Tying machine intelligent control method and tying machine thereof

Abstract

The invention relates to the technical field of tying devices, and particularly discloses a tying machine intelligent control method and a tying machine thereof, and the method comprises the steps: collecting the state data of a tying operation environment through a multi-mode sensing module; based on the state data, generating a target torque Ttarget and an optimal motion stroke Sooptimal of the motor through a tying decision model; based on the generated parameters, the motor is driven to operate, and motor output is dynamically adjusted through a feedback control algorithm during operation; after the tying action is completed, performing image analysis on the tying effect, and determining subsequent operation according to an analysis result; according to the invention, by introducing the multi-modal perception and intelligent decision algorithm, the self-adaptive precise regulation and control of the core parameters in the tying process are realized, the optimal tying strength and stroke are dynamically matched according to the physical characteristics of the object to be tied, and the problems of poor adaptability, easy puncturing or loose tying and the like caused by fixed parameters of the traditional equipment are solved.

Description

Technical Field

[0001] This invention belongs to the field of sealing device technology, specifically relating to an intelligent control method for sealing machines and the sealing machine itself. Background Technology

[0002] In existing technologies, sealing machines (or sealing devices) typically use mechanical transmission or simple motor drive to complete the sealing operation. The basic working principle of these devices is that, through manual operation by the user or motor-driven compression end, a U-shaped staple is pushed and squeezed through the sealing component, causing the U-shaped staple to bend and lock the bag opening. The parameters of the drive mechanism (such as the motor's output torque and stroke) are often preset fixed values, and its operation is an open-loop control, lacking the ability to sense and respond to external changes.

[0003] However, this open-loop control method based on fixed parameters has significant limitations. First, because the objects to be sealed vary greatly in terms of material, thickness, and filler fullness, fixed sealing force and stroke are difficult to adapt to all situations, easily leading to sealing failure: too little force may result in insufficient U-shaped nail insertion or loose sealing; too much force may cause the U-shaped nail to bend and deform or even puncture the packaging bag. Second, existing equipment relies entirely on the operator's experience to judge whether it is successful, and cannot automatically detect and ensure the sealing quality. In addition, mechanical parts inevitably wear down after long-term use, and the control system based on fixed parameters cannot compensate for this, leading to a gradual deterioration of sealing performance and insufficient reliability. These shortcomings seriously restrict the widespread application of sealing machines in modern working environments with increasingly higher requirements for automation and intelligence. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent control method for a sealing machine and a sealing machine thereof, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for intelligent control of a sealing machine includes the following steps:

[0007] S1. Collect status data of the tying operation environment through a multimodal sensing module;

[0008] S2. Based on the aforementioned state data, generate the target torque T of the motor using the tie-down decision model. target With the optimal motion distance S optimal ;

[0009] S3. Based on the generated parameters, drive the motor to run, and dynamically adjust the motor output during operation through a feedback control algorithm;

[0010] S4. After the tying action is completed, perform image analysis on the tying effect and determine the subsequent operation based on the analysis results.

[0011] Preferably, in S1, the target torque T target The specific algorithm is as follows:

[0012]

[0013] Among them, S texture It is based on the complexity of the surface texture of the object to be tethered, extracted from visual images;

[0014] S density It is the material density coefficient of the object to be ligated, estimated based on visual images; D prox It is the diameter or thickness of the object to be ligated detected by the proximity sensor; w1, w2, and K are the weights and coefficients obtained through machine learning training; T base This is the base torque value calibrated for this model of sealing machine.

[0015] Preferably, the material density coefficient S density Obtained through the following methods:

[0016] Control the clamping end to contact the surface of the object to be clamped with a preset detection force;

[0017] Record the instantaneous micro-displacement Δd generated at the clamping end under the aforementioned probing force;

[0018] calculate Where F probe The detection force is given by k, which is a conversion factor determined based on the structural stiffness of the closure machine.

[0019] Preferably, in step S3, the "dynamic adjustment of motor output through feedback control algorithm" adopts fuzzy PID control, with a proportional coefficient K. p The online adjustment rules are determined based on the sum and ratio of the absolute values ​​of the error e and the error change rate ec:

[0020] Calculate the intermediate decision variables α and β:

[0021] α = |e| + |ec|,

[0022] Here, δ is a minimal constant set to prevent division by zero; K is dynamically output by querying a predefined fuzzy rule table based on the values ​​of α and β. p Adjustment amount ΔK p .

