Surface grinding device and grinding method for metal parts processing

Abstract

The present invention discloses a surface grinding device and a grinding method for metal parts processing, which belongs to the technical field of metal parts processing, and includes a processing table and a vacuum suction cup, wherein the vacuum suction cup is arranged at the upper end of the processing table; a first motor, wherein the first motor is installed at a position close to one end of the upper surface of the processing table, and the first motor is used to drive the vacuum suction cup to move left and right on the processing table; and a support frame, wherein the support frame is installed at the upper end of the processing table. The present invention can realize that during the grinding process, the dynamic change of the grinding force is captured in real time by a pressure sensor. After grinding a short section of the path, the second camera immediately detects the grinding effect. Once it is found that the ideal standard is not reached, the cylinder parameters are immediately fine-tuned to increase the grinding force. This dynamic adjustment mechanism ensures that the whole grinding process is advanced toward the ideal effect, reduces the scrap rate, and the data of each adjustment will also be fed back to the neural network model to continuously optimize the model prediction accuracy.

Description

Technical Field

[0001] The invention relates to the field of metal parts processing, and more specifically, to a surface grinding device and a grinding method for metal parts processing. Background Art

[0002] In the metal parts processing industry, surface grinding is undoubtedly a crucial process. After a series of processing processes such as cutting, welding, and casting, metal parts often leave various defects on their surfaces, such as burrs, rust, unevenness, and processing marks. These surface defects not only damage the appearance of the parts, but more importantly, they may seriously reduce the matching accuracy of the parts, weaken their mechanical properties, and even cause malfunctions or early failures during the use of the parts, thereby affecting the normal operation and service life of the entire mechanical system.

[0003] Therefore, it is particularly important to perform meticulous grinding operations on processed parts. However, under existing technical conditions, there are many challenges in batch grinding of metal parts. For some parts with regular shapes and easy to grind, such as plate parts, automated grinding has been achieved, which greatly improves the grinding efficiency and consistency. However, for some special-shaped parts, due to their complex shapes and varied curves, automated grinding equipment is difficult to adapt, so small handheld grinders are often used for grinding operations.

[0004] This manual grinding method has many disadvantages. First, the labor intensity is high. Workers need to hold the grinder for a long time to work, which can easily lead to hand fatigue and injury. Second, the efficiency is low. The speed of manual grinding is far from comparable to that of automated equipment, and it is difficult to meet the needs of large-scale production. More importantly, manual grinding is difficult to guarantee the quality and consistency of grinding. Due to differences in workers' technical level and experience, as well as subjective factors in the grinding process, the grinding effects vary, making it difficult to achieve a unified quality standard. Summary of the invention

[0005] In view of the problems existing in the prior art, the purpose of the present invention is to provide a surface grinding device and a grinding method for metal parts processing, which can realize the real-time capture of dynamic changes in grinding force through a pressure sensor during the grinding process. After grinding a short path, the second camera immediately detects the grinding effect. Once it is found that the ideal standard is not met, the cylinder parameters are immediately fine-tuned to increase the grinding force. This dynamic adjustment mechanism ensures that the entire grinding process is carried out towards the ideal effect, reducing the scrap rate, and the data of each adjustment will be fed back to the neural network model to continuously optimize the model prediction accuracy, forming a virtuous circle, making subsequent grinding operations more accurate and efficient.

[0006] To solve the above problems, the present invention adopts the following technical solutions.

[0007] A surface grinding device for metal parts processing, comprising:

[0008] A processing table and a vacuum suction cup, wherein the vacuum suction cup is arranged at the upper end of the processing table;

[0009] A first motor, which is mounted on an upper surface of the processing table near one end, and is used to drive the vacuum suction cup to move left and right on the processing table;

[0010] A support frame, the support frame is installed on the upper end of the processing table;

[0011] A horizontal plate, the horizontal plate is fixedly mounted on the inner wall of the support frame;

[0012] A first cylinder and a second cylinder, wherein the first cylinder and the second cylinder are arranged in multiple groups, and the multiple groups of the first cylinder and the second cylinder are all installed on the upper end of the support frame;

[0013] Grinding pieces, wherein the grinding pieces are arranged in multiple groups, and the multiple groups of grinding pieces are correspondingly arranged at positions below the first cylinder and the second cylinder;

[0014] A second motor, which is mounted on the upper end of the support frame and is used to drive the grinding piece to rotate;

[0015] A first camera, which is mounted at the lower end of the support frame and is used to obtain feature data of a metal part passing below;

[0016] The control module controls the corresponding first cylinder and the second cylinder to drive the polishing part to move to a corresponding height according to the feature data obtained by the first camera, and applies a corresponding extrusion force to the corresponding surface of the metal part.

