Controllable automatic clamping hydraulic chucking device and control method thereof

Abstract

The application discloses a controllable automatic clamping hydraulic clamping device and a control method thereof, and belongs to the technical field of hydraulic clamping. The device comprises a clamping mechanism, a supporting mechanism, a driving mechanism, a pressing mechanism and a camera mechanism. The clamping mechanism clamps the workpiece through the cooperation of the clamping jaw A and the clamping jaw B; the driving mechanism adopts a servo motor to drive a lead screw, drives a movable seat to realize horizontal feeding, and realizes accurate positioning through an internal torque sensor; the pressing mechanism corrects the flatness of the surface of the workpiece through a cylinder-driven pressing plate; and the camera mechanism is used for collecting the image of the workpiece. The control method dynamically adjusts the pressing feeding amount by using a fuzzy PID controller through image preprocessing, feature point extraction and position deviation calculation, and combines historical data to perform adaptive optimization. The application realizes full automation and high-precision control of workpiece clamping, and effectively improves the machining quality and production efficiency.

Description

Technical Field

[0001] This invention relates to the field of hydraulic clamping technology, and in particular to a controllable automatic clamping hydraulic clamping device and its control method. Background Technology

[0002] In the parts processing, manufacturing, and assembly industries, clamping devices are widely used to ensure the stability of workpieces during machining, especially in high-precision machining processes such as surface grinders, CNC machine tools, and automated assembly lines, where clamping accuracy directly affects machining quality and efficiency. Traditional hydraulic clamping systems mostly rely on mechanical transmission (such as lead screw rotation) to achieve clamping, lacking precise control over workpiece positioning and pressing processes. This makes it difficult to accurately adjust the clamping force, easily leading to problems such as excessive or insufficient clamping force, which in turn causes workpiece surface damage, unacceptable flatness, or repeated clamping, affecting machining accuracy and increasing production costs.

[0003] In addition, existing equipment is usually less automated and cannot meet the requirements of high-speed, high-precision modern production lines, especially in terms of adaptability to workpieces with different structures (such as ring parts with large roundness deviations).

[0004] Therefore, there is an urgent need to develop an intelligent hydraulic clamping device with precise stroke control, adaptive clamping, and real-time feedback capabilities to improve clamping accuracy, efficiency, and product quality consistency. Summary of the Invention

[0005] To address the above problems, this invention provides a controllable automatic clamping hydraulic clamping device and its control method, aiming to solve the problems of low positioning accuracy, inaccurate clamping force control, and insufficient automation in existing clamping devices. By integrating visual inspection, intelligent control, and precision actuators, it achieves high-precision automatic positioning, adaptive clamping, and intelligent flatness correction of workpieces, effectively improving processing quality and production efficiency.

[0006] In a first aspect, the present invention provides a controllable automatic clamping hydraulic clamping device, comprising: The clamping mechanism includes a fixed-side clamp A and a movable-side clamp B. The lower bottom surface of clamp A is bolted to the upper surface of clamp V-block A and installed on one side of the hydraulic clamping device housing; the lower bottom surface of clamp B is bolted to the upper surface of clamp V-block B and installed on the side of the movable seat of the clamping device. The support mechanism includes a positioning base mounted on the worktable of the hydraulic clamping device housing, and a support base connected above the positioning base. The drive mechanism includes a lead screw installed at the lower end of the movable seat of the clamping device. The rotation of the lead screw drives the lateral feed of the jaw B through the movable seat of the clamping device. The pressing mechanism includes a column fixed to one side of the zero point position of the worktable of the surface grinder, a metal shell connected to the column by a fixing ring, a telescopic plate disposed on the slide rail on the bottom surface of the metal shell, a pneumatic motor that drives the telescopic plate, a cylinder installed at the end of the telescopic plate, and a pressing plate driven by the cylinder to perform the pressing action. The camera setup includes an industrial camera mounted on the column.

