Tea food medicine bag intelligent dispensing control system based on formula proportion

Abstract

The present application relates to the field of split charging control, and aims to solve the problems of poor control effect of traditional split charging control scheme on different forms of materials, lagging risk management and control of split charging action, and low formula switching efficiency, and specifically relates to a tea food medicine bag intelligent split charging control system based on formula proportion, which acquires material form categories in real time, automatically matches control parameters for materials with different physical characteristics, sets multiple parallel and linked quality detection and early warning points in the split charging process in a complex and changeable material environment, acutely warns potential problems caused by abnormal impact, material adhesion or sensor failure, discovers abnormalities earlier, responds through triggering early warning signals, and through the cooperative work of the modular formula switching mechanism and the linked monitoring and interception mechanism, quickly and automatically switches between different formulas, ensures no cross contamination, and meets the flexible production demand of small batch and multiple varieties.

Description

Technical Field

[0001] This invention relates to the field of dispensing control, specifically to an intelligent dispensing control system for tea, food, and medicine packets based on formula ratios. Background Technology

[0002] The production of tea, food, and medicine packets, especially those involving the mixing and packaging of multiple raw materials in specific proportions, requires extremely high precision in packaging, hygiene and safety, and batch consistency. Traditional packaging methods mainly rely on manual labor or semi-automatic weighing equipment, which suffer from problems such as low efficiency, high precision affected by human factors, and susceptibility to cross-contamination.

[0003] With the development of automation technology, quantitative dispensing machines based on weight feedback have emerged. These machines typically use single-point weighing sensors and PLC-controlled vibratory feeding or screw feeding to approximate the target weight. However, significant technical limitations still exist in practical applications. On the one hand, the system has poor adaptability to changes in the physical properties of materials. For example, the same control parameter has an unstable effect on the dispensing of materials in different forms, which can easily lead to deviations or low efficiency. On the other hand, the dispensing process lacks real-time intelligent judgment and early warning of the compliance of the materials themselves and the dispensing actions, resulting in lagging quality risk control. Furthermore, when different formulas need to be switched, it is often necessary to stop the machine and manually change materials and clean them, which cannot achieve flexible and continuous production.

[0004] Therefore, traditional tea, food and medicine packaging systems have shortcomings in terms of adaptive and precise control, real-time quality risk perception, and flexible formula switching, and urgently need a comprehensive solution that integrates visual recognition, intelligent weighing and model-based control. Summary of the Invention

[0005] This invention addresses the problems of poor control of materials in different forms, lagging risk management of dispensing operations, and low efficiency of formula switching in traditional dispensing control schemes. It proposes an intelligent dispensing control system for tea, food, and medicine bags based on formula ratios.

[0006] The objective of this invention can be achieved through the following technical solution: a smart packaging and control system for tea, food and medicine packets based on formula ratio, including a vision recognition module, a weight statistics module, a core control module and a multi-linkage execution module; The visual recognition module is used to collect image information of the packaged materials and to identify and classify the material shape and foreign objects in real time based on the image algorithm. The weight statistics module includes a high-precision array-type weighing unit, which is used to acquire dynamic weight data in real time during the packaging process, and at the same time perform error self-checking and early warning for the weight. The core control module is used to receive image information and dynamic weight data of the material. The core control module has a built-in fuzzy PID control algorithm, which is used to dynamically calculate and generate dispensing control parameters based on the material form, classification results and preset target weight value. The multi-linkage execution module is used to execute the dispensing control parameters. The multi-linkage execution module controls the modular formula switching mechanism and the linkage monitoring and interception mechanism to complete the dispensing of materials and formula switching operations. The full-process review and verification module is used to retrieve and review the timing data of the entire packaging process after a single packaging operation is completed, compare the timing data with the preset process standards, verify the correctness of the type, weight and packaging status of the packaged materials, and generate verification results. The core control module receives the verification results and decides whether the current package should flow into the next process or be rejected.

