A control method of a robot-based foundry cleaning intelligent equipment

Abstract

The application discloses a control method of casting cleaning intelligent equipment based on a robot, which comprises the following steps: automatically loading and clamping a casting, self-inspecting and starting a program, visually recognizing and planning a path, performing a plasma gas planing cutting process, performing a visual recognition verification process, performing a round corner cleaning process and performing a polishing process. A 3D visual camera, a plasma gas planer, a telescopic file and a strong grinder are respectively grabbed by a quick-change disc to realize visual measurement, gas planing cutting, round corner cleaning and polishing processes, so that the integrated degree is high, the robot is used to replace manual work to clean and polish the casting, the harm to the body of a worker in the cleaning process is reduced, and the requirements of intelligence, safety and reliability of the casting cleaning are met.

Description

Technical Field

[0001] This invention relates to the field of casting cleaning technology, and in particular to a control method for a robot-based intelligent equipment for casting cleaning. Background Technology

[0002] With the improvement of industrial level, people have higher and higher requirements for the manufacturing level and production efficiency of casting products. In the field of rail transit, the surface of coupler castings is generally rough after casting and there is a lot of casting residue. Therefore, it is necessary to clean and polish the surface of the castings.

[0003] In existing technologies, manual methods are generally used to clean and polish the surface of castings. However, the surface of castings contains a large amount of dust, and a large amount of dust is generated during the cutting and polishing process, which can cause great harm to workers. In addition, manual cleaning and polishing is inefficient. Therefore, there is an urgent need to develop a control system that uses robots to replace manual labor for the cleaning of castings. Summary of the Invention

[0004] To address the aforementioned problems, this invention proposes a control method for a robot-based intelligent equipment for casting cleaning.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] A control method for a robot-based intelligent foundry cleaning equipment includes the following steps:

[0007] Step 1: Open the sliding door of the safety room and place the casting workpiece to be cleaned onto the displacement fixture using the automatic feeding device. The intelligent equipment control system for casting cleaning controls the hydraulic press to drive the displacement fixture to press the casting workpiece to be cleaned.

[0008] Step 2: Close the sliding door of the safety room, start the preparation command of the intelligent equipment for casting cleaning, and the control system of the intelligent equipment for casting cleaning will perform a program self-check. After the program self-check is completed, the operation will be indicated on the control panel to allow the work and start the casting cleaning operation.

[0009] Step 3: The host computer controls a six-axis industrial robot to grab a 3D vision camera via a quick-change tray. The 3D vision camera scans and identifies the casting workpiece to be cleaned, acquiring image information of the casting workpiece and sending the image information to the host computer. The host computer processes the image information of the casting workpiece to obtain point cloud data, compares the overlap between the point cloud data and the 3D digital model of the casting workpiece sample, obtains the feature point positions of the parts to be cleaned, obtains the offset between the feature point positions and the reference path, and plans the cleaning trajectory based on the offset between the feature point positions and the reference path to obtain the cleaning trajectory planning path.

[0010] Step 4: The host computer controls a six-axis industrial robot to grab the plasma air planer through a quick-change disc and perform preliminary air planing cuts on the parts of the casting workpiece to be cleaned according to the planned path of the cleaning trajectory.

[0011] Step 5: After the initial air gouging is completed, the host computer controls the six-axis robot to place the plasma air gouging machine back onto the tool holder via the quick-change disc and re-grab the 3D vision camera to scan and identify the part of the casting workpiece to be cleaned after the initial air gouging. The robot obtains the point cloud data of the part of the casting workpiece to be cleaned after the initial air gouging, and obtains the feature point positions of the part of the casting workpiece to be cleaned after the initial air gouging based on the feature point positions. The robot also obtains the residual amount of the part to be cleaned after the initial air gouging based on the feature point positions. The robot determines whether the residual amount of the part to be cleaned after the initial air gouging is within the first set threshold range. If not, the cleaning trajectory is re-planned and the host computer controls the six-axis robot to grab the plasma air gouging machine for a second air gouging. If yes, proceed to step 6.

[0012] Step 6: The host computer controls the six-axis industrial robot to switch tools through the quick-change disc. The six-axis industrial robot grabs the telescopic file to clean the rounded corners of the casting workpiece to be cleaned. After the rounded corners are cleaned, proceed to step 7.

[0013] Step 7: The host computer controls the six-axis industrial robot to switch tools via a quick-change disc. The six-axis industrial robot grabs the high-power grinder to grind the residual material after air gouging and completes the cleaning of the casting workpiece to be cleaned.

[0014] Furthermore, the specific process of obtaining the cleanup trajectory planning path in step 3 based on the offset between the feature point position and the reference path is as follows:

[0015] 1) Obtain the highest point of the part of the casting workpiece to be cleaned as the starting point of the planned path. The coordinates of the highest point are (Px, Py, Pz).