[0007] Furthermore, the drive mechanism also includes a servo motor and a geared motor. The output end of the servo motor is connected to the lead screw via a synchronous pulley and a synchronous shaft. The servo motor is equipped with a torque sensor.

[0008] By introducing servo motors, synchronous wheel shaft drives, and built-in torque sensors, precise control and intelligent feedback of the driving process are achieved. It can accurately control the feed position of the moving side and determine in real time whether the workpiece is in contact with the target area through torque sensing, thereby avoiding workpiece surface damage or deformation caused by excessive pre-clamping force and significantly improving clamping accuracy and safety.

[0009] Furthermore, the servo motor and the geared motor are mounted and fixed by a motor mounting plate and a motor fixing plate. The servo motor and the geared motor are provided with a clamping device protective cover. A transparent PVC plate and a handle cover are installed on the side of the clamping device protective cover.

[0010] By incorporating a motor mounting plate, protective cover, transparent PVC panel, and handle cover, a stable mounting foundation is provided for the motor, ensuring transmission stability. The protective cover effectively isolates coolant and chips, extending the lifespan of the motor and sensors. The transparent PVC panel facilitates observation of the internal operating status, while the handle cover greatly simplifies daily maintenance and repair, enhancing the equipment's usability and reliability.

[0011] Furthermore, the gripper V-block A and gripper V-block B are inverted trapezoidal in shape; the support base is concave and has a drainage hole at the bottom; the bottom of the hydraulic clamping device housing is hollow and has drainage slots on both sides.

[0012] By employing special inverted trapezoidal and concave designs with drainage holes and perforated drainage grooves on the V-block, support base, and device housing, the self-cleaning and rust-prevention capabilities of the device are significantly improved. Cutting fluid and debris are difficult to accumulate and can be quickly discharged, keeping the clamping reference surface clean, reducing maintenance frequency, and fundamentally reducing the risk of internal components rusting due to fluid accumulation, thus ensuring long-term clamping accuracy.

[0013] Furthermore, the drive mechanism also includes a bearing housing and a bearing housing plate for supporting the lead screw.

[0014] By clearly defining the bearing housing and bearing housing plate for supporting the lead screw, a stable and reliable rotational support is provided for the lead screw, effectively suppressing radial runout and axial movement of the lead screw during transmission. This ensures that the movable seat of the clamping device can achieve smooth and precise linear feed, thereby improving the repeatability and long-term operational stability of the entire clamping process.

[0015] Secondly, the present invention also provides a control method for a controllable automatic clamping hydraulic clamping device, comprising the following steps: S1, acquires images of the workpiece and support base through an industrial camera, and preprocesses the images; S2, extract the first feature point set of the lower surface of the workpiece and the second feature point set of the upper surface of the support base from the preprocessed image; S3, calculate the positional deviation between the workpiece and the support base based on the first feature point set and the second feature point set, including distance deviation and angle deviation; S4, the fuzzy PID controller dynamically adjusts the PID parameters according to the magnitude and trend of the distance deviation and angle deviation, and outputs the feed amount of the pressing mechanism; S5: Control the pressing mechanism to perform the pressing action according to the feed amount, and continuously acquire images during the pressing process and repeat S1 to S4.

[0016] Furthermore, the method also includes: recording the feature point data and feed rate of the current workpiece, comparing it with the data of the previous workpiece, and dynamically adjusting the parameters of the fuzzy PID controller and the feed rate based on the comparison result.

[0017] The above solution provides an intelligent control method for the device, the core of which lies in realizing intelligent and adaptive optimization of the clamping process. Through visual inspection, fuzzy PID control, and data iterative learning, the system can actively sense and compensate for the position and posture deviations of the workpiece, and dynamically adjust the pressing amount. This not only significantly improves the flatness correction capability for different workpieces, but also enables the system to have a self-learning function, becoming more and more accurate with use, thus greatly improving production quality and automation level.

