Feeding control system and method for roller parts

Abstract

The invention discloses a feeding control system and method for roller parts. The feeding control system comprises a flexible vibration unit, a visual feedback unit, a four-axis self-adaptive transfer device, a dynamic sensing module and an intelligent decision center. The flexible vibration unit adjusts the poses of the roller parts through multi-frequency composite vibration, and the visual feedback unit recognizes the poses of the parts in the positioning grooves and generates a transfer grabbing instruction; and the transfer device executes adsorption carrying operation. The dynamic sensing module monitors material characteristics, vibration states, environmental parameters and groove shape data in real time, and the intelligent decision center constructs a positioning interference prediction model and outputs a repair instruction. The system forms a closed-loop process of vibration, detection, decision making and repairing, and high-precision and stable feeding of roller parts is achieved. The method has good self-adaptive regulation and control capability and positioning precision, and is suitable for intelligent manufacturing scenes in complex environments.

Description

Technical Field

[0001] This invention relates to the field of automated assembly control technology, and in particular to a feeding control system and method for roller parts. Background Technology

[0002] In automated manufacturing and intelligent assembly scenarios, roller components are key transmission components, and their loading accuracy directly affects the stability and efficiency of the entire assembly chain. Especially in mass production, how to quickly and accurately orient and place roller components has become a core technical challenge. To solve the loading deviation caused by the random posture of parts, existing methods generally rely on mechanical structure guidance or single-frequency vibration adjustment, but these methods have certain limitations in terms of adaptability and positioning accuracy.

[0003] First, due to the axially symmetrical shape and easy rolling of the roller components, conventional devices are prone to misidentification and missed loading when handling their attitude adjustment. Second, environmental changes (such as temperature and humidity gradients) have a significant impact on vibration response characteristics, making it difficult for traditional systems to maintain consistent positioning stability. In addition, changes in the material properties of the component's support platform (such as fatigue degradation or elastic loss) can also lead to uncontrolled vibration during long-term operation, thereby reducing the success rate of loading.

[0004] Furthermore, in complex production environments, the positioning groove structure undergoes micro-scale deformation due to frequent use. Traditional devices lack sophisticated monitoring and repair mechanisms, making it difficult to promptly detect and correct the impact of these "micro-defects" on positioning accuracy. These problems are widespread in practical engineering applications, necessitating a comprehensive solution with adaptive sensing capabilities, dynamic feedback control mechanisms, and in-situ micro-scale repair capabilities to achieve high-precision, robust, and intelligent feeding control for roller parts. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a feeding control system and method for roller parts, which can realize intelligent feeding control of roller parts with high precision, strong stability and adaptive repair capability.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a feeding control system for roller parts, comprising:

[0007] The collaborative control platform integrates:

[0008] A flexible vibration unit has an array of part positioning grooves on its bearing surface. The flexible vibration unit is configured to adjust the position and orientation of the roller parts through multi-frequency composite vibration.

[0009] A visual feedback unit, suspended above the flexible vibration unit, is configured to acquire the pose image of the roller parts in the positioning groove of each part after each vibration cycle, and to identify the pose image to form a pose recognition result. When the pose recognition result shows that the surface of the roller part is horizontal with the positioning groove of the part, a transfer and gripping command is generated.

[0010] A four-axis adaptive transfer device is located on the side of the flexible vibration unit. Its end effector integrates a negative pressure adsorption device. The four-axis adaptive transfer device drives the negative pressure adsorption device to move above the corresponding roller part according to the transfer gripping command, so as to adsorb the roller part.

[0011] The dynamic sensing module, connected to the collaborative control platform, includes:

[0012] A material property sensing unit is used to acquire the matrix material parameters of the flexible vibration unit bearing substrate in real time. The matrix material parameters include elastic modulus, damping coefficient, and fatigue life degradation coefficient.

[0013] A vibration state monitoring unit is used to detect vibration dynamic parameters in real time, including vibration frequency and three-dimensional amplitude vector.

[0014] An environmental coupling sensing unit is used to monitor environmental parameters in real time, including the distribution of ambient temperature and humidity.

[0015] The deformation tracking unit is used to acquire the topological data of the positioning slots of each part in real time through laser micron scanning;

[0016] The intelligent decision-making center, connected to the dynamic sensing module, includes:

[0017] Multiphysics coupling unit is used to construct a location disturbance prediction model by integrating elastic deformation theory, vibration energy transfer equation and environmental creep function;

[0018] An adaptive decision-making unit, connected to the multiphysics coupling unit, is used to input the real-time collected matrix material parameters, vibration dynamic parameters, environmental parameters, and trough topology data into the positioning interference prediction model and output a comprehensive positioning interference index; when the comprehensive positioning interference index is greater than a dynamic threshold, a trough repair instruction set is generated.

[0019] The in-situ repair execution unit is connected to the adaptive decision unit and is used to perform laser recladding or micro-stamping shaping on the target positioning groove according to the groove repair instruction set.

[0020] The visual feedback unit, the dynamic perception module, and the intelligent decision-making center form a closed-loop control circuit, driving the flexible vibration unit to perform an iterative optimization process of "vibration-detection-decision-repair".

[0021] Furthermore, a servo-driven rotation mechanism is provided between the end effector of the four-axis adaptive transfer device and the negative pressure adsorption device, and a positioning notch is provided on the axial end face of the roller part;

[0022] The feeding control system further includes a closed-loop alignment module, connected to the collaborative control platform, including:

[0023] The notch identification unit is mounted on a fixed bracket on the side of the flexible vibration unit and configured as follows:

[0024] After the negative pressure adsorption device adsorbs the roller part, an axial end face image containing the positioning notch is acquired;

[0025] The notch azimuth angle in the axial end face image is identified based on a convolutional neural network.

[0026] A rotation compensation command is generated when the azimuth angle of the notch deviates from the preset reference angle.

[0027] A transfer release command is generated when the azimuth angle of the notch matches a preset reference angle.

[0028] The pose correction unit, connected to the notch recognition unit and the servo-driven rotation mechanism, is configured as follows:

[0029] Analyze the angle deviation value in the rotation compensation command;

[0030] The servo-driven rotary mechanism is controlled to rotate the roller component around the Z-axis by the angular deviation value.

