Controllable automatic clamping hydraulic clamping device and control method thereof

Abstract

The invention 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 shooting mechanism. The clamping mechanism clamps a workpiece through cooperation of a clamping jaw A and a clamping jaw B; the driving mechanism adopts a servo motor to drive a screw rod to drive a movable seat to realize transverse feeding, and a torque sensor in the driving mechanism realizes accurate positioning; the downward pressing mechanism drives a downward pressing flat plate through an air cylinder to conduct flatness correction on the surface of the workpiece. The camera mechanism is used for collecting workpiece images. According to the control method, through image preprocessing, feature point extraction and position deviation calculation, the pressing feed amount is dynamically adjusted through a fuzzy PID controller, and self-adaptive optimization is carried out in combination with historical data. According to the invention, full automation and high-precision control of workpiece clamping are realized, and the machining quality and the production efficiency are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydraulic clamping, particularly to a controllable automatic clamping hydraulic clamping device and a control method thereof. BACKGROUND

[0002] In the parts processing, manufacturing and assembly industry, clamping devices are widely used to ensure the stability of workpieces during processing, especially in high-precision processing links such as surface grinding machines, numerical control machine tools, and automated assembly lines. The clamping precision directly affects the processing quality and efficiency. Traditional hydraulic clamping systems rely on mechanical transmission (such as screw rotation) to achieve clamping, lacking precise control over the positioning and pressing process of workpieces, resulting in difficulty in accurately adjusting the clamping force, and prone to problems such as excessive or insufficient clamping force, which in turn causes damage to the workpiece surface, unqualified flatness, or repeated clamping, affecting processing precision and increasing production costs.

[0003] In addition, existing devices generally have low automation, making it difficult to meet the needs of modern production lines with high speed and high precision, especially in terms of adaptability to different structural workpieces (such as ring-shaped parts with large roundness deviation).

[0004] Therefore, there is an urgent need to develop an intelligent hydraulic clamping device with precise travel control, self-adaptive clamping, and real-time feedback capability to improve clamping precision, efficiency, and product quality consistency. SUMMARY

[0005] To solve the problems of low positioning accuracy, inaccurate clamping force control, and insufficient automation of existing clamping devices, the present application provides a controllable automatic clamping hydraulic clamping device and a control method thereof, which integrates visual detection, intelligent control, and precise execution mechanism to achieve high-precision automatic positioning, self-adaptive clamping, and flatness intelligent correction of workpieces, effectively improving processing quality and production efficiency.

[0006] In a first aspect, the present application provides a controllable automatic clamping hydraulic clamping device, comprising: A clamping mechanism including a fixed-side clamping jaw A and a movable-side clamping jaw B, the lower bottom surface of the clamping jaw A is bolted to the upper surface of the clamping jaw V-block A and installed on one side of the hydraulic clamping device housing; the lower bottom surface of the clamping jaw B is bolted to the upper surface of the clamping jaw V-block B and installed on the side surface of the clamping device movable seat; A support mechanism including a positioning base installed on the workbench surface of the hydraulic clamping device housing, and a support base connected above the positioning base; A drive mechanism including a screw rod installed at the lower end of the clamping device movable seat, the rotation of the screw rod drives the transverse feed of the clamping jaw B through the clamping device movable seat; The down-pressing mechanism comprises a column fixed to one side of the zero position of the worktable of the surface grinder, a metal shell connected with the column through a fixing ring, an extendable flat plate arranged on the slide rail at the bottom of the metal shell, a pneumatic motor driving the extendable flat plate, a pneumatic cylinder mounted at the end of the extendable flat plate, and a down-pressing flat plate driven by the pneumatic cylinder to perform the down-pressing action. The camera mechanism comprises an industrial camera mounted on the column.

[0007] Further, the driving mechanism further comprises a servo motor and a speed reducer, the output end of the servo motor is in transmission connection with the lead screw through a synchronous wheel and a synchronous shaft, and a torque sensor is arranged inside the servo motor.

[0008] By introducing the servo motor, synchronous wheel shaft transmission and built-in torque sensor, accurate control and intelligent feedback of the driving process are realized. The feeding position of the movable side can be accurately controlled, and whether the workpiece is in contact is judged in real time through torque sensing, so that the surface damage or deformation of the workpiece caused by excessive pre-clamping force is avoided, and the precision and safety of clamping are significantly improved.