[0018] Furthermore, the image preprocessing specifically includes: An adaptive median filtering algorithm is used to analyze the image pixel by pixel to remove noise while preserving edge details; Dynamic histogram equalization is used to enhance image contrast. Image correction algorithms are used to correct distortion and adjust the viewing angle of images that are not taken from the front.

[0019] By employing adaptive median filtering, dynamic histogram equalization, and image correction techniques, clear, accurate, and high-contrast images can still be obtained even in complex industrial environments (such as uneven lighting, noise, and tilted shooting angles). This provides a high-quality data foundation for subsequent feature extraction and bias calculation, ensuring the robustness and reliability of the vision system.

[0020] Furthermore, S2 specifically includes: The adaptive Canny edge detection algorithm is used to obtain the edge information of the image, and the contour detection algorithm and deep learning model are used to extract the contours of the workpiece and the support base. A feature point recognition algorithm based on the Transformer architecture is used to detect key feature points in an image, and these feature points are compared with preset model feature points to locate and identify the workpiece.

[0021] By combining the efficiency of traditional edge detection with the powerful recognition capabilities of deep learning models, especially by introducing the Transformer architecture, the system can more accurately identify and locate workpiece feature points under complex or occluded conditions. This significantly improves the system's success rate in recognizing diverse workpieces and its positioning accuracy, enhancing the device's versatility.

[0022] Furthermore, the deep learning model is divided into a general model and a special model based on the structure of the workpiece, and is trained in the following way: General Model: The initial deep learning model is pre-trained using a general dataset labeled with feature points of standard circular workpieces; Special model: The parameters of the pre-trained model are transferred to the new model, and the new model is fine-tuned using a special dataset labeled with feature points of non-standard circular workpieces.

[0023] By designing a transfer learning training strategy that combines general and specific models, the system ensures both rapid and accurate processing of common standard workpieces (such as circular structures) and the ability to quickly adapt to new, non-standard shaped workpieces using a small number of samples. This training method reduces the need for a large amount of labeled data for specific workpieces, shortens the model deployment and optimization cycle, and enables the device to flexibly adapt to changing production tasks.

[0024] Compared with the prior art, the beneficial effects of the present invention are: 1. By integrating a high-resolution industrial camera and a precision actuator, real-time visual detection and closed-loop control of the workpiece position are achieved. Structurally, the design adopts an inverted trapezoidal V-block, a support base with drainage holes, and a hollow shell to effectively discharge cutting fluid and chips, while also providing the advantages of rust prevention and easy maintenance. The servo motor has a built-in torque sensor, which, combined with the "pre-clamping - retraction - hydraulic clamping" process, significantly improves the reliability and repeatability of the clamping process.

[0025] 2. Employing a fuzzy PID control algorithm, the system dynamically adjusts the downward feed amount based on real-time position deviation, achieving adaptive control of clamping force and flatness. A transfer learning mechanism is introduced, combining general and specific models to enable the system to quickly adapt to workpieces with different structures, enhancing its generalization ability. By recording historical data and comparing it with the current state, the system continuously optimizes control parameters and recognition models, possessing online learning and continuous evolution capabilities. The entire clamping process, from positioning and clamping to calibration, is fully automated, significantly improving production efficiency while ensuring consistent accuracy. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this drawing 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 drawing. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram of the hydraulic clamping device described in this invention; Figure 2 This is a schematic diagram of the pressing mechanism described in this invention; Figure 3 This is a flowchart of the method of the present invention.

[0028] In the diagram: 1. Gripper A; 2. Gripper B; 3. Gripper V-block A; 4. Gripper V-block B; 5. Support base; 6. Positioning base; 7. Lead screw; 8. Hydraulic clamping device housing; 9. Servo motor and geared motor; 10. Bearing housing and bearing housing plate; 11. Motor mounting plate; 12. Motor fixing plate; 13. Synchronous pulley; 14. Synchronous shaft; 15. Handle cover plate; 16. Clamping device protective cover; 17. Transparent PVC plate; 18. Clamping device movable seat; 19. Telescopic plate; 20. Cylinder; 21. Lower pressure plate; 22. Bottom slide rail; 23. Metal shell; 24. Fixing ring; 25. Column. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments provided by this invention without inventive effort are within the scope of protection of this invention.