[0017] Further, a first gear, wherein the first gear is mounted on an output shaft of the second motor;

[0018] A second gear, the second gear is rotatably mounted above the horizontal plate, the number of the second gears corresponds to the number of the first cylinders, a plurality of groups of the second gears are meshed and connected with each other, and one group of the second gears is meshed and connected with the first gear;

[0019] A first prism hole, wherein the first prism hole is opened through the center of the second gear;

[0020] A first prism, wherein the first prism is disposed inside the first prism hole, and the lower end of the first prism movably penetrates the horizontal plate and is fixedly connected to the corresponding polishing piece;

[0021] A connecting piece is installed between the upper end of the first prism and the first cylinder, and is used to ensure that the first prism rotates when the first cylinder pushes the first prism downward.

[0022] Further, a third gear, the third gear is arranged at the meshing connection of every two sets of second gears, and the third gear is meshed and connected with the corresponding second gear;

[0023] a second prism hole, wherein the second prism hole is formed through the center of the third gear;

[0024] The second prism is arranged in the second prism hole, the lower end of the second prism movably passes through the cross plate and is fixedly connected to the corresponding polishing piece, and the upper end of the second prism is connected to the second cylinder with a connecting piece of the same structure.

[0025] Furthermore, the connecting piece comprises:

[0026] A lower rotating box, wherein the lower rotating box is sleeved on the upper ends of the first prism and the second prism;

[0027] A first blocking plate, the first blocking plate is fixedly mounted on the upper ends of the first prism and the second prism, and the first blocking plate is arranged in the lower rotating box;

[0028] An upper rotating box, wherein the upper rotating box is sleeved on the lower ends of the first prism and the second prism, and the lower end of the upper rotating box is fixedly connected to the lower rotating box;

[0029] The second blocking plate is fixedly mounted on the lower ends of the first prism and the second prism, and the second blocking plate is arranged in the upper rotating box.

[0030] Furthermore, a pressure sensor is installed at the lower end of the second baffle plate, and the pressure sensor is electrically connected to the controller.

[0031] Furthermore, the polishing piece is composed of an elastic airbag and polishing bristles, the polishing bristles are evenly distributed on the outer surface of the elastic airbag, and the lower ends of the first prism and the second prism are fixedly connected to the corresponding elastic airbag.

[0032] Furthermore, a second camera is installed at the lower end of the horizontal plate and away from one end of the first camera, and the second camera is used to obtain characteristic data of the polished metal part.

[0033] Further, a screw rod, wherein the screw rod is fixedly mounted on the output shaft of the first motor;

[0034] A moving plate, wherein the moving plate is threadedly connected to a side surface of the lead screw, and the vacuum suction cup is mounted on the moving plate;

[0035] An optical axis, the optical axis being movably arranged at the front and rear ends of the movable plate;

[0036] The support blocks are arranged in multiple groups, and the support blocks are fixedly installed at the left and right ends of the optical axis. One group of support blocks is rotatably installed at the end of the screw rod away from the first motor, and the lower end of the support block is fixedly connected to the upper surface of the processing table.

[0037] The present invention also provides a grinding method, which is applicable to the above-mentioned surface grinding device for metal parts processing, and comprises the following steps:

[0038] Step 1, place the metal part on the vacuum suction cup, turn on the vacuum suction cup, and then drive the lead screw to rotate through the first motor to drive the vacuum suction cup to move from left to right;

[0039] Step 2, when the metal part moves with the vacuum suction cup to the bottom of the first camera, the first camera takes a picture of the passing metal part and obtains feature data of the metal part, wherein the feature data includes defect data and polishing area data on the surface of the metal part;

[0040] Step 3: The first camera controls the polishing piece to adjust to the corresponding height according to the acquired feature data. The neural network model outputs the required polishing force of the polishing piece according to the acquired feature data. The second motor drives the polishing piece to rotate and performs polishing operations on the passing metal parts.