[0016] 2) Calculate the deviation (Δx) between the highest point and the three-dimensional digital model of the cast workpiece sample. p Δy p Δz p );

[0017] 3) Use formula (1) to calibrate the coordinate system, transform the workpiece coordinate system into the tool coordinate system, and obtain the horizontal cleaning planning path:

[0018]

[0019] Where (Rx, Ry, Rz) is the starting target position of the robot's tool path;

[0020] 4) Obtain the coordinates Ze of the highest point of the area to be cleaned along the Z-axis and the coordinates Zt of the target position after cleaning, and calculate the amount e to be removed along the Z-axis. z =Ze-Zt;

[0021] If the amount to be removed in the Z-axis direction is e z When the value is ≥5, the Z-axis cleaning path is planned when fz = 2mm / s;

[0022] If the amount to be removed in the Z-axis direction is e z When <5, the Z-axis cleaning path is planned using the PID algorithm formula (2):

[0023] fz=Kp*e z (t)+Ki*∑e z (t)+K d *(e z (t)-e z (t-1)) (2)

[0024] Where fz is the feed amount in the Z-axis direction; t is the contact time between the cleaning tool and the workpiece for each feed. The time t depends on the form of the part to be cleaned. For large and thick gates and risers, t is generally selected as 2ms. For ordinary flash or burrs, t is generally selected as 1ms; Kp is the scaling factor, which is generally selected as 0.5 to 2; Ki is the integral factor, which is generally selected as 0.2 to 3; and kd is the differential factor, which is generally selected as 0 to 2.

[0025] Furthermore, during the initial horizontal path planning for air gouging, the deviation between the highest point and the three-dimensional digital model of the cast workpiece sample is set as (Δx). p Δy p Δz p -L1);

[0026] The coordinate system is calibrated using formula (3), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the preliminary air gouging cutting horizontal cleaning planning path:

[0027]

[0028] Where (Rx, Ry, Rz) are the starting target positions of the robot's tool path, and L1 is the first set threshold.

[0029] Furthermore, during the horizontal path planning process for secondary air gouging, the deviation between the highest point and the three-dimensional digital model of the casting workpiece sample is set as (Δx). p Δy p Δz p -L2);

[0030] The coordinate system is calibrated using formula (4), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the secondary air gouging cutting horizontal cleaning planning path:

[0031]

[0032] Where (Rx, Ry, Rz) is the starting target position of the robot's tool path, and L2 is the residual amount of the part to be cleaned after the secondary air gouging.

[0033] Furthermore, during the horizontal path planning process for high-intensity grinding, the deviation between the highest point and the three-dimensional digital model of the cast workpiece sample is set as (Δx). p Δy p Δz p -L3);

[0034] The coordinate system is calibrated using formula (5), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the secondary air gouging cutting horizontal cleaning planning path:

[0035]

[0036] Where (Rx, Ry, Rz) are the starting target positions of the robot's tool path, and L3 is the residual amount of the part to be cleaned after grinding by the high-intensity grinder.

[0037] Furthermore, during the cleaning process of the six-axis industrial robot gripping and cleaning tool on the part of the casting workpiece to be cleaned, the force control device collects force signal data during the cleaning process through a force sensor installed between the quick-change plate and the end effector of the six-axis industrial robot, and transmits the force signal data to the host computer via Ethernet communication. The host computer is used to determine whether the force on the end effector of the six-axis robot is within a second set threshold range based on the force signal data. If not, the host computer sends a stop command to the intelligent equipment control system for casting cleaning.

[0038] Furthermore, the force control device also includes collecting force signal data when the cleaning tool and the part to be cleaned just come into contact through a force sensor, and transmitting the force signal data when the cleaning tool and the part to be clean just come into contact to a host computer. The host computer is used to determine whether the force applied when the cleaning tool and the part to be clean just come into contact is less than a third preset threshold based on the force signal data when the cleaning tool and the part to be clean just come into contact. If so, a command to replace the cleaning tool is issued.

[0039] Furthermore, during steps 3 to 7, the intelligent equipment control system for casting cleaning detects whether the safety door of the intelligent equipment for casting cleaning is opened. If so, the intelligent equipment control system for casting cleaning cuts off the power supply to the equipment inside the opened safety door.

[0040] Furthermore, step 3 also includes the host computer inputting the image information of the casting workpiece to be cleaned into a convolutional neural network for deep learning to obtain a sample model of the casting workpiece, and storing the sample model of the casting workpiece in a sample library; the host computer compares the overlap of the point cloud data of the casting workpiece to be cleaned with all the sample models of casting workpieces in the sample library to obtain the feature point positions of the parts to be cleaned in the casting workpiece to be cleaned compared with the casting workpiece sample model with the largest overlap, as well as the offset of the feature point positions from the reference path, and performs cleaning trajectory planning based on the offset of the feature point positions from the reference path to obtain the cleaning trajectory planning path.