[0007] In a preferred embodiment of the present invention, the visual recognition module acquires images through an industrial camera and identifies the flake, granular, and powder forms of tea leaves and medicinal materials through a lightweight neural network model deployed in the edge computing unit, and classifies and detects foreign objects not included in the formula. When the visual recognition module detects a foreign object, it generates an early warning and sends a stop command to the core control module.

[0008] In a preferred embodiment of the present invention, the high-precision array weighing unit in the weight statistics module is composed of multiple miniature weighing sensors. After acquiring weight data each time, the weight statistics module temporarily stores the weight data and records it as a temporary weight. The weight statistics module compares the continuously acquired temporary weights to obtain the difference between adjacent weights. When the difference between adjacent weights is 0, the two adjacent temporary weights are recorded as stable weights. When the number of consecutive occurrences of stable weights exceeds a set threshold, the stable weight is recorded as the current material weight.

[0009] In a preferred embodiment of the present invention, the weight statistics module constructs a measured data sequence based on the acquisition time of the temporary weight and the stable weight, and performs error self-checking on the measured data sequence. The specific method is as follows: S1: Based on the sub-packaging process parameters, establish a theoretical model of the impact of material falling impact on dynamic weighing results; and obtain the expected curve of weight change over time through the theoretical model, which includes the instantaneous positive peak value caused by material impact; S2: During a single packaging process, the measured data sequence is time-series aligned and compared with the theoretical weight evolution model; S3: Extract key features from the measured data sequence, including the magnitude of the first weight jump, the time to reach the peak, and the rate of decay from the peak to the steady value; S4: If the key features of the measured data sequence deviate from the preset reasonable feature range, it is determined that there is unreasonable data in this weighing process caused by abnormal impact, material adhesion or sensor failure; the weight statistics module immediately generates a graded warning signal containing the abnormal feature description and the degree of deviation, and sends the warning signal to the core control module.

[0010] In a preferred embodiment of the present invention, the fuzzy PID control algorithm in the core control module has input variables including the deviation between the current material weight and the target weight, the rate of change of the deviation, and the material morphology category provided by the visual recognition module, and outputs the valve opening of the feeding mechanism in the multi-linkage execution module, so as to achieve adaptive and precise control of the dispensing speed of materials of different morphologies.

[0011] In a preferred embodiment of the present invention, when the core control module controls the valve opening of the feeding mechanism, it uses the deviation between the current material weight and the target weight and the rate of change of the deviation as parameters, obtains the feeding speed reference value through weight allocation, and then combines the material form category for correction and limitation, and converts the feeding speed reference value into the valve opening.

[0012] In a preferred embodiment of the present invention, the modular formula switching mechanism includes a rotating hopper group, with each hopper corresponding to a formula material; The linkage monitoring and interception mechanism consists of an electromagnetic valve and a high-speed vision sensor. When the core control module determines that the formula needs to be switched, it drives the rotating silo group to switch workstations and ensures that there is no cross-contamination or material leakage during the switching process through the linkage monitoring and interception mechanism.

[0013] As a preferred embodiment of the present invention, the time-series data reviewed by the full-process review and verification module includes: image snapshot sequence of each material in the formula during dispensing, dynamic weight accumulation curve corresponding to each stage, and action timestamp log of the modular formula switching mechanism. The comparison process employs data alignment and threshold judgment rules based on time windows.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. In this invention, by acquiring the material form category in real time and using it as one of the key input parameters of the fuzzy PID control algorithm, the core control module can dynamically adjust the packaging strategy. For materials with different physical properties, the system can automatically match the optimal control parameters, effectively overcoming the control instability caused by differences in material flowability and adhesion. Thus, even in complex and ever-changing material environments, the packaging weight can still meet high precision requirements, significantly improving product consistency and pass rate.