[0023] Preferably, in the "image analysis of the sealing effect", the quantitative standard for determining whether the sealing is qualified is:

[0024] Image recognition technology is used to locate the two prongs of the U-shaped nail;

[0025] Calculate the pin symmetry (Sym) and embedding degree (Emb);

[0026]

[0027] Where L1 and L2 are the lengths of the two pins, D1 and D2 are the embedment depths of the two pins, and T is the length of the pin. object Let be the average thickness of the object to be ligated; if Sym > θ simultaneously sym And θ emb_low <Emb<θ emb_high If the knot is sealed properly, then the knot is deemed acceptable.

[0028] Preferably, after the initial ligation attempt is deemed a failure, an adaptive religation strategy is implemented, including:

[0029] Based on the failure mode of the first closure, the target torque T is adjusted according to preset rules. target If the nail feet are bent, reduce the torque; if the nail is not properly embedded, increase the torque.

[0030] The re-tying operation is performed using the adjusted torque, and the success or failure of this operation is recorded in the training dataset of the tying decision model for online incremental learning of the model.

[0031] Preferably, it also includes predictive maintenance steps based on historical data:

[0032] Continuously record the average operating current I of the motor during each tying operation. avg and no-load starting current I start ;

[0033] When I avg Or I start The cumulative drift ΔI relative to its initial baseline value exceeds the maintenance threshold I. th At that time, a maintenance reminder signal is generated;

[0034]

[0035] A sealing machine, used to implement the intelligent control method for sealing machines described in any one of the above claims, comprising:

[0036] One-piece alloy die-cast housing;

[0037] A closure component is provided at the front end of the housing;

[0038] The sealing and extrusion end is slidably installed inside the housing and can move toward the sealing component;

[0039] The drive mechanism includes a motor and a multimodal sensing module;

[0040] A control unit, located inside the housing, is used to execute the intelligent control method.

[0041] Preferably, the multimodal sensing module integrates:

[0042] A CMOS vision sensor facing the closure position is used to acquire high-resolution images;

[0043] A laser rangefinder sensor, as the proximity sensor, is used to accurately detect the thickness of the object to be ligated;

[0044] A high-precision Hall current sensor is used to monitor the operating current of the motor in real time.

[0045] Compared with the prior art, the beneficial effects of the present invention are:

[0046] (1) By introducing multimodal perception and intelligent decision-making algorithms, the core parameters of the tying process are adaptively and precisely controlled. The optimal tying force and stroke are dynamically matched according to the physical characteristics of the object to be tying, overcoming the problems of poor adaptability, easy puncture or loose tying caused by fixed parameters in traditional equipment. At the same time, the closed-loop control strategy based on real-time feedback ensures the stability and consistency of the execution process. Combined with the visual quality detection after tying, a complete intelligent closed loop from perception, decision-making, execution to verification is formed.

[0047] (2) This invention completely liberates operators from relying on experience-based judgment, greatly reducing the operational threshold and error rate. Its built-in self-learning and predictive maintenance functions not only enable the equipment to become more intelligent with long-term use, but also provide early warning of potential faults, thereby improving the reliability and service life of the equipment. Attached Figure Description

[0048] Figure 1 This is a perspective view of the present invention;

[0049] Figure 2 This is a flowchart of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Example 1:

[0052] Please see Figure 1-2As shown, an intelligent control method for a sealing machine includes the following steps:

[0053] S1. Collect status data of the tying operation environment through a multimodal sensing module;

[0054] S2. Based on the aforementioned state data, generate the target torque T of the motor using the tie-down decision model. target With the optimal motion distance S optimal ;

[0055] S3. Based on the generated parameters, drive the motor to run, and dynamically adjust the motor output during operation through a feedback control algorithm;

[0056] S4. After the tying action is completed, perform image analysis on the tying effect and determine the subsequent operation based on the analysis results.