[0030] This invention provides a controllable automatic clamping hydraulic clamping device, such as... Figure 1 , Figure 2 As shown, it includes: The clamping mechanism includes a fixed-side clamp A1 and a movable-side clamp B2. The lower bottom surface of the clamp A1 is bolted to the upper surface of the clamp V-block A3 and installed on one side of the hydraulic clamping device housing 8. The lower bottom surface of the clamp B2 is bolted to the upper surface of the clamp V-block B4 and installed on the side of the movable seat 18 of the clamping device. The support mechanism includes a positioning base 6 mounted on the worktable of the hydraulic clamping device housing 8, and a support base 5 connected above the positioning base 6. The drive mechanism includes a lead screw 7 installed at the lower end of the movable seat 18 of the clamping device. The rotation of the lead screw 7 drives the lateral feed of the jaw B2 through the movable seat 18 of the clamping device. The pressing mechanism includes a column 25 fixed to one side of the zero point position of the worktable of the surface grinder, a metal shell 23 connected to the column 25 by a fixing ring 24, a telescopic plate 19 disposed on the slide rail 22 on the bottom surface of the metal shell 23, a pneumatic motor that drives the telescopic plate 19, a cylinder 20 installed at the end of the telescopic plate 19, and a pressing plate 21 driven by the cylinder 20 to perform the pressing action. The camera setup includes an industrial camera mounted on the column 25.

[0031] Furthermore, the drive mechanism also includes a servo motor and a geared motor 9, the output ends of which are connected to the lead screw 7 via a synchronous pulley 13 and a synchronous shaft 14; and a torque sensor is provided inside the servo motor and the geared motor 9.

[0032] Furthermore, the servo motor and the geared motor 9 are mounted and fixed by the motor mounting plate 11 and the motor fixing plate 12. The servo motor and the geared motor 9 are provided with a clamping device protective cover 16. A transparent PVC plate 17 and a handle cover 15 are installed on the side of the clamping device protective cover 16.

[0033] Furthermore, the gripper V-block A3 and gripper V-block B4 are inverted trapezoidal in shape; the support base 5 is concave in design and has a drainage hole at the bottom; the bottom of the hydraulic clamping device housing 8 is hollowed out and has drainage slots on both sides.

[0034] Furthermore, the drive mechanism also includes a bearing housing and a bearing housing plate 10 for supporting the lead screw 7.

[0035] Working Principle: The control system first activates the drive mechanism. Specifically, the servo motor and geared motor 9 begin to rotate, and their power is precisely transmitted to the lead screw 7 via the synchronous belt transmission system composed of the synchronous pulley 13 and the synchronous shaft 14. The lead screw 7 rotates under the stable support of the bearing seats at both ends and the bearing seat plate 10, driving the movable seat 18 of the clamping device to move smoothly towards the fixed side along the straight rail of the worktable. The jaws B2 and jaw V-blocks B4 fixed on the movable seat 18 of the clamping device move forward accordingly, cooperating with the jaws A1 and jaw V-blocks A3 on the fixed side to approach and contact the workpiece from both sides, completing the initial positioning and pre-clamping of the workpiece.

[0036] During the process of the drive mechanism pushing the workpiece, the torque sensor built into the servo motor and geared motor 9 continuously monitors the output torque of the motor. When the workpiece is pushed to the predetermined clamping position, the resistance increases, causing the torque to rise to a preset threshold. The torque sensor immediately feeds this signal back to the control system. The control system then issues a command to briefly reverse the servo motor and geared motor 9, causing the movable seat 18 of the clamping device and the movable side gripper to retract as a whole by a precise distance (e.g., 3mm). The purpose of this retraction action is to actively eliminate the feed error that may be caused by the intrusion of grinding debris or other small particles from the worktable surface, providing a clean and accurate reference position for the subsequent main clamping.