[0015] 2. This invention sets up multiple parallel and linked quality detection and early warning points during the packaging process, which can not only identify and intercept foreign objects in real time at the source, but also conduct a rational analysis of the impact law of falling material by constructing a measured data sequence and comparing it with the theoretical model. It can keenly warn of potential problems caused by abnormal impact, material adhesion or sensor failure. Compared with simple result threshold judgment, it can detect abnormalities earlier and more fundamentally, and feed back the early warning signal to the core control module in real time, which can immediately trigger responses such as deceleration, verification or shutdown.

[0016] 3. This invention also enables the system to quickly and automatically switch between different formulas through the coordinated operation of a modular formula switching mechanism and a linkage monitoring and interception mechanism, ensuring no cross-contamination and meeting the flexible production needs of small batches and multiple varieties. After the packaging is completed, the full-process review and verification module retrieves and reviews the time-series data of the entire chain and performs intelligent comparison with the process standards based on a time window. This not only accurately verifies the final result but also locates any minor deviations in the packaging process, providing a complete and reliable data chain for production management, process optimization, and quality problem traceability. Attached Figure Description

[0017] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0018] Figure 1 This is a system block diagram of the present invention; Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1: Please refer to Figure 1 - Figure 2As shown, an intelligent packaging and control system for tea, food, and medicine packets based on formula ratios is presented. The system uses a core control module as the central decision-making unit, coordinating the visual recognition module, weight statistics module, multi-linkage execution module, and verification module to work together to form a complete closed loop of perception, decision-making, execution, and verification. It aims to achieve fully automated, high-precision, and traceable packaging operations for multiple materials and formulas. The core control module communicates with each module through an industrial fieldbus and is equipped with an IoT interface, which can interact with the upper-level production management system for data exchange.

[0021] The visual recognition module is responsible for the initial quality inspection and classification of materials before packaging. Its hardware includes at least one or more sets of high-resolution industrial cameras and edge computing units connected to them, specifically including image acquisition, intelligent recognition and classification, and early warning and linkage functions. The image acquisition part of the vision recognition module uses industrial cameras set at key points in the material conveying or feeding path to capture surface and overall images of the material to be packaged. Intelligent recognition and classification involve deploying lightweight convolutional neural network models trained on specific datasets within edge computing units. These models can perform two tasks simultaneously: Material form classification: Automatically identifies whether the material in the image belongs to a preset category such as flake, granule or powder, and sends the classification result to the core control module in real time.

[0022] Foreign object detection: Identifies and selects foreign objects that are not within the scope of the current formulation, such as metal shavings, plastic pieces, hair, or other types of materials.

[0023] Early warning and linkage: When the neural network model detects a foreign object, the visual recognition module will immediately generate an early warning signal containing the image and location information of the foreign object, and send an emergency stop command to the core control module through the high-speed communication interface to prevent the problematic batch of materials from entering the packaging process and eliminate contamination at the source.

[0024] The weight statistics module is responsible for accurately measuring the weight of materials during the packaging process and has the ability to intelligently analyze the weighing process. Its core is a high-precision array-type weighing unit.

[0025] Weighing unit composition: This unit consists of multiple miniature high-precision weighing sensors arranged in a specific array to support the weighing platform. This design can evenly distribute the load and improve measurement stability and resistance to off-center loads.

[0026] Dynamic weight data acquisition and processing: The module acquires sensor data at high frequency and executes the following processing logic to obtain stable and reliable weight values: Data storage: The original weight value collected each time is temporarily stored as a temporary weight; Stability determination: The system compares the continuously acquired temporary weights in real time and calculates the difference between adjacent weights. When multiple consecutive adjacent weight differences are detected to be zero or below the minimum noise threshold, the system considers the weight readings to have entered a stable state and records these consistent weight values ​​as stable weights. The final result confirms that the system only recognizes a stable weight as a valid current material weight and outputs it to the core control module for control decision-making when the number of consecutive occurrences of the stable weight exceeds the preset reliability threshold.