[0057] In one embodiment of the present invention, in S1, the target torque T target The specific algorithm is as follows:

[0058]

[0059] Among them, S texture It is based on the complexity of the surface texture of the object to be tethered, extracted from visual images; S density It is the material density coefficient of the object to be ligated, estimated based on visual images; D prox It is the diameter or thickness of the object to be ligated detected by the proximity sensor; w1, w2, and K are the weights and coefficients obtained through machine learning training; T base This is the base torque value calibrated for this model of sealing machine.

[0060] In one embodiment of the present invention, the material density coefficient S density Obtained through the following methods:

[0061] Control the clamping end to contact the surface of the object to be clamped with a preset detection force;

[0062] Record the instantaneous micro-displacement Δd generated at the clamping end under the aforementioned probing force;

[0063] calculate Where F probe Let k be the detection force, and k be the conversion coefficient determined based on the structural stiffness of the closure machine.

[0064] In one embodiment of the present invention, in step S3, the "dynamic adjustment of motor output through feedback control algorithm" adopts fuzzy PID control, with a proportional coefficient K. p The online adjustment rules are determined based on the sum and ratio of the absolute values ​​of the error e and the error change rate ec:

[0065] Calculate the intermediate decision variables α and β:

[0066] α = |e| + |ec|,

[0067] Here, δ is a minimal constant set to prevent division by zero; K is dynamically output by querying a predefined fuzzy rule table based on the values ​​of α and β. p Adjustment amount ΔK p .

[0068] In one embodiment of the present invention, the quantitative standard for determining whether the closure is qualified in the "image analysis of the closure effect" is:

[0069] Image recognition technology is used to locate the two prongs of the U-shaped nail;

[0070] Calculate the pin symmetry (Sym) and embedding degree (Emb);

[0071]

[0072] Where L1 and L2 are the lengths of the two pins, D1 and D2 are the embedment depths of the two pins, and T is the length of the pin. object Let be the average thickness of the object to be ligated; if Sym > θ simultaneously sym And θ emb_low <Emb<θ emb_high If the knot is sealed properly, then the knot is deemed acceptable.

[0073] In one embodiment of the present invention, after the initial ligation is determined to be a failure, an adaptive re-ligation strategy is executed, including:

[0074] Based on the failure mode of the first closure, the target torque T is adjusted according to preset rules. target If the nail feet are bent, reduce the torque; if the nail is not properly embedded, increase the torque.

[0075] The re-tying operation is performed using the adjusted torque, and the success or failure of this operation is recorded in the training dataset of the tying decision model for online incremental learning of the model.

[0076] In one embodiment of the invention, a predictive maintenance step based on historical data is also included:

[0077] Continuously record the average operating current I of the motor during each tying operation. avg and no-load starting current I start ;

[0078] When I avg or (I) start The cumulative drift ΔI relative to its initial baseline value exceeds the maintenance threshold I. thAt that time, a maintenance reminder signal is generated;

[0079]

[0080] A sealing machine, used to implement the intelligent control method for sealing machines described in any one of the above claims, comprising:

[0081] One-piece alloy die-cast housing;

[0082] A closure component is provided at the front end of the housing;

[0083] The sealing and extrusion end is slidably installed inside the housing and can move toward the sealing component;

[0084] The drive mechanism includes a motor and a multimodal sensing module;

[0085] A control unit, located inside the housing, is used to execute the intelligent control method.

[0086] In one embodiment of the present invention, the multimodal sensing module integrates:

[0087] A CMOS vision sensor facing the closure position is used to acquire high-resolution images;

[0088] A laser rangefinder sensor, as the proximity sensor, is used to accurately detect the thickness of the object to be ligated;

[0089] A high-precision Hall current sensor is used to monitor the operating current of the motor in real time.

[0090] Example 2:

[0091] This embodiment provides a sealing machine that integrates an intelligent adaptive control system. Its mechanical structure and intelligent control system work together to achieve efficient and high-success-rate automated sealing operations.

[0092] Reference Figure 1 The sealing machine includes a one-piece die-cast alloy housing 1, which forms a robust and lightweight mounting base. A sealing component 2 is located at the front end of the housing 1. The sealing and pressing end 3 is slidably mounted inside the housing 1 via a precision guide rail, and can move precisely toward the sealing component 2 under the action of the drive mechanism to complete the sealing action.