[0041] According to the feature data acquired by the first camera, the polishing piece is controlled to be adjusted to a corresponding height, and the neural network model outputs the required polishing force of the polishing piece according to the acquired feature data, including:

[0042] Step 31, record the material of the metal part and the defect data on the surface of the metal part during historical polishing, record the polishing force of the polishing part on the polishing area and the evaluation score corresponding to the polishing force through the pressure sensor, and use the material, defect data, polishing force of the polishing area and the evaluation score corresponding to the polishing force of the metal part during historical polishing as training data of the neural network model. After the neural network model is trained, when the metal part passes under the first camera, the trained neural network model predicts the first cylinder and the second cylinder that need to be controlled accordingly and the required polishing force, and the control module sends the control signal to the corresponding first cylinder and the second cylinder.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] (1) This solution uses the first camera to obtain the three-dimensional data of the metal part and the information of the area to be polished. Combined with the material information entered through human-computer interaction and the precise quantitative analysis of surface defects, it can provide detailed and accurate input data for the polishing process. Based on these data, the trained neural network model is used to predict the polishing force. By controlling the cylinder stroke, the polished part can be accurately fitted to the polishing requirements of different metal parts. Whether it is a flat or uneven surface, fine polishing can be achieved, effectively improving the polishing quality and reducing over-polishing or under-polishing.

[0045] (2) During the grinding process, this solution uses a pressure sensor to capture the dynamic changes in grinding force in real time. After grinding a short distance, the second camera immediately detects the grinding effect. Once it is found that the ideal standard is not met, the cylinder parameters are immediately fine-tuned to increase the grinding force. This dynamic adjustment mechanism ensures that the entire grinding process moves towards the ideal effect and reduces the scrap rate. In addition, the data of each adjustment will be fed back to the neural network model to continuously optimize the model prediction accuracy, forming a virtuous circle and making subsequent grinding operations more accurate and efficient.

[0046] (3) This solution sets polishing effect evaluation indicators such as surface roughness and glossiness and converts them into quantitative scores ranging from 0 to 100, providing an intuitive and comparable measurement standard for the polishing effect. This not only makes it easier for operators to quickly judge the polishing quality, but also provides clear and unambiguous feedback for subsequent data-driven model training and optimization, helping to improve the stability and reliability of the entire polishing system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 This is an appearance view of the overall structure of the present invention;

[0049] Figure 2 It is a side view of the overall structure of the present invention;

[0050] Figure 3 It is a schematic diagram of the structure of the first prism of the present invention;

[0051] Figure 4 It is a structural schematic diagram of the second gear of the present invention;

[0052] Figure 5 is a schematic diagram of the distribution of the second gear and the third gear of the present invention;

[0053] Figure 6 It is a structural schematic diagram of the first cylinder of the present invention;

[0054] Figure 7 It is a structural schematic diagram of the grinding part of the present invention;

[0055] Figure 8 It is a structural schematic diagram of the connecting piece of the present invention;

[0056] Fig. 9 This is an appearance view of the metal part of the present invention;

[0057] Fig.10 It is a schematic diagram of grinding a metal part with a concave surface according to the present invention.

[0058] Description of the numbers in the figure:

[0059] 1. Processing table; 2. First motor; 3. Vacuum suction cup; 4. First cylinder; 5. Second camera; 6. Second cylinder; 7. Grinding piece; 71. Elastic airbag; 72. Grinding brush; 8. Second motor; 9. First gear; 10. Second gear; 11. First prism hole; 12. Third gear; 13. Second prism hole; 14. Connecting piece; 141. Lower rotating box; 142. Upper rotating box; 143. First blocking plate; 144. Second blocking plate; 15. First prism; 16. Second prism; 17. Support frame; 18. Horizontal plate; 19. First camera; 20. Screw; 21. Optical axis; 22. Moving plate; 23. Support block; 24. Pressure sensor; 25. Metal parts. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention; it is obvious that the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments, and all other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making creative work are within the scope of protection of the present invention.

