A blow and suck online locking screw method and system

Abstract

The application provides a kind of blow suction online lock screw method, comprising the following steps: step one, visual system scans workpiece identification code, calls corresponding process parameter set, real-time prediction workpiece thermal deformation based on deep learning model and generates lock trajectory pre-compensation data, feedback to PLC.The application calls process parameters by scanning identification code through visual system, real-time prediction thermal deformation combined with deep learning model and generates lock trajectory pre-compensation data feedback to PLC, which can improve lock precision, reduce manual intervention and debugging time, improve production efficiency and product quality.Screws are transported by compressed air cyclone suspension, avoiding screw touching pipe wall, correcting screw head posture by generating electrostatic adsorption field by electrode, ensuring screw smoothly entering chuck, preventing screw and electric screwdriver alignment failure leading to lock screw operation interruption.Locking program is real-time compensated and dynamically corrected by visual identification module and infrared temperature measurement module, which can be real-time adjusted and optimized according to actual production situation.

Description

Technical Field

[0001] This invention relates to an online screw-locking method, and more particularly to an online screw-locking method and system using blow-suction, belonging to the field of automated assembly technology. Background Technology

[0002] In many manufacturing industries such as electronics, automobiles, and home appliances, product assembly often involves the tightening of a large number of screws. Traditional manual screw tightening methods are not only labor-intensive and inefficient, but the tightening quality is also greatly affected by factors such as worker skill level and fatigue, making it difficult to guarantee product consistency and stability. As the manufacturing industry's requirements for production efficiency, product quality, and cost control continue to increase, automated screw tightening technology has emerged and is gradually becoming a key link in modern manufacturing production lines.

[0003] Currently, online screw fastening equipment has been applied in many enterprises. Among them, the blow-suction online screw fastening method uses a pneumatic system to blow screws from the feeder to the screwdriver nozzle, and then uses suction to adsorb the screws onto the screwdriver nozzle, realizing fast and accurate screw feeding and positioning, and then completes the fastening action with the help of an electric screwdriver.

[0004] In the operation of online screw fastening equipment, the workpiece needs to be precisely positioned and fixed beforehand to ensure that the screws are accurately fastened to the designated position. However, in actual production, there will be certain deviations in the size, shape, and position of the workpiece. Most existing equipment lacks the ability to automatically identify the workpiece and cannot adjust in real time according to the actual position of the workpiece. It can only perform the fastening operation according to the preset program. As a result, when there are deviations in the workpiece, the screw cannot be accurately fastened to the target position, leading to fastening failure.

[0005] Alignment of the screw and electric screwdriver is crucial for ensuring screw fastening quality. Existing blower-suction online screw fastening equipment frequently experiences alignment failures during screw feeding and positioning. Problems such as screw jamming or flipping in the feeder lead to incorrect screw posture, preventing smooth entry into the screwdriver nozzle, resulting in alignment failure, interruption of the screw fastening operation, and impact on production efficiency.

[0006] Most existing blow-suction online screw fastening equipment adopts traditional control methods and has relatively simple functions. During operation, the equipment mainly relies on preset programs and parameters, and cannot be adjusted and optimized in real time according to the actual production situation. Manual intervention is required, which not only increases the labor intensity of operators, but also leads to production interruption.

[0007] To address this, a method and system for online screw fastening using a blower and suction system is proposed. Summary of the Invention

[0008] In view of this, the present invention provides a method and system for online screw tightening by blowing and suction, to solve or alleviate one of the technical problems existing in the prior art, and at least provides a beneficial alternative.