[0031] The linkage control unit, connected to the notch recognition unit, is configured as follows:

[0032] In response to the transfer release command, the transfer trajectory of the negative pressure adsorption device to the next workstation tray is planned;

[0033] The negative pressure release signal is triggered when the target placement slot is directly above the target placement slot.

[0034] Furthermore, the in-situ repair execution unit includes:

[0035] The laser energy modulation subunit is configured as follows:

[0036] The target laser power reference value P is determined by consulting the laser power mapping table based on the fatigue life degradation coefficient λ. base ;

[0037] Deformation depth h calculated based on groove topology data d According to the dynamic power formula Adjust the output power, where κ represents the material correction factor;

[0038] The micro-stamping compensation subunit is configured as follows:

[0039] Real-time acquisition of vibration frequency f vib When f vib Activate impulse compensation at >50Hz: F comp =F0[1+0.02(f vib -50)], where F0 represents the reference punching force, i.e. f vib = Impulse force at 50Hz, F comp Indicates the impact force after compensation;

[0040] Based on the three-dimensional amplitude vector The transverse vibration component is obtained by decomposition, and a reverse compensating force is applied to the stamping head. Where K represents the stiffness coefficient. This represents the lateral vibration component.

[0041] Furthermore, the adaptive decision-making unit includes:

[0042] The threshold evolution subunit is used to perform dynamic threshold calculation:

[0043] T dyn =T0+α T ·ΔT+α H ·|ΔRH-40%|+α f ·log 10 (f vib / 10);

[0044] Among them, T dyn The dynamic threshold is represented by T0, the baseline threshold is represented by ΔT, the temperature change is represented by ΔRH, and the relative humidity deviation is represented by α. T α H α f These represent the preset first configuration coefficient, second configuration coefficient, and third configuration coefficient, respectively;

[0045] Repair the priority decision subunit and configure it as follows:

[0046] When the comprehensive positioning interference index δ > 1.2T dyn Emergency repair instructions are generated in a timely manner;

[0047] When T dyn <δ≤1.2T dyn A standby repair command is generated in time.

[0048] Furthermore, it also includes a density-adaptive vibration module, connected to the collaborative control platform, comprising:

[0049] The component density detection unit calculates the roller component density ρ based on the pixel grayscale distribution of the pose image. part ;

[0050] The spectrum optimization unit, connected to the part density detection unit, is configured as follows:

[0051] According to the density ρ of the roller parts part Matching vibration spectrum group [f min f opt f max ], where f min f opt f max These represent the minimum permissible frequency, the optimal operating frequency, and the maximum safe frequency, respectively.

[0052] The amplitude compensation unit, connected to the spectrum optimization unit, is used to receive the three-dimensional amplitude vector in real time. And execute:

[0053] when At that time, according to ΔV=β·(ρ part -ρ0) Corrects the driving voltage;

[0054] in ρ represents the theoretical amplitude, β represents the density-voltage coefficient, ρ0 represents the density of the reference part, and ΔV represents the amplitude deviation modulus.

[0055] Furthermore, the closed-loop alignment module and the visual feedback unit interact through a pose fusion unit:

[0056] The tilt angle θ of the part receiving the pose recognition result vib ;

[0057] The component tilt angle θ is calculated from the azimuth angle of the receiving notch. gap ;

[0058] When |θ vib -θ gap |When the angle is greater than 5°, a vibration parameter readjustment command is triggered;

[0059] The vibration-transfer co-controller responds to the vibration parameter readjustment command by executing:

[0060] Calculate the frequency adjustment amount Δf adj =0.1×|θ vib -θ gap |;

[0061] Calculate the transit delay time Δt delay=0.2×|θ vib -θ gap |

[0062] Furthermore, the visual feedback unit and the flexible vibration unit constitute a real-time feedback link, including:

[0063] The phase synchronization controller is configured as follows:

[0064] Let the current vibration phase angle be φ (φ∈[0,2π]), in Image acquisition is triggered when (k is an integer);

[0065] The pose prediction compensator, connected to the phase synchronization controller, performs the following:

[0066] Calculate the displacement vector of the part based on the pose images of adjacent vibration cycles.

[0067] According to the amplitude compensation formula Correct the amplitude of the next cycle, where denoted as the corrected amplitude vector, and μ represents the compensation coefficient.

[0068] Furthermore, the multiphysics coupling unit performs:

[0069] Environment creep function construction:

[0070]

[0071] Where, ∈ c σ represents creep strain, E0 represents material stress, Q represents the reference elastic modulus, R represents the activation energy, T represents the ambient temperature, t represents time, and n represents the stress exponent.

[0072] Constructing a dynamic damage model:

[0073] D d =1-(1-λ) 50ΔRH ;

[0074] Among them, D d The fatigue damage factor is represented by ΔRH, which represents the relative humidity deviation value, calculated based on the humidity distribution.

[0075] The creep strain ∈ c and fatigue damage factor D d The comprehensive positioning interference index δ is generated by inputting the positioning interference prediction model.

[0076] Furthermore, the in-situ repair execution unit includes:

[0077] The trough-shaped classification subunit classifies deformations based on the trough-shaped topology data into:

[0078] Microdeformation (h) d ≤0.1mm) → Activate laser thermal compensation mode;

[0079] Moderate deformation (0.1mm < h) d ≤0.3mm) → Activate micro-stamping mode;

[0080] Severe deformation (h) d >0.3mm) → Activate laser-stamping composite mode;

[0081] The energy optimization subunit is connected to the slot-shaped classification subunit and performs energy repair:

[0082]

[0083] Among them, W total The total repair energy is represented by k1, k2, and c, which represent the preset first mode coefficient, second mode coefficient, and third mode coefficient, respectively. α and β represent the first weighting factor and the second weighting factor, respectively.

[0084] A method for controlling the feeding of roller parts, applied to the aforementioned feeding control system for roller parts, includes:

[0085] Step S1: The flexible vibration unit adjusts the position and orientation of the roller parts through multi-frequency composite vibration; the visual feedback unit collects the position and orientation images of the roller parts in the positioning slots of each part after each vibration cycle, and identifies the position and orientation images to form a position and orientation recognition result. When the position and orientation recognition result shows that the surface of the roller parts is horizontal with the positioning slots of the parts, a transfer and gripping command is generated.