[0009] Further, the servo motor and the speed reducer are installed and fixed through a motor mounting plate and a motor fixing plate, a clamping device protective cover is arranged outside the servo motor and the speed reducer, and a transparent PVC plate and a handle cover plate are mounted on the side surface of the clamping device protective cover.

[0010] By arranging the motor fixing plate, the protective cover, the transparent PVC plate and the handle cover plate, a stable installation foundation is provided for the motor, the stability of transmission is ensured, the cooling liquid and the chips are effectively isolated through the protective cover, and the service life of the motor and the sensor is prolonged. The transparent PVC plate facilitates observation of the internal running state, and the handle cover plate greatly facilitates daily maintenance and repair, improving the ease of use and reliability of the equipment.

[0011] Further, the clamping jaw V-shaped block A and the clamping jaw V-shaped block B are inverted trapezoidal, the support base is concave and has a drainage hole at the bottom, and the hydraulic clamping device housing is hollow at the bottom and has drainage grooves at the two sides.

[0012] Through the special inverted trapezoidal, concave with drainage hole and hollow drainage groove design of the V-shaped block, the support base and the device housing, the self-cleaning and anti-rust ability of the device is significantly improved. The cutting fluid and the chips are difficult to accumulate and can be quickly discharged, the clamping reference surface is kept clean, the maintenance frequency is reduced, and the risk of rusting of the internal components due to liquid accumulation is fundamentally reduced, ensuring the long-term clamping accuracy.

[0013] Further, the driving mechanism further comprises a bearing seat and a bearing seat plate for supporting the lead screw.

[0014] The bearing seat and bearing seat plate for supporting the lead screw are arranged in a specific manner, thereby providing stable and reliable rotation support for the lead screw, effectively inhibiting the radial jumping and axial movement of the lead screw during transmission, which ensures that the movable seat of the clamping device can realize smooth and accurate linear feeding, thereby improving the repeat positioning accuracy and long-term operation stability of the entire clamping process.

[0015] In a second aspect, the present application also provides a control method of a controllable automatic clamping hydraulic clamping device, comprising the following steps: S1, acquiring the image of the workpiece and the support base through an industrial camera, and pre-processing the image; S2, extracting a first feature point set of the lower surface of the workpiece and a second feature point set of the upper surface of the support base from the pre-processed image; S3, calculating the position deviation between the workpiece and the support base according to 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 size and change trend of the distance deviation and the angle deviation, and outputs the feed amount of the pressing mechanism; S5, controlling the pressing mechanism to perform pressing action according to the feed amount, and continuously acquiring images during the pressing process and repeating S1 to S4.

[0016] Further, the method further comprises: recording the feature point data of the current workpiece and the feed amount, and comparing with the data of the previous workpiece, and dynamically correcting the parameters of the fuzzy PID controller and the feed amount according to the comparison result.

[0017] The above scheme provides an intelligent control method matched with the device, and the core is to realize the intelligentization and self-adaptive optimization of the clamping process. Through visual detection, fuzzy PID control and data iterative learning, the system can actively perceive and compensate the position and attitude deviation of the workpiece, dynamically adjust the pressing amount, not only significantly improve the flatness correction ability for different workpieces, but also make the system have self-learning function, can be more and more accurate, greatly improve the production quality and automation level.

[0018] Further, the pre-processing of the image specifically comprises: Using an adaptive median filtering algorithm to analyze the image pixel by pixel to remove noise while preserving edge details; Using dynamic histogram equalization technology to enhance the contrast of the image; Through an image correction algorithm, the image taken from a non-frontal view is corrected for distortion and adjusted for viewing angle.

[0019] Through adaptive median filtering, dynamic histogram equalization and image correction technology, clear, accurate and high-contrast images can still be obtained in complex industrial environments such as uneven lighting, noise and inclined shooting angle, providing a high-quality data basis for subsequent feature extraction and deviation calculation, and ensuring the robustness and reliability of the vision system.