[0009] The technical solution of this invention is implemented as follows: a method for online screw fastening by blowing and suction, comprising the following steps: Step 1: The vision system scans the workpiece identification code, retrieves the corresponding process parameter set, predicts the workpiece thermal deformation in real time based on the deep learning model, generates locking trajectory pre-compensation data, and feeds it back to the PLC. A high-resolution industrial camera and infrared thermal imager array are deployed above the fastening station. The vision system automatically scans the QR code or barcode on the workpiece surface and retrieves the process parameter set pre-stored in the local database. At the same time, the infrared sensor performs high-speed scanning of the workpiece surface at a frequency of ≥5Hz, generates a real-time temperature distribution thermal map and transmits it to the edge computing unit. The edge computing unit starts the deformation prediction algorithm and generates incremental compensation instructions. After receiving the compensation vector, the programmable logic controller corrects the XYZ values ​​of the fastening coordinates, drives the precision slide to the new coordinate position, and optimizes the torque curve according to the real-time temperature. Step 2: The PLC triggers the start of the air-blowing screw feeder. Compressed air drives the screw to be suspended and conveyed through the vortex generator pipeline to the screwdriver bit inlet of the locking mechanism. After the system starts, the PLC sends a pulse command to the air-blowing screw feeder, which pushes the screw into the vortex generator through the compressed air pipeline. The inner wall of the generator pipeline is uniformly arranged with turbulence vanes. When compressed air is injected, the airflow forms a high-speed rotating vortex field under the guidance of the vanes. After entering the pipeline, the screw is suspended in the center of the airflow due to centripetal force, avoiding attitude deviation caused by collision with the pipeline wall. Step 3: Activate the annular electrode on the inner wall of the screwdriver bit to generate an electrostatic adsorption field, correct the screw head posture, and simultaneously start the combined negative pressure of the main suction port at the root of the screwdriver bit and the auxiliary suction port at the tip to fix the screw column and suppress radial vibration. When the screw arrives at the screwdriver bit inlet, the PLC sends a 10ms pulse width high voltage pulse to the ring electrode controller, generating an electrostatic adsorption field on the ring electrode on the inner wall of the screwdriver bit, causing the screw head to be attracted and aligned with the center axis of the electrode, correcting the posture deviation from the initial maximum of 5° to within 0.2°.

[0010] The synchronously started negative pressure system performs two-stage control. The main suction port at the root of the screwdriver bit is maintained at a constant pressure of -80kPa by a vacuum pump, forming a strong suction force on the screw shaft. The secondary suction port at the tip generates a negative pressure of -20kPa, which helps to suppress radial vibration generated during equipment operation. Step 4: The PLC coordinates the control of the X / Y / Z axis motion mechanism and drives the locking mechanism to complete the surface adaptive locking based on the pre-compensation data; After receiving the visual pre-compensation data, the PLC coordinates the three-axis motion mechanism to perform surface adaptive locking.

[0011] The locking mechanism employs a three-stage sinusoidal speed change strategy. For curved workpieces, the system calculates the normal tilt angle compensation in real time and drives the electric screwdriver universal joint to adjust its posture, ensuring that the screw axis is always perpendicular to the tangent plane of the curved surface. Step 5: Collect misaligned screws and waste materials to a cyclone centrifugal separator, where centrifugal force separates metal dust from reusable screws; The system transports the mixture of misaligned screws and metal dust collected at the fastening station to a cyclone centrifugal separator via a negative pressure pipeline. The core of the separator is a conical separation tower connected to a high-pressure centrifugal fan at the bottom. The mixed waste material generates centrifugal acceleration under the action of the rotating airflow. The metal dust enters the bag filter through the top outlet with the rising airflow, while the intact screws slide down the tower wall and are separated into different specifications by the bottom vibrating screening module. Step 6: Collect the lock-attached voiceprint signal in real time through a microphone, and identify the characteristics of slipped teeth and missing lock defects through a convolutional neural network; A ring microphone array is installed on the side of the locking mechanism to collect wideband acoustic signals in real time during the operation of the electric screwdriver. After the raw audio is converted by an ADC, the transient acoustic emission signal in the 3000-5000Hz frequency band is extracted by a digital bandpass filter. The preprocessing module performs a short-time Fourier transform to generate a time-spectrum diagram as input data. When the edge computing unit identifies a slipped tooth feature or a missing lock feature, it immediately sends an error code to the PLC.

[0012] Step 7: Upload the locking torque curve and defect type to the blockchain node. When the production line switches to a new workpiece, the locking strategy for similar cases is matched from the cloud distributed database and automatically loaded.