[0037] After the movable side retracts to its position, the control system immediately activates the hydraulic clamping mechanism (hydraulic cylinders and other components not shown in detail in the figure). The hydraulic clamping mechanism drives the movable seat 18 of the clamping device to move forward again, firmly clamping the workpiece onto the support base 5. At this time, the workpiece is locked by high-rigidity hydraulic pressure, which is sufficient to cope with various cutting forces in the subsequent machining process, ensuring the stability of the machining.

[0038] To address the issue of slight warping of the workpiece due to clamping force, which could affect the flatness of its upper surface, this device incorporates a pressing mechanism. After hydraulic clamping is complete, the control system instructs the pressing mechanism to operate: First, a pneumatic motor drives the telescopic plate 19 to extend along the slide rail 22 at the bottom of the metal housing 23, moving the pressing plate 21 directly above the workpiece. Subsequently, the cylinder 20 drives the pressing plate 21 downwards, where the nylon or polyurethane protective plate on its contact surface buffers the impact and applies a uniform vertical downward pressure to the upper surface of the workpiece. While maintaining this downward pressure, the control system controls the hydraulic clamping force to perform a brief "release-re-clamp" cycle. This operation effectively releases the uneven stress inside the workpiece caused by the initial clamping, prompting a fine adjustment of the workpiece's vertical position, thereby correcting its upper surface flatness and ensuring it meets the requirements of high-precision machining. After correction, the pressing plate 21 retracts first, followed by the telescopic plate 19 returning to its initial position, completing the entire clamping process.

[0039] Throughout the process, the industrial camera installed on column 25 can monitor and capture images of the clamping status in real time, providing visual evidence for process quality control and subsequent data analysis.

[0040] This invention also provides a control method for a controllable automatic clamping hydraulic clamping device, such as... Figure 3 As shown, it includes the following steps: S1 acquires images of the workpiece and support base using an industrial camera and preprocesses the images.

[0041] Specifically, a high-resolution industrial camera is installed on the front and rear sides of the pressing mechanism column, and each camera is equipped with a ring LED light source (the wavelength of the light source matches the workpiece material to eliminate surface reflection). At the same time, the height of the camera can be adjusted according to the height of the support base to ensure that the images of the side surface of the workpiece and the support base can be clearly captured, and the frame rate of the camera is sufficient to capture the dynamic changes during the pressing process.

[0042] To improve the recognition accuracy of the vision system, multiple preprocessing steps were performed on the acquired raw images: First, an adaptive median filtering algorithm is used for noise reduction. This algorithm can dynamically adjust the size of the filtering window according to the noise situation in the local area of ​​the image, effectively suppressing noise while preserving the edge details of the workpiece and the support base to the maximum extent.

[0043] Building upon this, dynamic histogram equalization is used to enhance image contrast. This method adaptively determines the optimal pixel value mapping range, avoiding over-enhancement in local areas, thereby making the boundary contours between the lower surface of the workpiece and the upper surface of the support base clearer and sharper.

[0044] To address the issue of industrial cameras being unable to take frontal shots due to installation location limitations, image correction technology was introduced. Through image correction algorithms, image distortion and perspective deformation caused by angles are compensated and adjusted, uniformly correcting images to a standard frontal viewing angle benchmark. This ensures a high degree of consistency and comparability of image features acquired from different batches or at different times.

[0045] S2, extract the first feature point set of the lower surface of the workpiece and the second feature point set of the upper surface of the support base from the preprocessed image.