[0061] See also Figures 1 to 10A surface grinding device for metal parts processing includes a processing table 1 and a vacuum suction cup 3, the vacuum suction cup 3 is arranged at the upper end of the processing table 1; a first motor 2, the first motor 2 is installed at a position close to one end of the upper surface of the processing table 1, and the first motor 2 is used to drive the vacuum suction cup 3 to move left and right on the processing table 1; a support frame 17, the support frame 17 is installed at the upper end of the processing table 1; a cross plate 18, the cross plate 18 is fixedly installed on the inner wall of the support frame 17; a first cylinder 4 and a second cylinder 6, the first cylinder 4 and the second cylinder 6 are arranged in multiple groups, and the multiple groups of the first cylinder 4 and the second cylinder 6 are all installed at the upper end of the support frame 17; grinding parts 7, the grinding pieces 7 are arranged in multiple groups, and the multiple groups of grinding pieces 7 are correspondingly arranged at the lower positions of the first cylinder 4 and the second cylinder 6; the second motor 8, the second motor 8 is installed at the upper end of the support frame 17, and the second motor 8 is used to drive the grinding pieces 7 to rotate; the first camera 19, the first camera 19 is installed at the lower end of the support frame 17, and the first camera 19 is used to obtain the characteristic data of the metal piece 25 passing below; the control module controls the corresponding first cylinder 4 and the second cylinder 6 to drive the grinding piece 7 to move to the corresponding height according to the characteristic data obtained by the first camera 19, and applies the corresponding extrusion force to the corresponding surface of the metal piece 25;

[0062] A first gear 9, the first gear 9 is mounted on the output shaft of the second motor 8; a second gear 10, the second gear 10 is rotatably mounted above the cross plate 18, the number of the second gears 10 corresponds to the number of the first cylinders 4, a plurality of groups of second gears 10 are meshed and connected with each other, and one group of the second gears 10 is meshed and connected with the first gear 9; a first prism hole 11, the first prism hole 11 is penetrated and opened at the center position of the second gear 10; a first prism 15, the first prism 15 is arranged inside the first prism hole 11, the lower end of the first prism 15 movably penetrates the cross plate 18, and is fixedly connected with the corresponding grinding piece 7; a connecting piece 14, the connecting piece 14 is mounted between the upper end of the first prism 15 and the first cylinder 4, and the connecting piece 14 is used to ensure that the first prism 15 rotates when the first cylinder 4 pushes the first prism 15 downward;

[0063] A third gear 12, the third gear 12 is arranged at the meshing connection of every two sets of second gears 10, and the third gear 12 is meshed and connected with the corresponding second gear 10; a second prism hole 13, the second prism hole 13 is penetrated and opened at the center position of the third gear 12; a second prism 16, the second prism 16 is arranged in the second prism hole 13, the lower end of the second prism 16 movably penetrates the cross plate 18 and is fixedly connected to the corresponding grinding piece 7, and the upper end of the second prism 16 is connected to the second cylinder 6 with a connecting piece 14 of the same structure;

[0064] A screw rod 20, which is fixedly mounted on the output shaft of the first motor 2; a movable plate 22, which is threadedly connected to the side surface of the screw rod 20, and a vacuum suction cup 3 is mounted on the movable plate 22; an optical axis 21, which is movably arranged at the front and rear ends of the movable plate 22; support blocks 23, which are arranged in multiple groups, and the support blocks 23 are fixedly mounted on the left and right ends of the optical axis 21, wherein one group of support blocks 23 is rotatably mounted on one end of the screw rod 20 away from the first motor 2, and the lower end of the support block 23 is fixedly connected to the upper surface of the processing table 1.

[0065] By adopting the above technical solution, when it is necessary to grind the processed metal part 25, first place the metal part 25 on the vacuum suction cup 3, turn on the vacuum suction cup 3, the vacuum suction cup 3 will suck the metal part 25 placed on it, and then the first motor 2 drives the screw rod 20 to rotate, and the screw rod 20 will drive the movable plate 22 to move to the right when rotating, and the movable plate 22 will also drive the vacuum suction cup 3 to move synchronously when moving synchronously to the right, then the metal part 25 will pass through the first camera 19 when moving to the right, the first camera 19 will shoot the metal part 25, and will obtain the three-dimensional data of the metal part 25, and at the same time obtain the area on the surface of the metal part 25 that needs to be polished, as the area to be polished on the surface of the metal part 25 passes and moves to the bottom of the polishing part 7, the control module controls the first cylinder 4 and the second cylinder 6 to drive the corresponding polishing part 7 to move downward, the first cylinder 4 controls the first prism 15 to move downward, and the second cylinder 6 controls the second prism 16 to move downward, so that the polishing part 7 contacts the corresponding area to be polished, The second motor 8 drives the first gear 9 to rotate all the time. When the first gear 9 rotates, it drives the corresponding second gear 10 to rotate. When the second gear 10 rotates, it drives the entire second gear 10 to rotate. When the second gear 10 rotates, it drives the first prism 15 to rotate. When the first prism 15 rotates, it drives the grinding piece 7 at the lower end to rotate; when the second gear 10 rotates, it drives the third gear 12 to rotate. When the third gear 12 rotates, it drives the second prism 16 to rotate. When the second prism 16 rotates, it drives the corresponding grinding piece 7 at the lower end to rotate. The surface of the metal part 25 can be polished by the rotation of the grinding piece 7. If the surface of the metal part 25 is uneven, when the metal part 25 passes under the grinding piece 7, the corresponding grinding piece 7 is controlled to move up and down by the first cylinder 4 and the second cylinder 6, so that the grinding piece 7 is close to the surface of the metal part 25, and the uneven metal part 25 can be polished.