[0013] Further preferred embodiment: In step one, the process parameter set includes screw specifications, fastening coordinates, and torque curves. Workpiece temperature is measured using an infrared sensor. The infrared sensor scans the workpiece surface at a sampling frequency of ≥5Hz to generate a temperature distribution thermogram. A deformation prediction algorithm is used to calculate the workpiece's thermal deformation. This algorithm calculates the temperature gradient field based on the heat conduction equation, combines it with the workpiece material's thermal expansion coefficient α, and outputs a thermal deformation compensation vector ΔP=(Δx,Δy,Δz). Then, the thermal expansion coefficient α is matched according to the material property library, and the local deformation is calculated using the formula: The deformation trend within the next 2 seconds is predicted by using an LSTM neural network, and the compensation vector ΔP is dynamically updated.

[0014] A further preferred embodiment: In step two, the inner wall of the vortex generator pipe is provided with a 120° tilt angle baffle array.

[0015] Further preferred embodiment: In step three, the negative pressure of the main suction port at the root of the bit is -80 kPa, and the negative pressure of the secondary suction port at the tip of the bit is -20 kPa.

[0016] A further preferred embodiment: In step four, the locking mechanism employs a sinusoidal speed-changing strategy to control the locking rate during the locking phase.

[0017] A further preferred embodiment: In step five, the cyclone centrifugal separator generates a tangential airflow of >50m / s through a conical separation tower.

[0018] Further optimization: In step six, the transient acoustic emission signal in the 3-5kHz frequency band during the locking process is extracted, and the acoustic fingerprint patterns of slipped teeth and missing locks are classified using a ResNet-18 network.

[0019] A further preferred embodiment: In step seven, the cloud-based policy matching includes: Construct a process knowledge graph with screw specifications, workpiece materials, and fastening coordinates as keywords; The topological similarity between the new workpiece and historical cases is calculated using a graph neural network, and the top K-optimal locking parameter set is output. After verifying the validity of the parameters through a blockchain smart contract, they are sent to the edge executor.

[0020] A blow-suction online screw-locking system, applied to a blow-suction online screw-locking method, includes: Visual recognition module: Scans workpiece identification code and detects screw hole position in real time, outputs locking coordinates and thermal deformation compensation data; Pneumatic feeding module: Generates a cyclone flow field through a spiral baffle to suspend and convey the screw to the screwdriver bit; Online fastening module: Fastens screws to designated positions online; Negative pressure adsorption module: provides a combined negative pressure from the main suction port and the auxiliary suction port; Electrostatic adsorption component: Ring electrode to correct screw posture; X / Y / Z axis drive mechanism: Servo motor drives the movement of the online locking module; Voiceprint acquisition array: Multi-microphone circular array, real-time capture of 3-5kHz locked voiceprints; Cyclone centrifugal separator: generates tangential wind through a conical separation tower, and uses an electromagnetic flap door to classify and recycle screws; PLC main control unit: coordinates the timing of each module, processes sensor data, and executes compensation algorithms; Infrared temperature measurement module: Uses a 5Hz scanning frequency to scan the surface temperature of the workpiece.

[0021] Further preferred embodiment: the signal output terminals of the visual recognition module, the voiceprint acquisition array, and the infrared temperature measurement module are connected to the signal input terminals of the PLC main control unit, and the electrical output terminal of the PLC main control unit is electrically connected to the electrical input terminals of the pneumatic feeding module, the online locking module, the negative pressure adsorption module, the electrostatic adsorption component, the X / Y / Z axis drive mechanism, and the cyclone centrifugal separator.

[0022] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions: I. This invention retrieves process parameters by scanning identification codes using a vision system, and combines this with a deep learning model to predict thermal deformation in real time and generate pre-compensation data for the fastening trajectory, which is then fed back to the PLC. This can improve fastening accuracy, reduce manual intervention and debugging time, and improve production efficiency and product quality.

[0023] Second, this invention uses compressed air to suspend and transport screws in a cyclone, avoiding the screws from touching the tube wall. An electrostatic adsorption field is generated by the electrodes to correct the screw head posture, ensuring that the screws enter the screwdriver bit smoothly and preventing the screw tightening operation from being interrupted due to misalignment of the screw and the electric screwdriver.