[0086] Step S2: The four-axis adaptive transfer device drives the negative pressure adsorption device to move above the corresponding roller part according to the transfer gripping command, so as to adsorb the roller part.

[0087] Step S3: The material property sensing unit acquires the matrix material parameters of the flexible vibration unit bearing substrate in real time; the vibration state monitoring unit detects the vibration dynamic parameters; the environmental coupling sensing unit monitors the environmental parameters in real time; and the deformation tracking unit acquires the groove topology data of each part positioning groove in real time through laser micron scanning.

[0088] Step S4: The multiphysics coupling unit constructs a positioning interference prediction model by integrating elastic deformation theory, vibration energy transfer equation and environmental creep function; the adaptive decision unit inputs the real-time collected matrix material parameters, vibration dynamic parameters, environmental parameters and groove topology data into the positioning interference prediction model and outputs a comprehensive positioning interference index; when the comprehensive positioning interference index is greater than the dynamic threshold, a groove repair instruction set is generated.

[0089] Step S5: The in-situ repair execution unit performs laser recladding or micro-stamping shaping on the target positioning groove according to the groove repair instruction set.

[0090] Step S6: The visual feedback unit, the dynamic perception module, and the intelligent decision-making center form a closed-loop control circuit, driving the flexible vibration unit to execute the iterative optimization process of "vibration-detection-decision-repair".

[0091] The beneficial effects of this invention are:

[0092] (1) Improve the accuracy and efficiency of roller component posture adjustment: Multi-frequency composite vibration is achieved through flexible vibration unit, and the position and posture of roller component are identified in real time by visual feedback unit, which effectively improves the success rate of component orientation adjustment and overcomes the limitations of traditional vibrators in handling axisymmetric and easily rolling workpieces.

[0093] (2) Achieve intelligent and adaptive feeding control: The four-axis adaptive transfer device works in conjunction with the vision system to automatically complete the adsorption and gripping operation based on the recognition results. It has flexible and precise control capabilities, reducing the risk of manual intervention and misoperation.

[0094] (3) Enhance the robustness of the system to environmental changes: The dynamic sensing module can monitor external disturbance factors such as temperature, humidity, and vibration in real time, and sense the material properties and deformation data of the platform, so that the system has the ability to adapt to complex manufacturing environments and ensure the long-term stable operation of the system.

[0095] (4) Achieve in-situ repair and long-term accuracy maintenance of positioning structure: Through the positioning interference prediction model constructed by deformation tracking and intelligent decision center, the system can provide early warning of deformation of part positioning groove and automatically issue laser re-cladding or micro-stamping commands to extend the service life of equipment and ensure the continuous reliability of feeding accuracy.

[0096] (5) Constructing a closed-loop iterative optimization mechanism of "perception-analysis-repair": This invention forms an integrated closed-loop control process, realizes the intelligent evolution and self-optimization of the material feeding process, and adapts to the future development trend of intelligent manufacturing in industry. Attached Figure Description

[0097] Figure 1 This is a schematic diagram of the feeding control system for roller parts in this invention;

[0098] Figure 2 This is an isometric view of the collaborative control platform in this invention;

[0099] Figure 3 This is a flowchart of the steps in the feeding control method for roller parts in this invention.

[0100] Reference numerals: 1. Collaborative control platform; 11. Flexible vibration unit; 12. Visual feedback unit; 13. Four-axis adaptive transfer device; 2. Dynamic sensing module; 21. Material property sensing unit; 22. Vibration state monitoring unit; 23. Environmental coupling sensing unit; 24. Deformation tracking unit; 3. Intelligent decision-making center; 31. Multi-physics coupling unit; 32. Adaptive decision-making unit; 321. Threshold evolution subunit; 322. Repair priority decision-making subunit; 33. In-situ repair execution unit; 331. Laser energy control subunit; 332. Micro-stamping compensation subunit; 333. Groove classification subunit; 334. Energy optimization subunit; 4. Servo-driven rotation mechanism; 5. Closed-loop alignment module; 51. Notch recognition unit; 52. Pose correction unit; 53. Linkage control unit; 6. Density adaptive vibration module; 61. Part density detection unit; 62. Spectrum optimization unit; 63. Amplitude compensation unit. Detailed Implementation

[0101] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Identical components are denoted by the same reference numerals. It should be noted that the terms "front," "rear," "left," "right," "upper," and "lower" used in the following description refer to directions in the accompanying drawings, and the terms "bottom surface," "top surface," "inner," and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0102] Example 1, referring to Figure 1 This embodiment provides a feeding control system for roller parts, which is suitable for the directional feeding of axisymmetric roller parts in automated assembly lines, and is particularly suitable for intelligent manufacturing production lines with high requirements for feeding accuracy, adaptability and stability.

[0103] I. System Structure

[0104] like Figure 1 and Figure 2 As shown, the feeding control system includes the following modules and devices:

[0105] Collaborative control platform 1 integrates the following devices:

[0106] Flexible vibration unit 11: A programmable vibration platform with multi-axis drive. Its bearing surface is made of elastic polymer material (such as polyurethane composite matrix), and multiple roller part positioning grooves are arranged in an array on the surface. The drive source is a dual-channel piezoelectric vibrator, which supports multi-frequency superimposed vibration output to disturb the attitude of the roller parts and achieve self-correction under random posture.

[0107] Visual feedback unit 12: Installed above the flexible vibration unit 11, it uses an industrial camera (such as the Baslerace series) in conjunction with an LED surface light source to acquire high-contrast pose images. Through an image recognition module integrating deep learning algorithms (deploying a YOLOv5 micro-model), it determines whether the roller components in each positioning slot are in a horizontal and stable state.

[0108] Four-axis adaptive transfer device 13: A four-axis robotic arm with a SCARA structure (such as the ABBIRB910SC series) with a negative pressure adsorption component (vacuum suction cup with SMC vacuum generator) integrated at its end. This component is used to accurately adsorb roller parts with qualified posture after receiving the transfer gripping command and transfer them to the subsequent work station.

[0109] Dynamic sensing module 2 includes:

[0110] Material property sensing unit 21: An ultrasonic echo detector embedded in the bottom of the flexible vibration unit 11 is used to obtain material parameters such as the elastic modulus and damping coefficient of the bearing substrate, and to monitor the fatigue degradation trend during the operation cycle.