[0046] Specifically, the adaptive Canny edge detection algorithm is used, which automatically calculates the optimal threshold based on the gradient characteristics of different regions of the image, thereby accurately capturing the target edges. Subsequently, the findContours function of the contour detection algorithm is combined with a pre-trained deep learning model to collaboratively extract the contours of the workpiece and the support base. This method effectively focuses on key areas such as the contact edges between the lower surface of the workpiece and the upper surface of the support base, while suppressing and removing interference from other irrelevant contours.

[0047] Furthermore, to achieve high-precision recognition of various workpieces, the deep learning model is divided into general models and special models according to the structure of the workpiece, and the model is continuously optimized through transfer learning and online feedback mechanisms.

[0048] Specifically, we will construct general and specific models: General Model: Established for standard circular workpieces. The model is trained using images of the standard workpiece acquired under various lighting and angle conditions, with its center point and evenly distributed feature points along the edges labeled to give the model basic recognition capabilities.

[0049] Special models are created for workpieces with special structures such as roundness deviations (e.g., ellipses), parallelism errors, or surface runout. In addition to annotating the basic feature points, special features are additionally annotated, such as the endpoints of the major and minor axes of an ellipse, and protrusions that cause surface runout. Data augmentation techniques (rotation, scaling, translation) are used to expand the dataset to specifically identify these unconventional features.

[0050] The model training employs a transfer learning process to balance recognition efficiency and accuracy. The specific process is as follows: A convolutional neural network (CNN) model is fully trained using a general dataset to learn the basic visual features of a circular workpiece (such as arcs and edges).

[0051] The parameters of a pre-trained CNN model are transferred to a specialized model as initial parameters for fine-tuning. Subsequently, on a dataset with a specific structure, the parameters of the later layers of the model are finely tuned with a small learning rate. This strategy allows the specialized model to utilize general features while quickly focusing on learning the details of the specific structure, effectively avoiding overfitting caused by insufficient data.

[0052] The trained model is used to assist contour detection algorithms (such as findContours). The model first performs preliminary identification and classification, and then the algorithm extracts the contours accurately, so as to stably lock the key contours even in complex backgrounds and eliminate noise interference.

[0053] Built-in online learning mechanism to maintain and improve model performance: The image processing results acquired in real time during the compression process are compared with the model prediction. When the error is large, a feedback signal is triggered to analyze the cause and correct the model's input parameters in real time.

[0054] Newly acquired qualified workpiece images are automatically included in the training dataset. The system periodically uses the new data to fine-tune the model weights, enabling the system to adapt to slow changes in workpieces or new product types and maintain long-term recognition accuracy.

[0055] After obtaining a clear outline, the MatchFormer feature point recognition algorithm based on the Transformer architecture is used to detect key feature points in the image. This algorithm can effectively capture the global contextual relationships between feature points, improving recognition robustness. The system compares the identified feature points with preset model feature points, thereby achieving accurate positioning and recognition of the workpiece.

[0056] Among them, the feature points include the center point of the lower surface of the workpiece ( x c , y c , z c ), and 6 feature edge points, with the rightmost point on the horizontal axis as edge point 1 ( ), x 1, y 1, z1), take a feature edge point every 60° in a clockwise direction, which are edge points 2 ( x 2, y 2, z2), edge point 3 ( x 3, y 3, z3), edge point 4 ( x 4, y 4, z4), edge point 5 ( x 5, y 5, z5), edge point 6 ( x 6, y 6, z6); Center point of the upper surface of the support base ( x c0 , y c0 , z c0 ), and 6 feature points evenly distributed along the edge: ( x 10 , y 10 , z 10 ), ( x 20 , y 20 , z 20 ), ( x 30 , y30 , z 30 ), ( x 40 , y 40 , z 40 ), ( x 50 , y 50 , z 50 ), ( x 60 , y 60 , z 60 ).

[0057] S3, calculate the positional deviation between the workpiece and the support base based on the first feature point set and the second feature point set, including distance deviation and angle deviation.