[0024] Third, this invention uses a visual recognition module and an infrared temperature measurement module to compensate and dynamically correct the locking procedure in real time, which can be adjusted and optimized in real time according to the actual production situation without manual intervention, thus reducing the labor intensity of operators.

[0025] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart of the present invention; Figure 2 This is a system structure diagram of the present invention. Detailed Implementation

[0028] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0029] It should be understood that the following specific examples illustrate the implementation of this disclosure, and those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0030] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0031] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The drawings only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0032] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0033] Example 1 like Figure 1 As shown, this embodiment of the invention provides a method for online screw fastening using a blower and suction system, comprising the following steps: Step 1: The vision system scans the workpiece identification code and retrieves the corresponding process parameter set, which includes screw specifications, fastening coordinates, and torque curves. Based on a deep learning model, it predicts the workpiece's thermal deformation in real time and generates pre-compensation data for the fastening trajectory, which is then fed back to the PLC. The infrared sensor measures the workpiece temperature, scanning the workpiece surface at a sampling frequency of ≥5Hz to generate a temperature distribution thermogram. A deformation prediction algorithm is used to calculate the workpiece's thermal deformation. This algorithm calculates the temperature gradient field based on the heat conduction equation and, combined with the workpiece material's thermal expansion coefficient α, outputs a thermal deformation compensation vector ΔP=(Δx,Δy,Δz). Then, based on the material property library, the thermal expansion coefficient α is matched, and the local deformation is calculated using the formula: The deformation trend within the next 2 seconds is predicted by using an LSTM neural network, and the compensation vector ΔP is dynamically updated. A high-resolution industrial camera and infrared thermal imager array are deployed above the fastening station. When the conveyor belt carries the workpiece into the station, the vision system automatically scans the QR code or barcode on the workpiece surface and retrieves a set of process parameters pre-stored in the local database. This parameter set includes key data such as screw diameter and length specifications, three-dimensional coordinates of the fastening point, and staged torque curves. Simultaneously, the infrared sensor performs high-speed scanning of the workpiece surface at a frequency of ≥5Hz, generating a real-time temperature distribution thermal map and transmitting it to the edge computing unit. The edge computing unit initiates a deformation prediction algorithm. First, it constructs a temperature gradient field model inside the workpiece based on the heat conduction equation, accurately locating the high-temperature expansion zone and the low-temperature contraction zone. Then, it matches the coefficient of thermal expansion in the material property database according to the workpiece's unique identifier, and for the micro-region surrounding each fastening point, it calculates the coefficient of thermal expansion based on: Calculate the local deformation and integrate all micro-area data to generate a three-dimensional compensation vector. The values ​​of the vectors represent the deformation compensation amounts along the X, Y, and Z axes, respectively. For example, a positive value on the Z axis indicates that the compression depth needs to be increased.

[0034] To cope with dynamic temperature changes, the system inputs 20 consecutive frames of temperature data into a 128-node LSTM neural network to predict the deformation trend within the next two seconds and update the compensation value in real time. For example, when it predicts that the Z-direction deformation will increase from 0.05 mm to 0.07 mm in 0.5 seconds, an incremental compensation command is generated. After receiving the compensation vector, the programmable logic controller performs a triple adjustment: correcting the XYZ values ​​of the locking coordinates, driving the precision slide to the new coordinate position, and optimizing the torque curve based on the real-time temperature.

[0035] This invention retrieves process parameters by scanning identification codes using a vision system, and combines this with a deep learning model to predict thermal deformation in real time and generate pre-compensation data for the fastening trajectory, which is then fed back to the PLC. This improves fastening accuracy, reduces manual intervention and debugging time, and enhances production efficiency and product quality.