[0093] The multimodal sensing module is integrated into key locations inside the housing, including:

[0094] The vision sensor uses a miniaturized CMOS image sensor, facing the tying station, to capture the surface texture, preliminary shape, and image of the U-shaped nail after tying of the object to be tied (such as a garbage bag);

[0095] The proximity / range sensor uses a laser rangefinder to accurately measure the thickness or diameter D of the object to be tethered. prox ;

[0096] The current sensor uses a high-precision Hall current sensor, which is connected in series in the motor drive circuit to monitor the motor's operating current in real time. This current value is proportional to the motor's output torque.

[0097] The core of the control unit is a system-on-a-chip (SoC) that integrates an ARM Cortex-M7 core and a CNN accelerator. When the program instructions stored inside are executed, the following intelligent control process is implemented:

[0098] Step S101: Environmental perception and parameter calculation

[0099] When the user places the sealing machine over the bag opening, the multimodal sensing module is activated, the vision sensor captures an image of the bag opening, and the control unit runs an image processing algorithm to extract the surface texture complexity S. texture (For example, a smooth plastic bag has a thickness of about 0.1, while a rough woven bag has a thickness of about 0.8).

[0100] Simultaneously, the clamping end 3 performs a gentle "pre-compression" action with a constant probing force F. probe Contact the bag opening and measure the resulting micro-displacement Δd. Then, according to the formula... Calculate the material density coefficient S density (The small displacement of the fluffy plastic bag is large, S) density Small value; small micro-displacement of compacted garbage bags, S density (High value);

[0101] The laser rangefinder simultaneously provides the thickness D of the bag opening. prox ;

[0102] All this data is fed into a choke point decision model pre-installed in a CNN accelerator. The model performs calculations rapidly:

[0103]

[0104] For a half-full woven bag, the calculated torque ensures that the U-shaped nails can be fully pierced and fastened, while avoiding the nail feet bending or damaging the bag due to excessive torque.

[0105] Step S102: Adaptive Motion Control

[0106] Upon receiving the target torque command, the motor begins to drive the extrusion end 3 of the seam to move. Throughout the extrusion process, the current sensor continuously feeds back the actual torque T. current ;

[0107] The control unit calculates the error e = 5.2 - T.current and its change rate ec;

[0108] Subsequently, the fuzzy PID control algorithm is called:

[0109] Calculate α = |e| + |ec|,

[0110] Query the fuzzy rule table: When the system rapidly approaches the target value in the initial stage, α is larger and β is smaller, and the algorithm will automatically increase K p to make the motor respond quickly; when approaching the target value, both α and β become smaller, and the algorithm keeps K p stable, perfectly suppressing the overshoot phenomenon and making the bag-tying process smooth and precise;

[0111] Step S103: Quality verification and self-learning

[0112] After the bag-tying action is completed, the vision sensor takes a close-up of the bag-tying effect again;

[0113] The image recognition algorithm locates the two legs of the U-shaped staple, measures their lengths L1, L2 and the embedding depths D1, D2, and calculates:

[0114]

[0115] Compare the results with the preset thresholds (θ sym = 0.9, θ emb_low = 0.6, θ emb_high = 0.9). In this example, Sym > 0.9 and 0.6 < Emb < 0.9, and the system determines that the bag-tying is successful.

[0116] The operation data of this successful operation (environmental state and the torque used) will be used as a new positive sample and stored in the training data set. When a certain amount of such data accumulates, the system can start online incremental learning during idle time to fine-tune the parameters (such as w1, w2, K) of the bag-tying decision model, making its decision more accurate when facing similar scenarios in the future.

[0117] The system background continuously runs a maintenance monitoring program. After each bag-tying, it records the average working current and starting current of the motor. The controller regularly calculates its cumulative drift ΔI relative to the new machine reference value. When ΔI exceeds the threshold, it indicates that there may be wear or insufficient lubrication in the transmission mechanism, and the system will prompt the user to perform maintenance through the flashing of the indicator light, effectively avoiding sudden failures.