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

[0021] By combining the efficiency of traditional edge detection and the powerful recognition ability of deep learning model, especially introducing the Transformer architecture, it can more accurately identify and locate the feature points of the workpiece under complex or occluded conditions. This greatly improves the recognition success rate and positioning accuracy of the system for diversified workpieces, enhancing the versatility of the equipment.

[0022] Further, the deep learning model is divided into a general model and a special model according to 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 standard circular structure workpiece feature points; Special model: the pre-trained model parameters are migrated to a new model, and the new model is fine-tuned using a special dataset labeled with non-standard circular structure workpiece feature points.

[0023] By designing a transfer learning training strategy combining general and special models, the system not only ensures fast and accurate processing capability for common standard workpieces (such as circular structure), but also enables it to quickly adapt to new, non-standard shaped workpieces with a small amount of samples. This training method reduces the need for a large amount of special workpiece labeled data, shortens the model deployment and optimization cycle, and makes the device flexible to adapt to changing production tasks.

[0024] Compared with the prior art, the beneficial effects of the present application 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; the inverted trapezoidal V-shaped block, the support base with drainage holes and the hollow shell design effectively drain cutting fluid and debris, and have the advantages of rust prevention and easy maintenance; the servo motor is equipped with a torque sensor, combined with the "pre-clamping-retracting-hydraulic clamping" process, which significantly improves the reliability and repeatability of the clamping process.

[0025] 2. Adopting fuzzy PID control algorithm, dynamically adjusting the pressing feed according to real-time position deviation, realizing adaptive control of clamping force and flatness; introducing transfer learning mechanism, combining general and special models, enabling the system to quickly adapt to different structure workpieces, enhancing generalization ability; recording historical data and comparing with current state, continuously optimizing control parameters and identification model, having online learning and continuous evolution ability; the whole clamping process is automatically completed from positioning, clamping to correction, which greatly improves production efficiency while ensuring consistent accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments or prior art in the present drawings, the drawings needed to be used in the following embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the drawings, and other drawings can be obtained according to the structures shown in the drawings without creative labor for those skilled in the art.

[0027] Figure 1 The structure diagram of the hydraulic clamping device is shown in the present application. Figure 2 The structure diagram of the pressing mechanism is shown in the present application. Figure 3 The method flow chart of the present application.

[0028] In the figure: 1, jaw A; 2, jaw B; 3, jaw V block A; 4, jaw V block B; 5, support base; 6, positioning base; 7, lead screw; 8, hydraulic clamping device housing; 9, servo motor and reduction motor; 10, bearing seat and bearing seat plate; 11, motor mounting plate; 12, motor fixing plate; 13, synchronous wheel; 14, synchronous shaft; 15, handle cover plate; 16, clamping device protective cover; 17, transparent PVC plate; 18, clamping device movable seat; 19, telescopic flat plate; 20, air cylinder; 21, pressing flat plate; 22, bottom surface slide rail; 23, metal shell; 24, fixing ring; 25, stand column. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application is described and explained in the following with reference to the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application, and are not used to limit the present application. Based on the examples provided by the present application, all other examples obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0030] The present application provides a controllable automatic clamping hydraulic clamping device, as shown inFigure 1 、 Figure 2 as shown in the drawings, comprising: a clamping mechanism, including a fixed side clamping jaw A1 and a movable side clamping jaw B2, the lower bottom surface of the clamping jaw A1 is bolted to the upper surface of the clamping jaw V-shaped block A3 and is installed on one side of the hydraulic clamping device housing 8; the lower bottom surface of the clamping jaw B2 is bolted to the upper surface of the clamping jaw V-shaped block B4 and is installed on the side of the clamping device movable seat 18; a supporting mechanism, including a positioning base 6 installed on the workbench surface of the hydraulic clamping device housing 8, and a supporting base 5 connected above the positioning base 6; a driving mechanism, including 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 transverse feed of the clamping jaw B2 through the clamping device movable seat 18; a pressing mechanism, including a column 25 fixed to one side of the zero position of the workbench of the surface grinder, a metal shell 23 connected to the column 25 through a fixed ring 24, a telescopic plate 19 arranged on the slide rail 22 of the bottom surface of the metal shell 23, a pneumatic motor driving the telescopic plate 19, a pneumatic cylinder 20 installed at the end of the telescopic plate 19, and a pressing plate 21 driven by the pneumatic cylinder 20 to perform the pressing action; a camera mechanism, including an industrial camera installed on the column 25.