[0066] In addition, there is a gap between the grinding pieces 7 at the lower ends of every two groups of first prisms 15, which will cause some areas to be ungrinded. The third gear 12 is set at the meshing connection between every two groups of second gears 10, so the grinding piece 7 at the lower end of the second prism 16 is set in the gap left between every two groups of grinding pieces 7 on the first prism 15 before, thereby avoiding the problem of remaining parts that are not polished when grinding the metal part 25.

[0067] In some embodiments of the present invention, the connecting member 14 includes a lower rotating box 141, which is sleeved on the upper ends of the first prism 15 and the second prism 16; a first blocking plate 143, which is fixedly mounted on the upper ends of the first prism 15 and the second prism 16, and the first blocking plate 143 is arranged in the lower rotating box 141; an upper rotating box 142, which is sleeved on the lower ends of the first prism 15 and the second prism 16, and the lower end of the upper rotating box 142 is fixedly connected to the lower rotating box 141; and a second blocking plate 144, which is fixedly mounted on the lower ends of the first prism 15 and the second prism 16, and the second blocking plate 144 is arranged in the upper rotating box 142.

[0068] By adopting the above technical solution, when the first prism 15 and the second prism 16 rotate, the corresponding first blocking plate 143 will be driven to rotate in the lower rotating box 141. The first blocking plate 143 can ensure that the first prism 15 and the second prism 16 fall from the lower rotating box 141. In this way, when the first cylinder 4 and the second cylinder 6 push the corresponding first prism 15 and the second prism 16 to move up and down, the first prism 15 and the second prism 16 can be guaranteed to rotate normally.

[0069] In some embodiments of the present invention, a pressure sensor 24 is installed at the lower end of the second blocking plate 144, and the pressure sensor 24 is electrically connected to the controller.

[0070] By adopting the above technical solution, when the first cylinder 4 and the second cylinder 6 push the corresponding first prism 15 and the second prism 16 to move downward, the second blocking plate 144 will move downward in the upper rotating box 142 and squeeze the pressure sensor 24. By monitoring the value of the pressure sensor 24, it can be known how much squeezing force the corresponding grinding part 7 applies to the surface of the metal part 25.

[0071] In some embodiments of the present invention, the material of the metal part 25 and the defect data on the surface of the metal part 25 are recorded, the grinding force of the grinding part 7 on the grinding area and the evaluation score corresponding to the grinding force are recorded through the pressure sensor 24, and the material, defect data, grinding force of the grinding area and the evaluation score corresponding to the grinding force of the metal part 25 are used as training data of the neural network model. After the neural network model is trained, when the metal part 25 passes under the first camera 19, the trained neural network model predicts the first cylinder 4 and the second cylinder 6 that need to be controlled and the required grinding force, and the control module sends the control signal to the corresponding first cylinder 4 and the second cylinder 6;

[0072] The second camera 5 is installed at the lower end of the horizontal plate 18 and away from one end of the first camera 19 . The second camera 5 is used to obtain feature data of the polished metal part 25 .