[0058] Specifically, calculate the distance deviation between corresponding feature points on the lower surface of the workpiece and the upper surface of the support base. Taking the center point distance deviation as an example, the distance deviation value is calculated using the following formula: ; Where, Δ d The distance between the centers of the two circles and the preset target distance d The difference of 0; α This is the error correction factor.

[0059] Angular deviation is used to measure the degree of tilt of the workpiece relative to the support base. Specifically, it is the angle between the line connecting the two center points and the preset target direction (i.e., the horizontal axis) and the preset target angle. θ The difference Δ to 0 θ It can be calculated using the following formula: ; in,( x ref ，y ref ，z ref () is the unit vector of the preset target direction; β This is the correction factor for shape asymmetry.

[0060] S4, the fuzzy PID controller dynamically adjusts the PID parameters according to the magnitude and trend of the distance deviation and angle deviation, and outputs the feed amount of the pressing mechanism.

[0061] Specifically, in order to precisely control the action of the pressing mechanism, a fuzzy PID control algorithm is used to dynamically adjust the PID parameters based on the real-time detected distance and angle deviations and historical data, so as to achieve fast, stable and overshoot-free precise pressing, thereby effectively suppressing overshoot and oscillation.

[0062] The input to the fuzzy PID controller is the position deviation. e ( k )=Δ d ( k )+Δ θ ( k The feed rate of the pressing mechanism is calculated using the following formula: ; in, u ( k ) is the feed rate at the k-th sampling time; K p , K i , K d These are the parameters of a fuzzy PID controller; e ( k ) represents the positional deviation at the k-th sampling time; Δ d ( k ) is the distance deviation at the k-th sampling time; Δ θ ( k ) is the angle deviation at the k-th sampling time.

[0063] S5: Control the pressing mechanism to perform the pressing action according to the feed amount, and continuously acquire images during the pressing process and repeat S1 to S4.

[0064] Furthermore, to achieve continuous optimization, a historical data feedback mechanism was introduced, the specific process of which is as follows: Record the key parameters of each workpiece during the pressing process, including feature point coordinates, distance deviation, angle deviation, and feed rate; Compare the feature point data matrix of the current workpiece with the data matrix of the previous workpiece to analyze the offset direction, magnitude and trend of the feature points; Based on historical analysis results, a feedforward correction Δ is generated. u hist The feed rate is then adjusted, and the final feed rate command is: ; in, This is the corrected feed rate; It is a correction factor.

[0065] If historical data analysis indicates an overshoot trend in previous tasks, the system will automatically reduce... K p and K d The value of is adjusted to improve system stability.

[0066] Furthermore, to ensure operational safety and process reliability, the system is equipped with a multi-layered early warning mechanism: During the downward pressure process, the system continuously compares the distance deviation and angle deviation calculated in real time with the preset target tolerance; Set a pressure threshold. When the current pressure exceeds the threshold or the position deviation cannot be corrected, the system will immediately stop operation, issue an audible and visual alarm, and notify the operator to intervene and check. All key parameters of alarm events (such as deviation, feed rate, and pressure curve) are fully recorded, providing data support for tracing the root cause of problems and optimizing the process model.

[0067] It should be noted that the present invention is not limited to the above-described embodiments. The above embodiments are merely examples, and any embodiments that have the same structure and perform the same effects as the technical concept within the scope of the present invention are included within the scope of the present invention. Furthermore, various modifications that can be conceived by those skilled in the art to the embodiments, and other ways of constructing by combining some of the constituent elements of the embodiments, without departing from the spirit of the present invention, are also included within the scope of the present invention.