[0118] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent control of a seaming machine, characterized in that, Includes the following steps: S1. Collect status data of the tying operation environment through a multimodal sensing module; S2. Based on the aforementioned state data, generate the target torque T of the motor using the tie-down decision model. target With the optimal motion distance S optimal ; S3. Based on the generated parameters, drive the motor to run, and dynamically adjust the motor output during operation through a feedback control algorithm; S4. After the tying action is completed, perform image analysis on the tying effect and determine the subsequent operation based on the analysis results.

2. The intelligent control method for a sealing machine according to claim 1, characterized in that: In S1, the target torque T target The specific algorithm is as follows: Among them, S texture It is based on the complexity of the surface texture of the object to be tethered, extracted from visual images; S density It is the material density coefficient of the object to be ligated, estimated based on visual images; D prox It is the diameter or thickness of the object to be ligated detected by the proximity sensor; w1, w2, and K are the weights and coefficients obtained through machine learning training; T base This is the base torque value calibrated for this model of sealing machine.

3. The intelligent control method for a sealing machine according to claim 1, characterized in that: The material density coefficient S density Obtained through the following methods: Control the clamping end to contact the surface of the object to be clamped with a preset detection force; Record the instantaneous micro-displacement Δd generated at the clamping end under the aforementioned probing force; calculate Where F probe The detection force is given by k, which is a conversion factor determined based on the structural stiffness of the closure machine.

4. The intelligent control method for a sealing machine according to claim 1, characterized in that: In step S3, the "dynamic adjustment of motor output through feedback control algorithm" adopts fuzzy PID control, with a proportional coefficient K. p The online adjustment rules are determined based on the sum and ratio of the absolute values ​​of the error e and the error change rate ec: Calculate the intermediate decision variables α and β: α=|e|+|ec|, Here, δ is a minimal constant set to prevent division by zero; K is dynamically output by querying a predefined fuzzy rule table based on the values ​​of α and β. p Adjustment amount ΔK p .

5. The intelligent control method for a sealing machine according to claim 1, characterized in that: In the "image analysis of the sealing effect", the quantitative standard for determining whether the sealing is qualified is as follows: Image recognition technology is used to locate the two prongs of the U-shaped nail; Calculate the pin symmetry (Sym) and embedding degree (Emb); Where L1 and L2 are the lengths of the two pins, D1 and D2 are the embedment depths of the two pins, and T is the length of the pin. object Let be the average thickness of the object to be ligated; if Sym > θ simultaneously sym And θ emb_low <Emb<θ emb_high If the knot is sealed properly, then the knot is deemed acceptable.

6. The intelligent control method for a sealing machine according to claim 1, characterized in that: After the initial ligation attempt is deemed a failure, an adaptive religation strategy is implemented, including: Based on the failure mode of the first closure, the target torque T is adjusted according to preset rules. target If the nail feet are bent, reduce the torque; if the nail is not properly embedded, increase the torque. The re-tying operation is performed using the adjusted torque, and the success or failure of this operation is recorded in the training dataset of the tying decision model for online incremental learning of the model.

7. The intelligent control method for a sealing machine according to claim 1, characterized in that: It also includes predictive maintenance steps based on historical data: Continuously record the average operating current I of the motor during each tying operation. avg and no-load starting current I start ; When I avg Or I start The cumulative drift ΔI relative to its initial baseline value exceeds the maintenance threshold I. th At that time, a maintenance reminder signal is generated; 8. A sealing machine, used to implement the intelligent control method for a sealing machine as described in any one of claims 1 to 7, characterized in that, include: One-piece alloy die-cast housing; A closure component is provided at the front end of the housing; The sealing and extrusion end is slidably installed inside the housing and can move toward the sealing component; The drive mechanism includes a motor and a multimodal sensing module; A control unit, located inside the housing, is used to execute the intelligent control method.

9. The sealing machine according to claim 8, characterized in that, The multimodal sensing module integrates: A CMOS vision sensor facing the closure position is used to acquire high-resolution images; A laser rangefinder sensor, as the proximity sensor, is used to accurately detect the thickness of the object to be ligated; A high-precision Hall current sensor is used to monitor the operating current of the motor in real time.

Images