[0073] By adopting the above technical solution, when polishing the metal part 25, the material of the metal part 25, such as aluminum, copper, stainless steel, etc., can be selected from the preset material category list through the human-computer interaction interface, and the category information can be entered into the system in the form of text encoding as basic data for subsequent analysis. The defect data on the surface of the metal part 25 can be obtained by shooting with the first camera 19. The defect data includes scratch data, pit data, and oxide layer data, that is, polishing area data; for scratch data, the edge detection algorithm is used to accurately measure its length; for pit data, the pit depth is calculated in combination with the depth estimation model of deep learning; for oxide layer data, pixel statistics and area conversion are used. The algorithm is used to obtain the coverage area. These visual data are converted into numerical form for subsequent processing. The pressure value during the grinding process is collected in real time through the pressure sensor 24. This value directly reflects the grinding force to capture the dynamic changes of the grinding force. After the grinding of the metal part 25 is completed, as the vacuum suction cup 3 moves, the metal part 25 will pass under the second camera 5. The second camera 5 is used to capture the characteristic data of the metal part 25 after grinding, and the grinding effect of the metal part 25 is evaluated. The evaluation indicators can be set to surface roughness, glossiness, etc. The evaluation results can be converted into quantitative scores, for example, 0 to 100 points, 0 points for the worst, and 100 points for a perfect grinding effect. The specific process of converting the evaluation result into a quantitative score is to process and analyze the image captured by the second camera 5, extract the required feature data, analyze the texture and undulations of the surface of the metal part 25 through image processing technology, calculate the surface roughness value, set a roughness threshold range, and determine it according to actual needs. If the roughness is lower than the lower threshold, full marks are awarded; if the roughness is higher than the upper threshold, 0 points are awarded. If the roughness is within the threshold range, the score is assigned according to a linear or nonlinear relationship. The lower the roughness, the higher the score. According to the reflection in the image obtained by the second camera 5, the glossiness of the surface of the metal part 25 is evaluated. The higher the glossiness, the smoother the surface and the more uniform the reflection. The threshold range of glossiness is set. If the glossiness reaches the upper threshold or even exceeds the upper threshold, full marks are awarded; if the glossiness is lower than the lower threshold, 0 points are awarded. If the glossiness is within the threshold range, the score is assigned according to a linear relationship. The higher the glossiness, the higher the score. When it is necessary to comprehensively score the surface roughness and glossiness, different weights can be assigned to the surface roughness and glossiness, depending on their importance in a specific application scenario, and the quantitative scores of the surface roughness and glossiness can be calculated respectively. According to the assigned weights, the two scores are weighted averaged to obtain the final comprehensive score. It should be noted that within the shooting range of the second camera 5, it is necessary to ensure that the light is sufficient, uniform and stable.

[0074] Before the device is officially put into use, the neural network model needs to be fully trained. For this purpose, a large number of metal parts 25 are selected as training samples. For example, 500 metal parts 25 with different characteristics and defects can be selected. The material information, defect data, and the force used in the grinding process of these metal parts 25, as well as the evaluation scores corresponding to these grinding forces, will be systematically input into the neural network model for training. Before the training begins, some key parameters are set in advance, including the threshold range of roughness and the threshold range of glossiness. These threshold ranges are determined based on the expectations of the final product quality. According to these threshold ranges of roughness and glossiness, the corresponding evaluation score standards are set.

[0075] During the model training and application process, after analyzing the image captured by the second camera 5, the evaluation score obtained is lower than 80 points, which means that the current grinding intensity may not be sufficient to meet the set quality standards. In this case, the system will automatically adjust the grinding intensity the next time the metal part 25 with the same material and defect data is polished. Specifically, a weight for increasing the grinding intensity can be set according to the difference between the evaluation score and 80 points. This weight will determine the amplitude of adjusting the grinding intensity, thereby ensuring that the grinding process can more accurately meet the requirements of product quality. Through such a mechanism, the grinding process can be continuously optimized to improve the product qualification rate and overall quality. After the neural network model is trained, when the first camera 19 shoots a new metal part 25, the material and defect data are obtained and input into the trained neural network model, and the predicted grinding intensity value can be quickly output. The control module controls the corresponding telescopic stroke of the first cylinder 4 and the second cylinder 6 according to this value to make the grinding part 7 reach the preset grinding intensity.

[0076] Each time the metal part 25 completes a short section of the polishing path and passes under the second camera 5, if the second camera 5 captures that the polished metal part 25 does not meet the ideal polishing standard (such as the effect score is less than 80 points), the control module fine-tunes the cylinder parameters in real time, such as increasing the preset polishing force by 5%, to increase the polishing force until the effect meets the standard, and inputs the above data into the neural network model again. By continuously adjusting the neural network model in real time, the prediction accuracy of the neural network model is improved. The neural network model of the present invention can adopt a fully connected neural network model or a convolutional neural network model.