[0027] In addition, the weight statistics module also performs error self-checking and early warning based on the weighing process. The specific error self-checking and early warning process is as follows: Establish a theoretical model: Based on material characteristics (density, particle size) and process parameters such as drop height, the system pre-establishes or learns a material drop impact dynamics model. This model can simulate and generate a theoretical weight-time change curve under ideal conditions, from the start of drop, through the impact peak, to the oscillation decay and then to stability. Actual sequence alignment: In each actual packaging, the weight statistics module uses the acquired temporary weight and stable weight to construct an actual weight data sequence in chronological order.

[0028] Feature extraction and rationality analysis: The system aligns the measured sequence with the theoretical model in time. Subsequently, it extracts key dynamic features from the measured sequence, including: the slope of the initial weight jump (reflecting the magnitude of the impact force), the time point at which the weight peak is reached, and the decay rate from the peak to the stable value. These features collectively describe the dynamic behavior of this material drop.

[0029] Intelligent early warning: The system presets a reasonable feature range based on theoretical models and a large amount of normal data statistics. If any one or more key features extracted deviate significantly from the reasonable range, it is determined that there are potential risks such as abnormal impact, material adhesion or sensor transient failure in the weighing process. At this time, the module will generate a graded early warning signal with a specific description of the deviation features and send it to the core control module.

[0030] The core control module is implemented using a high-performance industrial controller. The main control processes include information fusion, adaptive fuzzy PID control flow, and control output generation. Information fusion: The module receives material form category from the visual recognition module in real time, as well as dynamic weight data and its early warning signals from the weight statistics module; Adaptive fuzzy PID control: The module runs the core fuzzy proportional-integral-derivative control algorithm, whose input variables are: The deviation between the current material weight and the target weight.

[0031] The rate of change of this deviation (representing the rate at which weight increases).

[0032] Material form categories provided by the visual recognition module.

[0033] Control output generation: The algorithm first uses a set of preset fuzzy rule bases (e.g., "if the deviation is large and the rate of change of deviation is large, then the output opening is greatly reduced") to reason and obtain a preliminary reference value for the feeding speed.

[0034] Next, the system corrects and limits this reference value according to the material form category. For example, for powdery materials that are prone to dust, the algorithm will limit the maximum opening to prevent splashing, while for sheet-like materials with poor flowability, a more refined pulse feeding strategy may be adopted when the value is close to the target value.

[0035] Ultimately, the corrected feed rate reference value is converted into precise control commands for the valve opening or vibration frequency of the precision feeding mechanism in the multi-linkage execution module. This control strategy, which integrates material shape feedforward and weight deviation feedback, enables rapid, accurate, and smooth dispensing of different materials.

[0036] The multi-linkage execution module is responsible for accurately executing the action commands issued by the core control module and realizing automatic formula switching, including a modular formula switching mechanism and a linkage monitoring and interception mechanism.

[0037] The modular formula switching mechanism consists of a multi-station rotating hopper group. Each independent hopper is pre-filled with the materials required for a formula. The entire hopper group can be precisely indexed and rotated under the drive of a servo motor.

[0038] The linkage monitoring and interception mechanism is designed to ensure pollution-free switching processes. It consists of a linkage mechanism composed of high-speed electromagnetic valves and miniature high-speed vision sensors installed at the discharge port of each hopper and the common material drop channel.

[0039] Formula switching process: When the core control module needs to change materials according to the production formula: Close the solenoid valve of the current workstation hopper to ensure no residual material leaks.

[0040] The rotating hopper assembly is driven to rotate, accurately aligning the hoppers containing the materials required for the next formula with the feeding station.

[0041] During this process, high-speed vision sensors capture images of the common material feeding channel in real time. The core control module analyzes the images to confirm that there is no cross-contamination or material leakage. Only after confirmation is that there are no problems can the solenoid valve of the new material hopper be opened to begin the dispensing of the next material.