[0111] Vibration state monitoring unit 22: Real-time acquisition of vibration frequency, triaxial amplitude and modal changes through laser vibrometer and three-dimensional accelerometer, and feedback of vibration response state.

[0112] Environmental coupling sensing unit 23: A temperature and humidity sensor array (such as the SHT35 series) is arranged around the system to monitor the influence of environmental temperature gradient and humidity distribution on material behavior.

[0113] Deformation tracking unit 24: Employs a high-precision laser profile scanner (such as the KEYENCELJ-X8000 series) to acquire the microscopic groove topology profile data of each part's positioning groove in real time, forming a groove health record.

[0114] Intelligent Decision Center 3 includes:

[0115] Multiphysics coupling unit 31: By integrating finite element modules, the elastic deformation model, vibration energy transfer equation and hydrothermal coupling creep function are fused to construct a dynamic positioning interference prediction model;

[0116] Adaptive decision unit 32: Runs edge computing modules (such as NVIDIA Jetson Xavier NX), inputs the perceived data into the prediction model in real time, and outputs the comprehensive interference index of the current positioning slot;

[0117] In-situ repair execution unit 33: When the comprehensive interference index exceeds the threshold, it drives the precision laser remelting device (1064nm wavelength laser) or micro stamping module to perform in-situ repair on the groove-shaped defect area.

[0118] II. Working Principle of Example 1

[0119] The roller parts are randomly distributed onto the bearing surface of the flexible vibration unit 11. By setting parameters, a composite vibration mode is triggered, causing the parts to undergo attitude disturbance in the array positioning slot.

[0120] After each vibration cycle, the visual feedback unit 12 acquires the current image and identifies the posture of the roller parts in each positioning slot. If the detected posture is "horizontal upward", the system generates a transfer and grasping command with the corresponding coordinates.

[0121] The four-axis adaptive transfer device 13 quickly performs positioning and gripping, adsorbs the qualified roller parts, and sends them to the downstream assembly stage.

[0122] Meanwhile, the dynamic sensing module 2 continuously collects data including matrix material parameters, vibration response characteristics, environmental parameters, and trough shape data, and inputs them into the intelligent decision-making center 3.

[0123] The intelligent decision-making center 3 outputs a comprehensive positioning interference index based on the constructed multiphysics model. If the interference index of a positioning slot exceeds the threshold due to micro-deformation, material degradation, or environmental disturbance, a repair command is automatically generated, and the in-situ repair execution unit 33 is driven to perform micro-area repair operations.

[0124] III. Technical Effects of Example 1

[0125] This system uses multi-frequency composite vibration combined with visual recognition to achieve rapid and precise adjustment of the posture of roller parts, effectively improving feeding efficiency;

[0126] The coordinated control of the sensing module and the decision-making center enables the system to have environmental adaptability and fault early warning capability, thus enhancing its robustness;

[0127] The introduction of the in-situ repair unit ensures the long-term stability and reusability accuracy of the positioning structure, extending the equipment lifespan;

[0128] The closed-loop control mechanism realizes an intelligent iterative optimization process of "vibration-detection-decision-repair", which is suitable for automated factory environments with high requirements for production line stability.

[0129] Example 2 is the second embodiment of the present invention. Unlike the previous embodiment, this embodiment further introduces a closed-loop alignment module 5 based on Example 1, and enhances the end effector structure of the four-axis adaptive transfer device 13, so that the system can accurately identify and correct the axial end face positioning notch of the roller part, thereby realizing assembly-level attitude control.

[0130] I. System Structure

[0131] The feeding control system for roller parts includes all the modules in Example 1, with the following additional devices:

[0132] Servo-driven rotary mechanism 4: Located between the end effector of the four-axis adaptive transfer device 13 and the negative pressure adsorption device, it adopts a high-precision direct-drive rotary platform (such as the HarmonicDriveFHA series) to control the adsorbed roller parts to rotate precisely around the Z-axis in order to adjust the orientation of their notches.

[0133] Closed-loop positioning module 5 includes:

[0134] Notch Recognition Unit 51: An industrial camera (resolution ≥ 5MP) and a high-brightness backlight assembly mounted on the lateral fixed bracket of the flexible vibration unit 11. It is used to acquire axial end-face images of the roller parts after they are adsorbed by the negative pressure adsorption device, and to identify the azimuth angle of the notch using a convolutional neural network (such as a simplified version of ResNet18).

[0135] Position correction unit 52: Connects notch recognition unit 51 and servo drive rotation mechanism 4, used to receive the deviation value between notch azimuth angle and preset reference angle, and control rotation mechanism to perform compensatory rotation adjustment.

[0136] Linkage control unit 53: Generates a "transfer release command" based on the identified status, and links the four-axis robotic arm and end-effector adsorption device to complete the handling and precise placement actions.

[0137] Pose fusion unit: connects visual feedback unit 12 and notch recognition unit 51, and is used to fuse two types of angle recognition data to perform pose consistency judgment.

[0138] Vibration-transfer coordination controller: connects the pose fusion unit with the flexible vibration unit 11 and the four-axis adaptive transfer device 13, and is used to perform parameter adjustment and rhythm coordination.

[0139] II. Working Principle of Example 2

[0140] The roller parts are randomly distributed on the flexible vibration unit 11 and enter the array positioning slot under the action of multi-frequency composite vibration. The visual feedback unit 12 identifies their pose and extracts the tilt angle θ of the parts. vib ;

[0141] Once the initial posture determination conditions are met (such as the upper surface of the part being approximately horizontal), the negative pressure adsorption device adsorbs the target roller part.

[0142] The notch recognition unit 51 acquires an axial end face image of the roller part, identifies the notch azimuth angle, and converts it into the part tilt angle θ. gap ;

[0143] Pose fusion unit compares component tilt angle θ vib with the tilt angle θ of the part gapIf the difference exceeds 5°, it is considered that the attitude fusion error is significant, triggering the vibration parameter readjustment command.

[0144] The vibration-transport co-controller responds to this command and calculates the frequency adjustment amount:

[0145] Δf adj =0.1×|θ vib -θ gap |, Transit delay time Δt delay =0.2×|θ vib -θ gap | Then proceed with the next optimization cycle;

[0146] If the posture error is within the allowable range, the posture correction unit controls the servo drive rotation mechanism to perform precise angle compensation so that the notch faces the preset assembly requirements.