[0036] Step 2: The PLC triggers the start of the air-blowing screw feeder. The inner wall of the vortex generator pipe is equipped with a 120° tilting baffle array. Compressed air drives the screw to form a vortex suspension and convey it through the vortex generator pipe, reaching the screwdriver bit inlet of the locking mechanism. After the system starts, the PLC sends a pulse command to the air-blowing screw feeder. The screw selection mechanism inside the feeder separates a single screw from the vibrating hopper and pushes it into the vortex generator through the compressed air pipeline. The inner wall of the generator pipeline is uniformly arranged with turbulence vanes, each group of vanes spirally distributed at a 120-degree angle. When compressed air is injected at a pressure of 0.5 MPa, the airflow forms a high-speed rotating vortex field under the guidance of the vanes. After entering the pipeline, the screw is suspended in the center of the airflow by centripetal force, avoiding attitude deviation caused by collision with the pipe wall.

[0037] When the screw arrives at the screwdriver bit inlet, the infrared photoelectric sensor at the end of the pipe confirms the arrival signal, and at the same time the air circuit system switches to low-pressure mode to prevent the screw from bouncing due to airflow impact.

[0038] To ensure the safe operation of the system, the air pressure monitoring module continuously compares the pressure values ​​at the feeding end and the nozzle end. When the pressure difference between the two ends exceeds 30 kPa, an alarm is automatically triggered and the air supply is cut off. The maintenance cycle is set to be after 500,000 screws have been continuously run, after which the pipeline needs to be disassembled and the dust accumulated on the baffle plate is removed by an ultrasonic cleaner.

[0039] Step 3: Activate the annular electrode on the inner wall of the bit to generate an electrostatic adsorption field, correct the screw head posture, and simultaneously start the combined negative pressure of the main suction port at the root of the bit and the auxiliary suction port at the tip of the bit to fix the screw column and suppress radial vibration. The negative pressure of the main suction port at the root of the bit is -80kPa, and the negative pressure of the auxiliary suction port at the tip of the bit is -20kPa. When the screw arrives at the screwdriver bit inlet, the PLC sends a 10ms pulse width high voltage pulse to the ring electrode controller, generating an electrostatic adsorption field on the copper-tungsten alloy ring electrode on the inner wall of the screwdriver bit, causing the screw head to be attracted and aligned with the center axis of the electrode, correcting the posture deviation from the initial maximum of 5° to within 0.2°.

[0040] The synchronously started negative pressure system performs two-stage control. The main suction port at the root of the screwdriver bit is maintained at a constant pressure of -80kPa by a vacuum pump, forming a strong suction force on the screw shaft. The secondary suction port at the tip generates a negative pressure of -20kPa, which helps to suppress radial vibration generated during equipment operation.

[0041] To ensure adsorption stability, the main suction port adopts an annular cavity design with a diameter of 0.3 mm, while the secondary suction port is an array of twelve micropores with a diameter of 0.5 mm. The two are connected by independent gas paths to avoid pressure crosstalk.

[0042] During operation, the embedded pressure sensor monitors the pressure difference between the main and auxiliary suction ports in real time. When the pressure at the main suction port is higher than -75 kPa, the system immediately triggers a red alarm and suspends the clamping process.

[0043] For screws of different specifications, operators can adjust parameter combinations through the human-machine interface to change the air pressure according to the actual situation.

[0044] Step 4: The PLC coordinates the X / Y / Z axis motion mechanism and drives the locking mechanism to complete the surface adaptive locking based on the pre-compensation data. The locking mechanism adopts a sine wave speed change strategy to control the locking rate during the locking stage. After receiving the visual pre-compensation data, the PLC coordinates the three-axis motion mechanism to perform surface adaptive locking.

[0045] The locking mechanism employs a three-stage sinusoidal speed-changing strategy: High-speed rotation stage: The rotation speed stabilizes at 200 RPM, and the torque threshold reaches 70% of the set value; During the oscillating seating stage: the rotational speed fluctuates between 40-120 RPM according to the formula n=40+160·sin(2π·20·t), and is combined with ±5° angle micro-oscillation to eliminate thread clearance; Final tightening stage: The rotation speed is linearly increased to 100 RPM to complete the final tightening.