[0031] Further, the driving mechanism further comprises a servo motor and a reduction motor 9, the output end of the servo motor 9 is drivingly connected with the lead screw 7 through a synchronous wheel 13 and a synchronous shaft 14; and the servo motor 9 is internally provided with a torque sensor.

[0032] Further, the servo motor and the reduction motor 9 are installed and fixed through a motor mounting plate 11 and a motor fixing plate 12, the servo motor and the reduction motor 9 are externally provided with a clamping device protective cover 16, and the side surface of the clamping device protective cover 16 is installed with a transparent PVC plate 17 and a handle cover plate 15.

[0033] Further, the clamping jaw V-shaped block A3 and the clamping jaw V-shaped block B4 are inverted trapezoidal; the supporting base 5 is concave and provided with a drainage hole at the bottom; and the bottom of the hydraulic clamping device housing 8 is hollow and provided with drainage grooves on both sides.

[0034] Further, the driving mechanism further comprises a bearing block and a bearing block plate 10 for supporting the lead screw 7.

[0035] Working principle: The control system first starts the driving mechanism. Specifically, the servo motor and the reduction motor 9 begin to rotate, and their power is accurately transmitted to the lead screw 7 through the synchronous belt transmission system composed of the synchronous wheel 13 and the synchronous wheel shaft 14. The lead screw 7 rotates under the stable support of the bearing seat and bearing seat plate 10 at both ends, driving the clamping device movable seat 18 to move smoothly along the straight rail of the fixed side of the workbench surface. The clamping jaw B2 and the clamping jaw V block B4 fixed on the clamping device movable seat 18 advance with the front, and the clamping jaw A1 and the clamping jaw V block A3 of the fixed side cooperate to approach and contact the workpiece from both sides, completing the preliminary positioning and pre-clamping of the workpiece.

[0036] During the process of driving the workpiece by the driving mechanism, the torque sensor built into the servo motor 9 continuously monitors the motor output torque. When the workpiece is pushed to the predetermined clamping position, the resistance increases, causing the torque to rise to the preset threshold, and the torque sensor immediately feeds back this signal to the control system. The control system then issues an instruction to control the servo motor 9 to briefly reverse, driving the clamping device movable seat 18 and the movable side clamping jaw to retreat a precise distance (e.g. 3mm). The purpose of this retreat action is to actively eliminate the possible feed error caused by the invasion of workbench surface dust or other small particles, providing a pure and accurate reference position for subsequent main clamping.

[0037] After the movable side retreats to the position, the control system immediately starts the hydraulic clamping mechanism (components such as hydraulic cylinders not shown in detail). The hydraulic clamping mechanism drives the clamping device movable seat 18 to advance again, firmly clamping the workpiece on 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 solve the problem that the workpiece may be slightly warped due to clamping force, affecting the flatness of the upper surface, the device is designed with a pressing mechanism. After hydraulic clamping is completed, the control system instructs the pressing mechanism to act: first, the pneumatic motor drives the telescopic flat plate 19 to extend along the slide rail 22 at the bottom of the metal shell 23, causing the pressing flat plate 21 to move to the top of the workpiece. Then, the cylinder 20 drives the pressing flat plate 21 to descend, and the nylon or polyurethane protective plate on its contact surface cushions the impact and applies a uniform vertical pressing force to the upper surface of the workpiece. Under this pressing force, the control system controls the hydraulic clamping force to perform a short cycle of "loose-again clamping". This operation can effectively release the uneven stress inside the workpiece caused by initial clamping, prompting the workpiece to adjust its position in the vertical direction, thereby correcting the flatness of its upper surface and ensuring that it meets the requirements of high-precision machining. After the correction is completed, the pressing flat plate 21 is retracted first, and then the telescopic flat plate 19 is retracted to the initial position, and the clamping process is completed.

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

[0040] The application also provides a control method of the controllable automatic clamping hydraulic clamping device, as shown in the figure, comprising the following steps: Figure 3 S1, collecting the image of the workpiece and the support base by the industrial camera, and pre-processing the image.