[0147] After calibration, the linkage control unit 53 generates a transfer release command, and the four-axis adaptive transfer device 13 moves the part to the target tray and accurately releases it into the designated slot.

[0148] III. Technical Effects of Example 2

[0149] Enhanced accuracy of roller component loading posture control: By identifying end face positioning gaps and implementing active angle correction, the control of the "rotational degree of freedom" of roller components is supplemented, meeting the requirements of precision assembly positioning.

[0150] Enhance the intelligence and fault tolerance of the feeding system: The pose fusion mechanism enables the system to automatically determine the contradiction between visual recognition and gap recognition, trigger adaptive vibration optimization, and effectively avoid misidentification and missed detection.

[0151] Supports application requirements for complex assembly scenarios: This embodiment forms a collaborative closed-loop control process in trajectory planning, rotation control, image recognition, etc., which is particularly suitable for high-precision feeding scenarios that require angular alignment, such as motor shaft cores and needle roller bearings.

[0152] Maintaining the system's adaptive optimization capabilities and stability: Through a real-time fine-tuning mechanism for frequency and rhythm, the material feeding process can evolve on its own and maintain operational consistency, effectively extending the system's operating cycle and reducing maintenance frequency.

[0153] Example 3 is the third embodiment of the present invention. Unlike the previous embodiment, based on Example 1 and Example 2, this embodiment further introduces dynamic energy regulation of the in-situ repair mechanism, environmental adaptive threshold judgment mechanism, and roller component density-adaptive vibration strategy to improve the robustness, stability and service life of the system under various working conditions and component differences.

[0154] I. Supplementary Explanation of System Structure

[0155] Example 3 adds and expands the following content based on the existing structure:

[0156] In-situ repair execution unit 33 includes:

[0157] Laser energy control subunit 331: Built-in tunable fiber laser (such as 1064nm IPGYLP series), configured with laser power lookup table mapping logic and exponential dynamic power control module, used to automatically adjust laser output according to deformation depth and material correction coefficient;

[0158] Micro-stamping compensation subunit 332: It uses a micro servo punch (such as the PIN-310 series) and a three-axis force sensor to dynamically compensate for the punching force and direction based on the vibration frequency and three-dimensional amplitude vector feedback.

[0159] The adaptive decision unit 32 is extended as follows:

[0160] Threshold evolution subunit 321: Performs multi-parameter coupled calculations based on ambient temperature change ΔT, relative humidity deviation ΔRH, and current vibration frequency f. vib Dynamically generate judgment threshold T dyn ;

[0161] Repair priority decision subunit 322: Based on the comprehensive positioning interference index δ and the dynamic threshold T dyn The relationship between them automatically generates a "repair emergency" or "standby repair" decision result.

[0162] Density Adaptive Vibration Module 6 New Features:

[0163] Part density detection unit 61: Estimates the roller part density ρ based on the grayscale distribution of the image acquired by visual feedback unit 12. part ;

[0164] Spectrum optimization unit 62: Based on the roller component density ρ part Matching the optimal vibration spectrum set [f min f opt f max ];f min f opt f max These represent the minimum permissible frequency, the optimal operating frequency, and the maximum safe frequency, respectively.

[0165] Amplitude compensation unit 63: Receives the current amplitude vector in real time. Compared with theoretical value Comparison, drive voltage correction is performed: drive voltage ΔV = β·(ρ) part -ρ0).

[0166] II. Working Principle of Example 3

[0167] Once the roller parts are adjusted on the flexible vibration unit 11 to a position that basically meets the gripping conditions, they are transferred to the assembly area according to the aforementioned process.

[0168] Dynamic sensing module 2 collects deformation depth h in real time d The material fatigue degradation coefficient λ is transferred to the laser energy control subunit, and a reference laser power reference value P is generated according to the mapping relationship. base And according to the exponential relationship The output repair laser pulse, where κ represents the material correction coefficient;

[0169] If the current vibration frequency f vib When the frequency is greater than 50Hz, the micro-impact compensation subunit 332 activates the impact force compensation, according to formula F. comp =F0[1+0.02(f vib Adjust the vertical impact force and combine it with the three-dimensional amplitude vector. Decompose the transverse component and apply a reverse compensating force. To stabilize the stamping direction; where F0 represents the reference stamping force, i.e., f vib = Impulse force at 50Hz, F comp This represents the compensated impact force, and K represents the stiffness coefficient. This represents the lateral vibration component.

[0170] Threshold evolution subunit 321 calculates the current judgment threshold in real time:

[0171] T dyn =T0+α T ·ΔT+α H ·|ΔRH-40%|+α f ·log 10 (f vib / 10), and compared with the comprehensive interference index δ; T dyn The dynamic threshold is represented by T0, the baseline threshold is represented by ΔT, the temperature change is represented by ΔRH, and the relative humidity deviation is represented by α. T α H α f These represent the preset first configuration coefficient, second configuration coefficient, and third configuration coefficient, respectively;

[0172] If δ > 1.2T dyn Generate an emergency repair command to immediately activate the repair process;

[0173] If T dyn <δ≤1.2T dyn If so, a standby repair command is generated and added to the repair queue;

[0174] Simultaneously, the system acquires part images from the visual feedback unit 12 and calculates the part density ρ by comparing the image grayscale distribution. part The signal is transmitted to the spectrum optimization unit 62 to select the optimal vibration frequency f. opt ;

[0175] Amplitude compensation unit 63 calculates the difference between the current amplitude and the theoretical amplitude. At that time, voltage correction ΔV = β·(ρ) is performed. part -ρ0), thereby achieving dynamic adaptive vibration compensation for parts with different densities; ρ represents the theoretical amplitude, β represents the density-voltage coefficient, ρ0 represents the density of the reference part, and ΔV represents the amplitude deviation modulus.