[0046] For curved workpieces, the system calculates the normal tilt angle compensation in real time: The height values ​​of four points around the locking point are obtained using a laser rangefinder. Through formula Calculate the XY plane tilt angle (d=5mm is the distance between measuring points), drive the electric screwdriver universal joint to adjust the posture, and ensure that the screw axis is always perpendicular to the tangent plane of the curved surface.

[0047] Step 5: Collect misaligned screws and waste materials to the cyclone centrifugal separator. The cyclone centrifugal separator generates a tangential airflow of >50m / s through a conical separation tower, and separates metal dust and reusable screws by centripetal force. The system transports the mixture of misaligned screws and metal dust collected at the fastening station to a cyclone centrifugal separator via a negative pressure pipeline. The core of the separator is a stainless steel conical separation tower, with a high-pressure centrifugal fan at the bottom generating a 52 m / s tangential airflow. After the mixed waste is injected through the tangential inlet at the top of the tower, it generates centrifugal acceleration under the action of the rotating airflow. The metal dust, due to its low density, is thrown towards the central area and enters the bag filter through the top outlet with the rising airflow. The intact screws, due to their high density, slide down the tower wall tightly under the action of centrifugal force, and are separated into different specifications by the bottom vibrating screening module.

[0048] Step 6: Collect the acoustic fingerprint signal of the lock in real time through the microphone, identify the defects of slipped teeth and missing locks through the convolutional neural network, extract the transient acoustic emission signal in the 3-5kHz frequency band during the locking process, and use the ResNet-18 network to classify the acoustic fingerprint patterns of slipped teeth and missing locks. A ring microphone array is installed on the side of the locking mechanism to collect wideband acoustic fingerprint signals in real time during the operation of the electric screwdriver. The raw audio is converted by a 24-bit ADC, and then a digital bandpass filter is used to extract transient acoustic emission signals in the 3000-5000Hz frequency band, which contains the characteristic acoustic fingerprints of screw and thread engagement. The preprocessing module performs a short-time Fourier transform to generate a time-frequency spectrum as input data. When the edge computing unit identifies a slippage feature (a narrowband howling sound at 4.2kHz exceeding 300ms in the spectrum) or a missing lock feature (energy loss >40% at 3.5kHz), it immediately sends an error code to the PLC.

[0049] Step 7: Upload the locking torque curve and defect type to the blockchain node. When the production line switches to a new workpiece, the locking strategy for similar cases is matched from the cloud-based distributed database and automatically loaded. The cloud-based strategy matching includes: Construct a process knowledge graph with screw specifications, workpiece materials, and fastening coordinates as keywords; The topological similarity between the new workpiece and historical cases is calculated using a graph neural network, and the top K-optimal locking parameter set is output. After verifying the validity of the parameters through a blockchain smart contract, they are sent to the edge executor.

[0050] After fastening is complete, the PLC encrypts and packages the torque curve data and defect codes, and uploads them to the Hyperledger Fabric blockchain network via the OPC UA protocol. Each block contains: workpiece hash ID, screw specifications, material tag, fastening coordinates (X, Y, Z), ambient temperature and humidity, and quality judgment result. The cloud server node types cover screw specifications, workpiece materials, fastening coordinates, edge relationships, and other weighted parameters.

[0051] When the production line switches to a new workpiece, the system performs graph neural network matching.

[0052] Example 2 A blow-suction online screw-locking system, applied to a blow-suction online screw-locking method, comprising: Visual recognition module: Scans workpiece identification code and detects screw hole position in real time, outputs locking coordinates and thermal deformation compensation data; Pneumatic feeding module: Generates a cyclone flow field through a spiral baffle to suspend and convey the screw to the screwdriver bit; Online fastening module: Fastens screws to designated positions online; Negative pressure adsorption module: provides a combined negative pressure from the main suction port and the auxiliary suction port; Electrostatic adsorption component: Ring electrode to correct screw posture; X / Y / Z axis drive mechanism: Servo motor drives the movement of the online locking module; Voiceprint acquisition array: Multi-microphone circular array, real-time capture of 3-5kHz locked voiceprints; Cyclone centrifugal separator: generates tangential wind through a conical separation tower, and uses an electromagnetic flap door to classify and recycle screws; PLC main control unit: coordinates the timing of each module, processes sensor data, and executes compensation algorithms; Infrared temperature measurement module: Uses a 5Hz scanning frequency to scan the surface temperature of the workpiece.