[0077] In some embodiments of the present invention, the polishing piece 7 is composed of an elastic airbag 71 and polishing bristles 72, the polishing bristles 72 are evenly distributed on the outer surface of the elastic airbag 71, and the lower ends of the first prism 15 and the second prism 16 are fixedly connected to the corresponding elastic airbag 71.

[0078] By adopting the above technical solution, when encountering a concave cavity on the side of the metal part 25, the elastic airbag 71 is deformed and enters the concave cavity by increasing the grinding force. When the elastic airbag 71 rotates, it drives the grinding bristles 72 to rotate. The rotating grinding bristles 72 grind the concave cavity on the side of the metal part 25, thereby improving the adaptability of the grinding device of the present invention.

[0079] The present invention also provides a grinding method, which is applicable to the above-mentioned surface grinding device for metal parts processing, and comprises the following steps:

[0080] Step 1, place the metal part 25 on the vacuum suction cup 3, turn on the vacuum suction cup 3, and then drive the lead screw 20 to rotate through the first motor 2 to drive the vacuum suction cup 3 to move from left to right;

[0081] Step 2, when the metal part 25 moves with the vacuum suction cup 3 to the bottom of the first camera 19, the first camera 19 takes a picture of the passing metal part 25 and obtains feature data of the metal part 25;

[0082] In step 3, the first camera 19 controls the grinding piece 7 to adjust to a corresponding height according to the acquired feature data. The neural network model outputs the required grinding force of the grinding piece 7 according to the acquired feature data. The second motor 8 drives the grinding piece 7 to rotate and performs a grinding operation on the passing metal piece 25.

[0083] The above are only preferred specific implementations of the present invention; however, the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solutions and improved concepts of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A surface grinding device for metal parts processing, characterized in that: include: A processing table (1) and a vacuum suction cup (3), wherein the vacuum suction cup (3) is arranged at the upper end of the processing table (1); a first motor (2), the first motor (2) being mounted on the upper surface of the processing table (1) at a position close to one end, the first motor (2) being used to drive the vacuum suction cup (3) to move left and right on the processing table (1); A support frame (17), wherein the support frame (17) is mounted on the upper end of the processing table (1); A transverse plate (18), wherein the transverse plate (18) is fixedly mounted on the inner wall of the support frame (17); A first cylinder (4) and a second cylinder (6), wherein the first cylinder (4) and the second cylinder (6) are provided in multiple groups, and the multiple groups of the first cylinder (4) and the second cylinder (6) are all mounted on the upper end of the support frame (17); Grinding pieces (7), wherein the grinding pieces (7) are provided in multiple groups, and the multiple groups of grinding pieces (7) are correspondingly provided at positions below the first cylinder (4) and the second cylinder (6); a second motor (8), the second motor (8) being mounted on the upper end of the support frame (17), the second motor (8) being used to drive the grinding piece (7) to rotate; A first camera (19), the first camera (19) being mounted at the lower end of the support frame (17), the first camera (19) being used to acquire characteristic data of a metal part (25) passing below; A control module controls the corresponding first cylinder (4) and second cylinder (6) to drive the grinding member (7) to move to a corresponding height and to apply a corresponding extrusion force to a corresponding surface of the metal member (25) according to the characteristic data acquired by the first camera (19); A first gear (9), the first gear (9) being mounted on an output shaft of the second motor (8); a second gear (10), the second gear (10) being rotatably mounted above the horizontal plate (18), the number of the second gears (10) corresponding to the number of the first cylinders (4), a plurality of groups of the second gears (10) being meshedly connected to each other, wherein one group of the second gears (10) is meshedly connected to the first gear (9); a first prism hole (11), the first prism hole (11) being formed through and at a central position of the second gear (10); A first prism (15), wherein the first prism (15) is arranged inside the first prism hole (11), and the lower end of the first prism (15) movably penetrates the transverse plate (18) and is fixedly connected to the corresponding polishing piece (7); a connecting member (14), the connecting member (14) being installed between the upper end of the first prism (15) and the first cylinder (4), the connecting member (14) being used to ensure that the first prism (15) rotates when the first cylinder (4) pushes the first prism (15) downward; A third gear (12), the third gear (12) being arranged at a meshing connection between each two sets of second gears (10), the third gear (12) being meshingly connected with the corresponding second gear (10); a second prism hole (13), the second prism hole (13) being formed through and at a central position of the third gear (12); A second prism (16), wherein the second prism (16) is arranged in the second prism hole (13), wherein the lower end of the second prism (16) movably passes through the transverse plate (18) and is fixedly connected to the corresponding grinding piece (7), and a connecting piece (14) of the same structure is connected between the upper end of the second prism (16) and the second cylinder (6), and the connecting piece (14) comprises: A lower rotating box (141), wherein the lower rotating box (141) is sleeved on the upper ends of the first prism (15) and the second prism (16); a first blocking plate (143), the first blocking plate (143) being fixedly mounted on the upper ends of the first prism (15) and the second prism (16), the first blocking plate (143) being arranged in the lower rotating box (141); An upper rotating box (142), the upper rotating box (142) being sleeved on the lower ends of the first prism (15) and the second prism (16), the lower end of the upper rotating box (142) being fixedly connected to the lower rotating box (141); a second blocking plate (144), the second blocking plate (144) being fixedly mounted on the lower ends of the first prism (15) and the second prism (16), the second blocking plate (144) being arranged in the upper rotating box (142); a pressure sensor (24), the pressure sensor (24) being mounted at the lower end of the second blocking plate (144), the pressure sensor (24) being electrically connected to the controller, the polishing member (7) being composed of an elastic airbag (71) and polishing bristles (72), the polishing bristles (72) being evenly distributed on the outer surface of the elastic airbag (71), and the lower ends of the first prism (15) and the second prism (16) being fixedly connected to the corresponding elastic airbag (71); A second camera (5), the second camera (5) being mounted at the lower end of the horizontal plate (18) and away from one end of the first camera (19), the second camera (5) being used to obtain characteristic data of the metal part (25) after grinding, and when the metal part (25) completes a small section of the grinding path and passes under the second camera (5), if the second camera (5) finds that the metal part (25) after grinding does not meet the ideal grinding standard, the control module fine-tunes the cylinder parameters in real time.