[0176] III. Technical Effects of Example 3

[0177] Achieving high-precision groove repair control: Based on fatigue parameters and topology data, a laser power control logic is constructed to make the laser repair response more precise and avoid overheating or insufficient power;

[0178] Enhanced stamping stability and directional control: Stamping compensation is achieved through dual-dimensional input of vibration frequency and amplitude, making it particularly suitable for fine correction in high-speed operation scenarios;

[0179] An environment-adaptive judgment mechanism is introduced: dynamic threshold generation and repair priority judgment mechanism improve system fault tolerance and prevent erroneous or missed repairs;

[0180] Compatible with multi-density parts processing requirements: For roller parts of different materials or structures, adaptively match vibration spectrum and drive compensation parameters to improve the versatility and adaptability of the whole machine;

[0181] Enhanced system stability and long-term maintainability: This embodiment strengthens the response capability to "structural fatigue and environmental disturbances during long-term use", effectively extending the system maintenance cycle and ensuring continuous operation of the production line.

[0182] Example 4 is the fourth embodiment of the present invention. Unlike the previous embodiment, based on Examples 1 to 3, this embodiment further introduces a vibration-visual phase synchronization mechanism, an environmental coupling creep and fatigue modeling mechanism, and a groove level classification and repair energy optimization strategy, thereby enhancing the system's response accuracy, identification stability, and repair efficiency of roller parts under dynamic working conditions.

[0183] I. Supplementary Explanation of System Structure

[0184] This embodiment expands the existing modules with the following functional units:

[0185] Phase synchronization controller: integrated between visual feedback unit 12 and flexible vibration unit 11, connected to real-time vibration driver, and obtains the current vibration phase angle φ∈[0,2π] through high-precision encoder;

[0186] Pose prediction compensator: Deployed in the embedded computing core, it connects the phase synchronization controller and the amplitude control unit to calculate the displacement vector of the roller parts between continuous vibration cycles and correct the amplitude of the next cycle;

[0187] Two new modeling components have been added to Multiphysics Coupled Unit 31:

[0188] The environmental creep function construction module is used to calculate the creep strain caused by temperature and time. c ;

[0189] The dynamic damage modeling module is used to construct the fatigue damage factor D. d ;

[0190] In-situ repair execution unit 33 is expanded as follows:

[0191] Trough-shaped classification subunit 333, based on deformation depth h d They are classified into three categories: slight deformation, moderate deformation, and severe deformation.

[0192] Energy optimization subunit 334, based on the classification results, performs total energy calculation and control for the corresponding repair strategy. II. Working Principle of Example 4

[0193] Phase-visual synchronization feedback mechanism:

[0194] The flexible vibration unit 11 outputs an excitation signal at a composite frequency, and simultaneously records the current vibration phase φ (φ∈[0,2π]);

[0195] The phase synchronization controller sets the trigger phase conditions: (k is an integer) When the image is acquired at this phase, the displacement of the part is instantaneously still, and the image stability is optimal.

[0196] The pose prediction compensator compares two adjacent period images and calculates the planar displacement vector of the roller component on the bearing surface.

[0197] According to the amplitude correction formula Adjusting the three-dimensional amplitude command for the next cycle enables prediction and adaptive compensation for dynamic disturbances; among which... denoted as the corrected amplitude vector, and μ represents the compensation coefficient.

[0198] Multiphysics coupling modeling and localization interference prediction:

[0199] Multiphysics coupling unit introduces environmental creep function:

[0200]

[0201] Where, ∈ c σ represents creep strain, E0 represents material stress, Q represents the reference elastic modulus, R represents the activation energy, T represents the ambient temperature, t represents time, and n represents the stress exponent.

[0202] Simultaneous calculation of humidity-driven fatigue damage factors:

[0203] D d =1-(1-λ) 50ΔRH ;

[0204] Among them, D d The fatigue damage factor is represented by ΔRH, which represents the relative humidity deviation value, calculated based on the humidity distribution.

[0205] creep strain ∈ c and fatigue damage factor D d Input the location interference prediction model and output the interference index δ to guide the intelligent decision-making center in determining the repair priority.

[0206] Deformation level classification and repair energy optimization mechanism:

[0207] The deformation depth h obtained by the trough-shaped classification subunit based on the trough topology data d Classification:

[0208] Microdeformation (h) d ≤0.1mm) → Activate laser thermal compensation mode;

[0209] Moderate deformation (0.1mm < h) d ≤0.3mm) → Activate micro-stamping mode;

[0210] Severe deformation (h) d >0.3mm) → Activate laser-stamping composite mode;

[0211] Energy optimization subunit 334 calculates the total repair energy W based on the classification results. total The following piecewise function is used:

[0212]

[0213] Among them, W total The total repair energy is represented by k1, k2, and c, which represent the preset first mode coefficient, second mode coefficient, and third mode coefficient, respectively. α and β represent the first weighting factor and the second weighting factor, respectively.

[0214] The parameters for each repair method (such as laser power and impact force) are determined by the upper-level scheduling controller based on W. total Reverse adjustment ensures optimal repair efficiency with minimal energy consumption.

[0215] III. Technical Effects of Example 4

[0216] Enhance visual recognition stability and pose compensation accuracy: Based on phase synchronization control and pose prediction compensation mechanism, the system can achieve accurate image acquisition and amplitude self-adjustment during dynamic vibration, avoiding false detection and false capture;

[0217] Achieving coupled modeling and prediction of multiple environmental factors: For the first time, environmental creep function and humidity fatigue model were introduced into the roller feeding system, and a three-dimensional disturbance evolution path of material-time-environment was established to improve prediction accuracy;

[0218] Improve energy utilization and repair strategy adaptability: Intelligently match repair strategies according to deformation depth, execute differentiated energy allocation, effectively reduce energy consumption, improve repair efficiency, and prevent over-processing or under-repair.

[0219] Enhancing the long-term stability and adaptability of the system: Example 4 provides the system with higher-dimensional perception and response capabilities, and has the ability to self-evolve and self-repair in the face of complex and ever-changing production environments.

[0220] A method for controlling the feeding of roller parts, applied to the aforementioned feeding control system for roller parts, with reference to... Figure 3 ,include:

[0221] Step S1: The flexible vibration unit 11 adjusts the position and orientation of the roller parts through multi-frequency composite vibration; the visual feedback unit 12 collects the position and orientation images of the roller parts in the positioning groove of each part after each vibration cycle, and identifies the position and orientation images to form a position and orientation recognition result. When the position and orientation recognition result shows that the surface of the roller parts is horizontal with the positioning groove of the parts, a transfer and gripping command is generated.