[0041] Specifically, one high-resolution industrial camera is installed on the lower pressing mechanism column in front and rear directions, and each is provided with a ring-shaped LED light source (the wavelength of the light source matches the material of the workpiece, and the surface reflection is eliminated), and the height of the camera can be adjusted according to the height of the support base, so that the image 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 in the pressing process.

[0042] In order to improve the recognition accuracy of the vision system, the collected original image is pre-processed in multiple steps: First, the adaptive median filter algorithm is used for denoising, which can dynamically adjust the filter window size according to the noise condition of the local area of the image, effectively suppress the noise, and maximize the retention of the edge details of the workpiece and the support base.

[0043] On this basis, the dynamic histogram equalization technology is used to enhance the contrast of the image. This method can adaptively determine the optimal pixel value mapping range, avoid excessive enhancement of local areas, and make the boundary profile of the lower surface of the workpiece and the upper surface of the support base more clear and sharp.

[0044] In view of the problem that the industrial camera cannot be frontally photographed due to the limitation of the installation position, the image correction technology is introduced. Through the image correction algorithm, the image distortion and perspective deformation caused by the angle are compensated and adjusted, the image is uniformly corrected to the standard normal angle reference, so that the image features collected in different batches or at different times have high consistency and comparability.

[0045] S2, extracting 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 pre-processed image.

[0046] ​Specifically, an adaptive Canny edge detection algorithm is used, which can automatically calculate the optimal threshold according to the gradient characteristics of different regions of the image, thereby accurately capturing the target edge. Subsequently, in combination with the findContours function of the contour detection algorithm and the pre-trained deep learning model, the contours of the workpiece and the supporting base are extracted. This method can effectively focus on the contact edge between the lower surface of the workpiece and the upper surface of the supporting base, and at the same time suppress and remove the interference of other irrelevant contours.

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

[0048] Specifically, the general model and the special model are constructed: General model: established for standard circular structure workpieces. Standard workpiece images collected under various lighting and angles are used for training, and the center point and evenly distributed feature points of the edge are labeled, so that the model has basic recognition ability.

[0049] Special model: established for workpieces with special structures such as roundness deviation (e.g. ellipse), parallelism error or surface jump. In addition to labeling basic feature points, special features such as the endpoints of the major and minor axes of an ellipse, and the raised parts causing surface jump are labeled, and data augmentation techniques (rotation, scaling, translation) are used to expand the data set to specifically identify these irregular features.

[0050] The model training adopts 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 general data sets, so that it learns the basic visual features of ring-shaped workpieces (such as circular arc and edge).

[0051] The pre-trained CNN model parameters are transferred to the special model as the initial parameters for fine-tuning. Then, on the special structure data set, the parameters of the last few layers of the model are fine-tuned with a small learning rate. This strategy allows the special model to utilize general features while quickly focusing on learning the details of special structures, effectively avoiding overfitting due to insufficient data.

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

[0053] An online learning mechanism is built in to maintain and improve the performance of the model: The processing result of the image collected in real time during the pressing process is compared with the model prediction, and when the error is large, a feedback signal is triggered, the reason is analyzed, and the input parameters of the model can be corrected in real time.

[0054] The newly collected qualified workpiece image is automatically included in the training data set, and the system regularly fine-tunes the model weight using new data, so that the system can adapt to the slow changes of workpieces or new product types, and maintain long-term recognition accuracy.

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

[0056] Among them, the feature points include the center point of the lower surface of the workpiece (x0, y0, z0), and 6 feature edge points, with the first point on the right of the horizontal axis as the edge point 1 (x1, y1, z1), and every 60° in clockwise direction. Take a feature edge point, in turn, edge point 2 (x2, y2, z2), edge point 3 (x3, y3, z3), edge point 4 (x4, y4, z4), edge point 5 (x5, y5, z5), and edge point 6 (x6, y6, z6). x c , y c , z c ), x 1, y 1, z1), x 2, y 2, z2), x 3, y 3, z3), x 4, y 4, z4), x 5, y 5, z5), x 6, y 6, z6). The center point of the upper surface of the support base (x0, y0, z0), and 6 feature points evenly distributed on the edge: (x1, y1, z1), (x2, y2, z2), (x3, y3, z3), (x4, y4, z4), (x5, y5, z5), (x6, y6, z6). x c0 , y c0 , z c0 ), 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, calculating the position deviation between the workpiece and the support base according to the first feature point set and the second feature point set, including distance deviation and angle deviation.