2. The surface grinding device for metal parts processing according to claim 1, characterized in that: include: A screw rod (20), the screw rod (20) being fixedly mounted on an output shaft of the first motor (2); A movable plate (22), wherein the movable plate (22) is threadedly connected to a side surface of the screw rod (20), and the vacuum suction cup (3) is mounted on the movable plate (22); An optical axis (21), the optical axis (21) being movably arranged to penetrate the front and rear ends of the movable plate (22); Support blocks (23), the support blocks (23) being arranged in a plurality of groups, the support blocks (23) being fixedly mounted at the left and right ends of the optical axis (21), one group of the support blocks (23) being rotatably mounted at an end of the screw rod (20) away from the first motor (2), and the lower end of the support block (23) being fixedly connected to the upper surface of the processing table (1).

3. A grinding method, applicable to the surface grinding device for metal parts processing according to claim 2, characterized in that: The steps include: Step 1, placing the metal part (25) on the vacuum suction cup (3), turning on the vacuum suction cup (3), and then driving the lead screw (20) to rotate by the first motor (2) to drive the vacuum suction cup (3) to move from left to right; Step 2, when the metal part (25) moves with the vacuum suction cup (3) to the bottom of the first camera (19), the first camera (19) photographs the passing metal part (25) and obtains characteristic data of the metal part (25), wherein the characteristic data includes defect data and polishing area data on the surface of the metal part (25); Step 3: The first camera (19) controls the grinding piece (7) to adjust to a corresponding height based on the acquired feature data; the neural network model outputs the grinding force required for the grinding piece (7) based on the acquired feature data; the second motor (8) drives the grinding piece (7) to rotate, and performs a grinding operation on the passing metal piece (25).

4. The grinding method according to claim 3, characterized in that: According to the feature data acquired by the first camera (19), the grinding piece (7) is controlled to be adjusted to a corresponding height, and the neural network model outputs the grinding force required for the grinding piece (7) according to the acquired feature data, including: Step 31, recording the material of the metal part (25) and the defect data on the surface of the metal part (25) during historical polishing, recording the polishing force of the polishing part (7) on the polishing area and the evaluation score corresponding to the polishing force through the pressure sensor (24), and using the material of the metal part (25) during historical polishing, the defect data, the polishing force of the polishing area and the evaluation score corresponding to the polishing force as training data for the neural network model. After the neural network model is trained, when the metal part (25) passes under the first camera (19), the trained neural network model predicts the first cylinder (4) and the second cylinder (6) to be controlled and the required polishing force, and the control module sends the control signal to the corresponding first cylinder (4) and the second cylinder (6).

Images