and the mixing ready signal, and to perform dynamic reallocation of packaging available quality feedforward and single package filling target; The feeding control module, the mixing endpoint determination module, and the collaborative control module are integrated into the collaborative controller.

9. The powder weighing, mixing, and packaging control equipment according to claim 8, characterized in that, The feeding control module includes a weighing acquisition subunit, a flow identification subunit, a drop prediction subunit, and an inter-batch learning subunit connected in sequence. The control output terminal of the drop prediction subunit is connected to the feeding mechanism and the filling feeding mechanism for controlling the switching between coarse feeding and fine feeding, as well as the stopping of fine feeding. The mixing endpoint determination module includes a load acquisition subunit, a fluctuation calculation subunit, and an endpoint criterion subunit connected in sequence. The mixing ready signal output terminal of the endpoint criterion subunit is connected to the collaborative control module.

10. The powder weighing, mixing, and packaging control equipment according to claim 8, characterized in that, The collaborative control module includes a total mass feedforward subunit and a single-package target redistribution subunit. The total mass feedforward subunit is connected to the weighing device of the transition buffer silo to receive the available package mass. The output of the single-package target redistribution subunit is connected to the feeding control module to provide the single-package filling target. A batch isolation control valve is provided between the discharge port of the ribbon mixer and the transition buffer silo, and / or between the transition buffer silo and the packaging scale. The batch isolation control valve is connected to the collaborative controller. The collaborative controller includes a programmable logic controller, or includes a programmable logic controller and a host computer. The collaborative controller is equipped with a cycle scheduling unit. The trigger end of the cycle scheduling unit is connected to the mixing ready signal output end of the mixing endpoint determination module, and is used to trigger the weighing and batching of the next batch of materials after mixing is completed and the current batch of materials is isolated.

Images