Claims

1. A control method for a controllable automatic clamping hydraulic clamping device, characterized in that, The hydraulic clamping device includes: The clamping mechanism includes a fixed-side clamp A (1) and a movable-side clamp B (2). The lower bottom surface of the clamp A (1) is bolted to the upper surface of the clamp V-block A (3) and installed on one side of the hydraulic clamping device housing (8). The lower bottom surface of the clamp B (2) is bolted to the upper surface of the clamp V-block B (4) and installed on the side of the movable seat (18) of the clamping device. The support mechanism includes a positioning base (6) mounted on the worktable of the hydraulic clamping device housing (8) and a support base (5) connected above the positioning base (6). The drive mechanism includes a lead screw (7) installed at the lower end of the clamping device movable seat (18). The rotation of the lead screw (7) drives the jaw B (2) to feed laterally through the clamping device movable seat (18). The pressing mechanism includes a column (25) fixed to one side of the zero point position of the surface grinder table, a metal shell (23) connected to the column (25) by a fixing ring (24), a telescopic plate (19) set on the slide rail (22) on the bottom surface of the metal shell (23), a pneumatic motor that drives the telescopic plate (19), a cylinder (20) installed at the end of the telescopic plate (19), and a pressing plate (21) driven by the cylinder (20) to perform the pressing action. Camera setup, including an industrial camera mounted on the column (25); The method includes the following steps: S1, acquires images of the workpiece and support base through an industrial camera, and preprocesses the images; S2, extract the first feature point set of the lower surface of the workpiece and the second feature point set of the upper surface of the support base from the preprocessed image; S3, calculate the positional deviation between the workpiece and the support base based on the first feature point set and the second feature point set, including distance deviation and angle deviation; S4, the fuzzy PID controller dynamically adjusts the PID parameters according to the magnitude and trend of the distance deviation and angle deviation, and outputs the feed amount of the pressing mechanism; S5: Control the pressing mechanism to perform the pressing action according to the feed amount, and continuously acquire images during the pressing process and repeat S1 to S4.

2. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 1, characterized in that, The drive mechanism also includes a servo motor and a geared motor (9). The output end of the servo motor and the geared motor (9) is connected to the lead screw (7) via a synchronous pulley (13) and a synchronous shaft (14). The servo motor and the geared motor (9) are equipped with a torque sensor.

3. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 2, characterized in that, The servo motor and geared motor (9) are mounted and fixed by the motor mounting plate (11) and the motor fixing plate (12). The servo motor and geared motor (9) are provided with a clamping device protective cover (16) on the outside. A transparent PVC plate (17) and a handle cover plate (15) are installed on the side of the clamping device protective cover (16).

4. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 1, characterized in that, The gripper V-block A (3) and gripper V-block B (4) are inverted trapezoidal in shape; the support base (5) is concave and has a drainage hole at the bottom; the hydraulic clamping device housing (8) has a hollow bottom and drainage slots on both sides.

5. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 1, characterized in that, The drive mechanism also includes a bearing housing and a bearing housing plate (10) for supporting the lead screw (7).

6. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 1, characterized in that, The image preprocessing specifically includes: An adaptive median filtering algorithm is used to analyze the image pixel by pixel to remove noise while preserving edge details; Dynamic histogram equalization is used to enhance image contrast. Image correction algorithms are used to correct distortion and adjust the viewing angle of images that are not taken from the front.

7. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 1, characterized in that, S2 specifically includes: The adaptive Canny edge detection algorithm is used to obtain the edge information of the image, and the contour detection algorithm and deep learning model are used to extract the contours of the workpiece and the support base. A feature point recognition algorithm based on the Transformer architecture is used to detect key feature points in an image, and these feature points are compared with preset model feature points to locate and identify the workpiece.

8. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 7, characterized in that, The deep learning model is divided into a general model and a special model based on the structure of the workpiece, and is trained in the following way: General Model: The initial deep learning model is pre-trained using a general dataset labeled with feature points of standard circular workpieces; Special model: The parameters of the pre-trained model are transferred to the new model, and the new model is fine-tuned using a special dataset labeled with feature points of non-standard circular workpieces.

9. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 1, characterized in that, The method further includes: Record the feature point data and feed rate of the current workpiece, and compare them with the data of the previous workpiece. Based on the comparison results, dynamically adjust the parameters of the fuzzy PID controller and the feed rate.

Images