[0222] Step S2: The four-axis adaptive transfer device 13 drives the negative pressure adsorption device to move above the corresponding roller parts according to the transfer gripping command, so as to adsorb the roller parts.

[0223] Step S3: Material property sensing unit 21 acquires the matrix material parameters of the flexible vibration unit 11 bearing substrate in real time; vibration state monitoring unit 22 detects vibration dynamic parameters; environmental coupling sensing unit 23 monitors environmental parameters in real time; and deformation tracking unit 24 acquires the groove topology data of each part positioning groove in real time through laser micron scanning.

[0224] In step S4, the multi-physics coupling unit 31 constructs a positioning interference prediction model by integrating elastic deformation theory, vibration energy transfer equation and environmental creep function; the adaptive decision unit 32 inputs the real-time collected matrix material parameters, vibration dynamic parameters, environmental parameters and trough topology data into the positioning interference prediction model and outputs a comprehensive positioning interference index; when the comprehensive positioning interference index is greater than the dynamic threshold, a trough repair instruction set is generated.

[0225] Step S5: The in-situ repair execution unit 33 performs laser recladding or micro-stamping shaping on the target positioning groove according to the repair instruction set.

[0226] In step S6, the visual feedback unit 12, the dynamic perception module 2, and the intelligent decision-making center 3 form a closed-loop control circuit, driving the flexible vibration unit 11 to execute the iterative optimization process of "vibration-detection-decision-repair".

[0227] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A feeding control system for roller parts, characterized in that, include: The collaborative regulation platform (1) integrates: The flexible vibration unit (11) has an array of part positioning grooves on its bearing surface. The flexible vibration unit (11) is configured to adjust the position of the roller part through multi-frequency composite vibration. The visual feedback unit (12) is suspended above the flexible vibration unit (11) and is configured to acquire the position and pose images of the roller parts in the positioning groove of each part after each vibration cycle, and to identify the position and pose images to form a position and pose recognition result. When the position and pose recognition result shows that the surface of the roller part is horizontal with the positioning groove of the part, a transfer and gripping command is generated. The four-axis adaptive transfer device (13) is located on the side of the flexible vibration unit (11). Its end effector integrates a negative pressure adsorption device. The four-axis adaptive transfer device (13) drives the negative pressure adsorption device to move above the corresponding roller part according to the transfer gripping command, so as to adsorb the roller part. The dynamic sensing module (2), connected to the collaborative control platform (1), includes: The material property sensing unit (21) is used to acquire the matrix material parameters of the flexible vibration unit (11) bearing substrate in real time. The matrix material parameters include elastic modulus, damping coefficient and fatigue life degradation coefficient. The vibration state monitoring unit (22) is used to detect vibration dynamic parameters in real time, including vibration frequency and three-dimensional amplitude vector. An environmental coupling sensing unit (23) is used to monitor environmental parameters in real time, including environmental temperature and humidity distribution. The deformation tracking unit (24) is used to acquire the groove topology data of the positioning groove of each part in real time by laser micron scanning; The intelligent decision-making center (3), connected to the dynamic sensing module (2), includes: Multi-physics coupling unit (31) is used to construct a positioning interference prediction model by integrating elastic deformation theory, vibration energy transfer equation and environmental creep function; An adaptive decision unit (32) is connected to the multi-physics coupling unit (31) and is used to input the real-time collected matrix material parameters, vibration dynamic parameters, environmental parameters and trough topology data into the positioning interference prediction model and output a comprehensive positioning interference index; when the comprehensive positioning interference index is greater than the dynamic threshold, a trough repair instruction set is generated. The in-situ repair execution unit (33) is connected to the adaptive decision unit (32) and is used to perform laser recladding or micro-stamping shaping on the target positioning groove according to the groove repair instruction set; The visual feedback unit (12), the dynamic perception module (2), and the intelligent decision-making center (3) form a closed-loop control circuit, driving the flexible vibration unit (11) to perform an iterative optimization process of "vibration-detection-decision-repair".

2. The feeding control system for roller parts according to claim 1, characterized in that: The end effector of the four-axis adaptive transfer device (13) is provided with a servo-driven rotation mechanism (4) between the negative pressure adsorption device and the end effector of the roller part. The axial end face of the roller part is provided with a positioning notch. The feeding control system further includes a closed-loop alignment module (5), connected to the collaborative control platform (1), comprising: The notch identification unit (51) is disposed on a fixed bracket on the side of the flexible vibration unit (11) and configured as follows: After the negative pressure adsorption device adsorbs the roller part, an axial end face image containing the positioning notch is acquired; The notch azimuth angle in the axial end face image is identified based on a convolutional neural network. A rotation compensation command is generated when the azimuth angle of the notch deviates from the preset reference angle. A transfer release command is generated when the azimuth angle of the notch matches a preset reference angle. The pose correction unit (52), connected to the notch recognition unit (51) and the servo-driven rotation mechanism (4), is configured as follows: Analyze the angle deviation value in the rotation compensation command; The servo-driven rotation mechanism (4) controls the roller component to rotate around the Z-axis by the angle deviation value; The linkage control unit (53), connected to the notch recognition unit (51), is configured as follows: In response to the transfer release command, the transfer trajectory of the negative pressure adsorption device to the next workstation tray is planned; The negative pressure release signal is triggered when the target placement slot is directly above the target placement slot.

3. The feeding control system for roller parts according to claim 1, characterized in that: The in-situ repair execution unit (33) includes: The laser energy modulation subunit (331) is configured as follows: The target laser power reference value P is determined by consulting the laser power mapping table based on the fatigue life degradation coefficient λ. base ; Deformation depth h calculated based on groove topology data d According to the dynamic power formula Adjust the output power, where K represents the material correction factor; The micro-stamping compensation subunit (332) is configured as follows: Real-time acquisition of vibration frequency f vib When f vib Activate impulse compensation at >50Hz: F comp =F0[1+0 . 02(f vib -50)], where F0 represents the reference punching force, i.e. f vib = Impulse force at 50Hz, F comp Indicates the impact force after compensation; Based on the three-dimensional amplitude vector The transverse vibration component is obtained by decomposition, and a reverse compensating force is applied to the stamping head. Where K represents the stiffness coefficient. The transverse vibration component o is represented.