[0058] Specifically, the distance deviation value of the corresponding feature points of the lower surface of the workpiece and the upper surface of the support base is calculated, taking the central point distance deviation value as an example, which is calculated by the following formula: ; Where Δ d is the difference between the distance between the two centers and the preset target distance d 0; α is the error correction coefficient.

[0059] The angle deviation is used to measure the degree of inclination of the workpiece relative to the support base, specifically the difference Δ θ between the included angle of the line connecting the two center points and the preset target direction (i.e. the horizontal axis) and the preset target angle θ 0, which is calculated by the following formula: ; Where (e x ref ，y ref ，z ref ) is the unit vector of the preset target direction; β is the shape asymmetry correction coefficient.

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

[0061] Specifically, to accurately control the action of the pressing mechanism, a fuzzy PID control algorithm is adopted to dynamically adjust the PID parameters according to the real-time detected distance deviation and angle deviation and historical data, so as to realize precise pressing with rapidity, stability and no overshoot, thereby effectively inhibiting overshoot and oscillation.

[0062] The input of the fuzzy PID controller is the position deviation e ( k )=Δ d ( k )+Δ θ ( k ), and the feed amount of the pressing mechanism is calculated by the following formula: ; Wherein, u ( k ) is the feed amount at the kth sampling time; K p 、 K i 、 K d are parameters of the fuzzy PID controller; e ( k ) is the position deviation at the kth sampling time; Δ d ( k ) is the distance deviation at the kth sampling time; Δ θ ( k ) is the angle deviation at the kth sampling time.

[0063] S5, according to the feed amount, the pressing mechanism is controlled to perform the pressing action, and the image is continuously collected during the pressing process and S1 to S4 are repeated.

[0064] Further, to realize continuous optimization, a historical data feedback mechanism is introduced, and the specific process is as follows: Record the key parameters of each workpiece during the pressing process, including feature point coordinates, distance deviation, angle deviation and feed amount; Compare the feature point data matrix of the current workpiece with the data matrix of the previous workpiece, analyze the offset direction, amplitude and trend of the feature points; According to the historical analysis result, a feedforward correction amount Δ u hist is generated, and the feed amount is corrected, and the final feed amount instruction is: ; Wherein, is the corrected feed amount; is a correction coefficient.

[0065] If the historical data analysis shows that there is an overshoot trend in the previous task, the system will automatically reduce the value of K p and K d to improve the stability of the system.

[0066] Further, to ensure operation safety and process reliability, the system sets up a multi-layer early warning mechanism: During the pressing process, the system continuously compares the distance deviation, angle deviation calculated in real time with the preset target tolerance; Set the pressure threshold, when the pressing force exceeds the threshold, or the position deviation cannot be corrected continuously, the system immediately suspends the operation, issues an audible and visual alarm, and notifies the operator to intervene and check; All key parameters of the alarm event (such as deviation, feed amount, pressure curve) are recorded completely, providing data support for subsequent tracing of problem sources and optimization of process model.

[0067] It should be noted that the present application is not limited to the above embodiments. The above embodiments are only examples, and embodiments having the same technical idea and playing the same role and effect within the scope of the technical solutions of the present application are all included in the technical scope of the present application. In addition, within the scope of the main idea of the present application, various modifications of the embodiments that can be thought of by those skilled in the art, and other ways constructed by combining part of the constituent elements in the embodiments are also included in the scope of the present application.

Claims

1. A controllable automatic clamping hydraulic clamping device, characterized in that, include: 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. The camera setup includes an industrial camera mounted on the column (25).

2. The 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 (9) is connected to the lead screw (7) via a synchronous pulley (13) and a synchronous shaft (14). The servo motor (9) is equipped with a torque sensor.

3. The 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 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 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. A control method for a controllable automatic clamping hydraulic clamping device according to any one of claims 1-5, characterized in that, 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.

7. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 6, 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.

8. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 6, 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.

9. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 8, 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.

10. The control method for a controllable automatic clamping hydraulic clamping device as described in claim 6, 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