[0053] In this embodiment, specifically: the signal output terminals of the visual recognition module, the voiceprint acquisition array, and the infrared temperature measurement module are connected to the signal input terminals of the PLC main control unit; the electrical output terminals of the PLC main control unit are electrically connected to the electrical input terminals of the pneumatic feeding module, the online locking module, the negative pressure adsorption module, the electrostatic adsorption component, the X / Y / Z axis drive mechanism, and the cyclone centrifugal separator.

[0054] The visual recognition module scans the workpiece's QR code using an industrial camera, retrieves a pre-stored set of process parameters, and the infrared temperature measurement module scans the workpiece surface temperature field data at a 5Hz frequency. A deep learning model then predicts the thermal deformation and generates pre-compensation parameters for the fastening trajectory. The PLC main control unit activates the pneumatic feeding module, where spiral baffles create a spiral cyclone flow field within the conveying pipe, causing the screw to be suspended and conveyed to the bit inlet.

[0055] When the screw reaches the screwdriver bit, the negative pressure adsorption module activates the composite negative pressure system, while the annular electrode of the electrostatic adsorption component applies a high-frequency pulsed electric field to correct the screw's axial posture. The X / Y / Z axis drive mechanism, based on the compensation coordinates output by the vision module, drives the servo motor assembly to perform three-axis linkage interpolation motion.

[0056] During the locking process, the microphone ring array of the acoustic fingerprint acquisition array captures acoustic fingerprint signals in the 3-5kHz frequency band, and uses beamforming algorithms to locate abnormal sound sources and monitor faults such as slippage and floating lock in real time.

[0057] By scanning the identification code with a vision system to retrieve process parameters, and combining this with a deep learning model to predict thermal deformation in real time and generate pre-compensation data for the fastening trajectory, the fastening accuracy can be improved, manual intervention and debugging time can be reduced, and production efficiency and product quality can be increased.

[0058] This invention uses compressed air to suspend and transport screws, preventing them from touching the tube wall. An electrostatic adsorption field is generated by electrodes to correct the screw head posture, ensuring that the screw enters the screwdriver bit smoothly and preventing the screw tightening operation from being interrupted due to misalignment between the screw and the electric screwdriver.

[0059] This invention uses a visual recognition module and an infrared temperature measurement module to compensate for and dynamically correct the locking procedure in real time. It can be adjusted and optimized in real time according to the actual production situation without manual intervention, thus reducing the labor intensity of operators.

[0060] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0061] The steps in the method of this invention can be adjusted, combined, or deleted according to actual needs.

[0062] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented using electronic pre-installed hardware, or a combination of computer software and electronic pre-installed hardware. Whether these functions are implemented in pre-installed hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0063] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for online screw tightening using a blower and suction system, characterized in that, Includes the following steps: Step 1: The vision system scans the workpiece identification code, retrieves the corresponding process parameter set, predicts the workpiece's thermal deformation in real time based on a deep learning model, generates pre-compensation data for the locking trajectory, and feeds it back to the PLC. In step one, the process parameter set includes screw specifications, fastening coordinates, and torque curves. Workpiece temperature is measured using an infrared sensor that scans the workpiece surface at a sampling frequency of ≥5Hz to generate a temperature distribution thermogram. A deformation prediction algorithm is used to calculate the workpiece's thermal deformation. This algorithm calculates the temperature gradient field based on the heat conduction equation and, combined with the workpiece material's thermal expansion coefficient α, outputs a thermal deformation compensation vector ΔP=(Δx,Δy,Δz). Then, the thermal expansion coefficient α is matched according to a material property library, and the local deformation is calculated using the following formula: The deformation trend within the next 2 seconds is predicted by using an LSTM neural network, and the compensation vector ΔP is dynamically updated. Step 2: The PLC triggers the start of the air-blowing screw feeder. Compressed air drives the screw to be suspended and conveyed through the vortex generator pipeline to the screwdriver bit inlet of the locking mechanism. Step 3: Activate the annular electrode on the inner wall of the screwdriver bit to generate an electrostatic adsorption field, correct the screw head posture, and simultaneously start the combined negative pressure of the main suction port at the root of the screwdriver bit and the auxiliary suction port at the tip to fix the screw column and suppress radial vibration. Step 4: The PLC coordinates the control of the X / Y / Z axis motion mechanism and drives the locking mechanism to complete the surface adaptive locking based on the pre-compensation data; Step 5: Collect misaligned screws and waste materials to a cyclone centrifugal separator, where centrifugal force separates metal dust from reusable screws; Step 6: Collect the lock-attached voiceprint signal in real time through a microphone, and identify the characteristics of slipped teeth and missing lock defects through a convolutional neural network; Step 7: Upload the locking torque curve and defect type to the blockchain node. When the production line switches to a new workpiece, the locking strategy for similar cases is matched from the cloud distributed database and automatically loaded.