4. The feeding control system for roller parts according to claim 3, characterized in that: The adaptive decision-making unit (32) includes: Threshold evolution subunit (321) is used to perform dynamic threshold calculation: T dyn =T0+α T ·ΔT+αH·|ΔRH-40%|+α f ·log 10 (f vib / 10); Among them, T dyn The dynamic threshold is represented by T0, the baseline threshold is represented by ΔT, the temperature change is represented by ΔRH, and the relative humidity deviation is represented by α. T α H α f These represent the preset first configuration coefficient, second configuration coefficient, and third configuration coefficient, respectively; Repair the priority decision subunit (322) and configure it as follows: When the comprehensive positioning interference index δ > 1.2T dyn Emergency repair instructions are generated in a timely manner; When T dyn <δ≤1.2T dyn A standby repair command is generated in time.

5. The feeding control system for roller parts according to claim 1, characterized in that: It also includes a density-adaptive vibration module (6), connected to the collaborative control platform (1), comprising: The part density detection unit (61) calculates the roller part density ρpart through the pixel grayscale distribution of the pose image; The spectrum optimization unit (62), connected to the part density detection unit (61), is configured as follows: According to the density ρ of the roller parts part Matching vibration spectrum group [f min f opt f maax ], where f min f opt f max These represent the minimum permissible frequency, the optimal operating frequency, and the maximum safe frequency, respectively. An amplitude compensation unit (63) is connected to the spectrum optimization unit (62) and is used to receive the three-dimensional amplitude vector in real time. And execute: when At that time, according to ΔV=β·(ρ part -ρ0) Corrects the driving voltage; in ρ represents the theoretical amplitude, β represents the density-voltage coefficient, ρ0 represents the density of the reference part, and ΔV represents the amplitude deviation modulus.

6. The feeding control system for roller parts according to claim 2, characterized in that: The closed-loop alignment module (5) and the visual feedback unit (12) interact through the pose fusion unit: The tilt angle θ of the part receiving the pose recognition result vib ; The component tilt angle θ is calculated from the azimuth angle of the receiving notch. gap ; When |θ vib -θ gap |When the angle is greater than 5°, a vibration parameter readjustment command is triggered; The vibration-transfer co-controller responds to the vibration parameter readjustment command by executing: Calculate the frequency adjustment amount Δf adj =0.1×|θ vib -θ gap |; Calculate the transit delay time Δt delay =0.2×|θ vib -θ gap | 7. The feeding control system for roller parts according to claim 3, characterized in that: The visual feedback unit and the flexible vibration unit constitute a real-time feedback link, including: The phase synchronization controller is configured as follows: Let the current vibration phase angle be φ (φ∈[0, 2π]), in Image acquisition is triggered when (k is an integer); The pose prediction compensator, connected to the phase synchronization controller, performs the following: Calculate the displacement vector of the part based on the pose images of adjacent vibration cycles. According to the amplitude compensation formula Correct the amplitude of the next cycle, where denoted as the corrected amplitude vector, and μ represents the compensation coefficient.

8. The feeding control system for roller parts according to claim 3, characterized in that: The multiphysics coupling unit (31) performs: Environmental creep function construction: Where, ∈ c σ represents creep strain, E0 represents material stress, Q represents the reference elastic modulus, R represents the activation energy, T represents the ambient temperature, t represents time, and n represents the stress exponent. Constructing a dynamic damage model: D d =1-(1-λ) 50ΔRH ; Among them, D d ΔRH represents the fatigue damage factor and the relative humidity deviation value. The creep strain ∈ c and fatigue damage factor D d The comprehensive positioning interference index δ is generated by inputting the positioning interference prediction model.

9. The feeding control system for roller parts according to claim 3, characterized in that: The in-situ repair execution unit (33) includes: The trough-shaped classification subunit (333) classifies the deformation into: based on the trough-shaped topology data. Micro-deformation (h) d ≤0.1mm) → Activate laser thermal compensation mode; Moderate deformation (0.1mm < h) d ≤0.3mm) → Activate micro-stamping mode; Severe deformation (h) d >0.3mm) → Activate laser-stamping composite mode; The energy optimization subunit (334) is connected to the slot-shaped classification subunit (333) and performs energy repair: Among them, W total The total repair energy is represented by k1, k2, and c, which represent the preset first mode coefficient, second mode coefficient, and third mode coefficient, respectively. α and β represent the first weighting factor and the second weighting factor, respectively.

10. A method for controlling the feeding of roller parts, applied to the feeding control system for roller parts as described in any one of claims 1-9, characterized in that, include: Step S1, the flexible vibration unit (11) realizes the position and posture adjustment of the roller parts through multi-frequency composite vibration; the visual feedback unit (12) collects the position and posture images of the roller parts in the positioning groove of each part after each vibration cycle, and identifies the position and posture images to form a position and posture recognition result. When the position and posture recognition result shows that the surface of the roller parts is horizontal with the positioning groove of the part, a transfer and gripping command is generated. Step S2, the four-axis adaptive transfer device (13) drives the negative pressure adsorption device to move above the corresponding roller part according to the transfer gripping command, so as to adsorb the roller part; Step S3: The material property sensing unit (21) acquires the matrix material parameters of the flexible vibration unit (11) bearing substrate in real time; the vibration state monitoring unit (22) detects the vibration dynamic parameters; the environmental coupling sensing unit (23) monitors the environmental parameters in real time; and the deformation tracking unit (24) acquires the groove topology data of each part positioning groove in real time through laser micron scanning. Step S4, the multi-physics coupling unit (31) constructs a positioning interference prediction model by integrating elastic deformation theory, vibration energy transfer equation and environmental creep function; The adaptive decision unit (32) inputs the real-time collected matrix material parameters, vibration dynamic parameters, environmental parameters and trough topology data into the positioning interference prediction model and outputs a comprehensive positioning interference index; when the comprehensive positioning interference index is greater than the dynamic threshold, a trough repair instruction set is generated. Step S5, the in-situ repair execution unit (33) performs laser recladding or micro-stamping shaping on the target positioning groove according to the groove repair instruction set; In step S6, the visual feedback unit (12), the dynamic perception module (2), and the intelligent decision-making center (3) form a closed-loop control circuit, driving the flexible vibration unit (11) to execute the iterative optimization process of "vibration-detection-decision-repair".

Images