[0042] Example 2: Please refer to Figure 1 - Figure 2As shown, the full-process review and verification module is activated after a single packaging operation is completed. It performs the final full-link data review, retrieving multi-dimensional time-series data synchronized by timestamps from the system cache to form a review evidence chain, which mainly includes: Image snapshot sequence: Keyframe images captured by the visual recognition module before and after each type of material is packaged.

[0043] Dynamic weight cumulative curve: From start to finish, the weight statistics module records all weight change curves, which can clearly show the weight steps after each material is added.

[0044] Action timestamp log: Records the precise time of each solenoid valve opening and closing and hopper rotation in the multi-linkage execution module.

[0045] The end-to-end review and verification module employs time-window-based data alignment technology to precisely correlate the aforementioned multi-source data along the timeline. Then, it automatically compares the data using preset threshold judgment rules. For example, the module checks whether the weight step after each material is added meets the formulation requirements; it checks whether the appearance of the materials in the image snapshot is consistent with expectations; it checks whether the action log conforms to the established process sequence; and finally, the module generates a comprehensive verification result.

[0046] The core control module receives the verification results and makes a final decision based on them: if the result is qualified, it instructs the conveyor belt to flow the current tea, food and medicine package into the packaging and next process; if it is unqualified, it instructs the sorting mechanism to automatically reject it into the waste bin. All raw data and verification results are uploaded to the cloud database through the Internet of Things interface for quality traceability and production analysis.

[0047] In summary, after the system of this invention starts, the core control module, following the formula sequence, first instructs the visual recognition module to inspect the materials to be packaged. After the inspection is passed, combining the target weight and material form, the feeding mechanism in the multi-linkage execution module is controlled by a fuzzy PID algorithm to accurately add materials. Simultaneously, the weight statistics module monitors the weighing process in real time and performs health checks. Once one material is packaged, the system automatically switches to the next material hopper, repeating the above process until all materials in the formula have been added. Finally, the verification module performs a full-process data review and adjudication to determine the product's destination, thus completing a highly reliable and traceable intelligent packaging cycle.

[0048] Thresholds, preset values, preset ranges, etc. are set for result comparison and analysis to determine whether they are good or bad. The value of these thresholds is determined by a combination of large-scale model analysis of sample data and human experience. They can also be adjusted appropriately based on seasonal or common-sense influences. Furthermore, the settings for weighting ratios, influence factors, etc., are based on the magnitude of each parameter's influence on the results. The specific values ​​are allocated to ultimately reflect the impact on the results. The settings for input and storage are also determined by a combination of large-scale model analysis of sample data and human experience. Appropriate adjustments can also be made based on seasonal or rational influence conditions.

[0049] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A smart packaging and control system for tea, food, and medicine packets based on formula ratios, characterized in that, It includes a vision recognition module, a weight statistics module, a core control module, and a multi-linkage execution module; The visual recognition module is used to collect image information of the packaged materials and to identify and classify the material shape and foreign objects in real time based on the image algorithm. The weight statistics module includes a high-precision array-type weighing unit, which is used to acquire dynamic weight data in real time during the packaging process, and at the same time perform error self-checking and early warning for the weight. The core control module is used to receive image information and dynamic weight data of the material. The core control module has a built-in fuzzy PID control algorithm, which is used to dynamically calculate and generate dispensing control parameters based on the material form, classification results and preset target weight value. The multi-linkage execution module is used to execute the dispensing control parameters. The multi-linkage execution module controls the modular formula switching mechanism and the linkage monitoring and interception mechanism to complete the dispensing of materials and formula switching operations. The full-process review and verification module is used to retrieve and review the timing data of the entire packaging process after a single packaging operation is completed, compare the timing data with the preset process standards, verify the correctness of the type, weight and packaging status of the packaged materials, and generate verification results. The core control module receives the verification results and decides whether the current package should flow into the next process or be rejected.