2. The method for online screw tightening by blowing and suction according to claim 1, characterized in that, In step two, the inner wall of the vortex generator pipe is provided with a 120° tilt angle baffle array.

3. The method for online screw tightening by blowing and suction according to claim 1, characterized in that, In step three, the negative pressure at the main suction port at the root of the bit is -80 kPa, and the negative pressure at the secondary suction port at the tip of the bit is -20 kPa.

4. The method for online screw tightening by blowing and suction according to claim 1, characterized in that, In step four, the locking mechanism employs a sinusoidal speed-changing strategy to control the locking rate during the locking phase.

5. The method for online screw fastening by blowing and suction according to claim 1, characterized in that, In step five, the cyclone centrifugal separator generates a tangential airflow of >50m / s through a conical separation tower.

6. The method for online screw tightening by blowing and suction according to claim 1, characterized in that, In step six, transient acoustic emission signals in the 3-5kHz frequency band are extracted during the locking process, and the acoustic fingerprint patterns of slipped teeth and missing locks are classified using a ResNet-18 network.

7. The method for online screw fastening by blowing and suction according to claim 1, characterized in that, In step seven, the matching of lock-attachment strategies for similar cases in the cloud-based distributed database includes: Construct a process knowledge graph with screw specifications, workpiece materials, and fastening coordinates as keywords; The topological similarity between the new workpiece and historical cases is calculated using a graph neural network, and the top K-optimal locking parameter set is output. After verifying the validity of the parameters through a blockchain smart contract, they are sent to the edge executor.

8. A blow-suction online screw-locking system, applied to the blow-suction online screw-locking method as described in any one of claims 1-7, characterized in that, include: Visual recognition module: Scans workpiece identification code and detects screw hole position in real time, outputs locking coordinates and thermal deformation compensation data; Pneumatic feeding module: Generates a cyclone flow field through a spiral baffle to suspend and convey the screw to the screwdriver bit; Online fastening module: Fastens screws to designated positions online; Negative pressure adsorption module: provides a combined negative pressure from the main suction port and the auxiliary suction port; Electrostatic adsorption component: Ring electrode to correct screw posture; X / Y / Z axis drive mechanism: Servo motor drives the movement of the online locking module; Voiceprint acquisition array: Multi-microphone circular array, real-time capture of 3-5kHz locked voiceprints; Cyclone centrifugal separator: generates tangential wind through a conical separation tower, and uses an electromagnetic flap door to classify and recycle screws; PLC main control unit: coordinates the timing of each module, processes sensor data, and executes compensation algorithms; Infrared temperature measurement module: Uses a 5Hz scanning frequency to scan the surface temperature of the workpiece. The signal output terminals of the visual recognition module, the voiceprint acquisition array, and the infrared temperature measurement module are connected to the signal input terminals of the PLC main control unit. The electrical output terminal of the PLC main control unit is electrically connected to the electrical input terminals of the pneumatic feeding module, the online locking module, the negative pressure adsorption module, the electrostatic adsorption component, the X / Y / Z axis drive mechanism, and the cyclone centrifugal separator.

Images