2. The intelligent packaging control system for tea, food, and medicine packets based on formula ratios according to claim 1, characterized in that, The visual recognition module acquires images through an industrial camera and identifies the flake, granular, and powder forms of tea leaves and medicinal herbs using a lightweight neural network model deployed in the edge computing unit. It also classifies and detects foreign objects not included in the formula. When a foreign object is detected, the visual recognition module generates an early warning and sends a stop command to the core control module.

3. The intelligent packaging control system for tea, food, and medicine packets based on formula ratios according to claim 1, characterized in that, The high-precision array weighing unit in the weight statistics module is composed of multiple miniature weighing sensors. After acquiring weight data each time, the weight statistics module temporarily stores the weight data and records it as a temporary weight. The weight statistics module compares the continuously acquired temporary weights to obtain the difference between adjacent weights. When the difference between adjacent weights is 0, the two adjacent temporary weights are recorded as stable weights. When the number of consecutive occurrences of stable weights exceeds a set threshold, the stable weight is recorded as the current material weight.

4. The intelligent packaging control system for tea, food, and medicine packets based on formula ratios according to claim 3, characterized in that, The weight statistics module constructs a measured data sequence based on the acquisition time of the temporary weight and the stable weight, and performs error self-checking on the measured data sequence. The specific method is as follows: S1: Based on the sub-packaging process parameters, establish a theoretical model of the impact of material falling impact on dynamic weighing results; and obtain the expected curve of weight change over time through the theoretical model, which includes the instantaneous positive peak value caused by material impact; S2: During a single packaging process, the measured data sequence is time-series aligned and compared with the theoretical weight evolution model; S3: Extract key features from the measured data sequence, including the magnitude of the first weight jump, the time to reach the peak, and the rate of decay from the peak to the steady value; S4: If the key features of the measured data sequence deviate from the preset reasonable feature range, it is determined that there is unreasonable data in this weighing process caused by abnormal impact, material adhesion or sensor failure; the weight statistics module immediately generates a graded warning signal containing the abnormal feature description and the degree of deviation, and sends the warning signal to the core control module.

5. The intelligent packaging control system for tea, food, and medicine packets based on formula ratios according to claim 1, characterized in that, The fuzzy PID control algorithm in the core control module has input variables including the deviation between the current material weight and the target weight, the rate of change of the deviation, and the material shape category provided by the visual recognition module. The output is the valve opening of the feeding mechanism in the multi-linkage execution module, so as to achieve adaptive and precise control of the dispensing speed of materials of different shapes.

6. The intelligent packaging control system for tea, food, and medicine packets based on formula ratios according to claim 1, characterized in that, When the core control module controls the valve opening of the feeding mechanism, it uses the deviation between the current material weight and the target weight, and the rate of change of the deviation as parameters. Through weight allocation, it obtains a reference value for the feeding speed, and then combines it with the material form category for correction and limitation, converting the reference value for the feeding speed into the valve opening.

7. The intelligent packaging control system for tea, food, and medicine packets based on formula ratios according to claim 1, characterized in that, The modular formula switching mechanism includes a rotating hopper group, with each hopper corresponding to a formula material; The linkage monitoring and interception mechanism consists of an electromagnetic valve and a high-speed vision sensor. When the core control module determines that the formula needs to be switched, it drives the rotating silo group to switch workstations and ensures that there is no cross-contamination or material leakage during the switching process through the linkage monitoring and interception mechanism.

8. The intelligent packaging control system for tea, food, and medicine packets based on formula ratios according to claim 1, characterized in that, The time-series data reviewed by the full-process review and verification module includes: image snapshot sequences of each material in the formula during dispensing, dynamic weight accumulation curves corresponding to each stage, and action timestamp logs of the modular formula switching mechanism. The comparison process employs data alignment and threshold judgment rules based on time windows.

Images