[0041] Beneficial effects: The present invention discloses a control method for intelligent casting cleaning equipment based on robots. By using a quick-change disc to grasp different tools, it realizes processing steps such as visual measurement, path planning, cutting and grinding. The control system monitors, logically controls and protects the entire working process, meeting the requirements of intelligent, safe and reliable future casting cleaning production lines. It has a high degree of integration and realizes the use of robots to replace manual labor, reducing the harm to workers' bodies during the cleaning process, improving processing efficiency and enhancing processing accuracy. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a control method for a robot-based intelligent equipment for casting cleaning, as disclosed in an embodiment of the present invention.

[0044] Figure 2 This is a structural diagram of a robot-based intelligent equipment for casting cleaning disclosed in an embodiment of the present invention.

[0045] In the picture: 1. Six-axis industrial robot, 2. Quick-change disc, 3. Force control equipment, 4. Plasma air planer, 5. High-intensity grinder, 6. Telescopic file machine, 7. Dust collector, 8. 3D vision camera, 9. Host computer and control system, 10. Safety room, 11. Sliding door, 12. Safety protection door. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] like Figure 2 The diagram shows the robot-based intelligent equipment for casting cleaning in this invention, including a six-axis industrial robot 1, a quick-change disc 2, a force control device 3, a plasma air planer 4, a high-intensity grinder 5, a telescopic file machine 6, a dust collector 7, a 3D vision camera 8, a host computer and control system 9, and a safety room 10. The quick-change disc is installed at the end of the industrial robot, and the force control device is installed between the end of the industrial robot and the quick-change disc. The plasma air planer, high-intensity grinder, and telescopic file machine are installed on the tool quick-change bracket as tools for different processes. The vision camera can be docked with the quick-change disc, and the industrial robot can grasp the vision camera to perform 3D visual recognition of the casting workpiece. The host computer is connected to the industrial robot, control system, vision camera, and digital amplifier via Ethernet communication through a switch. The control system includes PLC, safety relays, circuit breakers, contactors, and servo drivers. The safety room includes sheet metal frame, sliding doors 11, safety doors 12, a light curtain, and safety door locks.

[0048] This embodiment provides a control method for robot-based intelligent equipment for casting cleaning, such as... Figure 1 As shown, it includes the following steps:

[0049] Step 1: Open the sliding door of the safety room and place the casting workpiece to be cleaned onto the displacement fixture using the automatic feeding device. The intelligent equipment control system for casting cleaning controls the hydraulic press to drive the displacement fixture to press the casting workpiece to be cleaned.

[0050] Specifically, the intelligent equipment for casting cleaning by the robot includes a safety chamber with a sliding door. When cleaning of the casting workpiece is required, the sliding door of the safety chamber is opened, and the casting workpiece to be cleaned is placed on the displacement fixture through an automatic feeding device. The intelligent equipment control system controls the hydraulic press to drive the displacement fixture to clamp the casting workpiece to be cleaned. In this application, the control system PLC outputs instructions to control the hydraulic press to drive the fixture to clamp the casting. At the same time, the PLC can also output instructions to the servo driver to control the X-axis and Y-axis motors of the displacement machine (displacement fixture) to drive the casting to rotate, so as to achieve cleaning of different surfaces of the hook with the help of the robot.

[0051] Step 2: Close the sliding door of the safety room, initiate the preparation command for the intelligent casting cleaning equipment, and the intelligent casting cleaning equipment control system will perform a program self-check. After the program self-check is completed, a work permission indication will be displayed on the control panel, and the casting cleaning work will be started. The program self-check ensures that all signals and safety protections are normal.

[0052] Step 3: The host computer controls a six-axis industrial robot to grab a 3D vision camera via a quick-change tray. The 3D vision camera scans and identifies the casting workpiece to be cleaned, acquiring image information of the casting workpiece and sending the image information to the host computer. The host computer processes the image information of the casting workpiece to obtain point cloud data, compares the overlap between the point cloud data and the 3D digital model of the casting workpiece sample, obtains the feature point positions of the parts to be cleaned, obtains the offset between the feature point positions and the reference path, and plans the cleaning trajectory based on the offset between the feature point positions and the reference path to obtain the cleaning trajectory planning path.

[0053] In a specific embodiment, the specific process of obtaining the cleanup trajectory planning path in step 3 based on the offset between the feature point position and the reference path is as follows:

[0054] 1) Obtain the highest point of the part of the casting workpiece to be cleaned as the starting point of the planned path. The coordinates of the highest point are (Px, Py, Pz).

[0055] 2) Calculate the deviation (Δx) between the highest point and the three-dimensional digital model of the cast workpiece sample. p Δy p Δz p );

[0056] 3) Use formula (1) to calibrate the coordinate system, transform the workpiece coordinate system into the tool coordinate system, and obtain the horizontal cleaning planning path:

[0057]

[0058] Where (Rx, Ry, Rz) is the starting target position of the robot's tool path;

[0059] 4) Obtain the coordinates Ze of the highest point of the area to be cleaned along the Z-axis and the coordinates Zt of the target position after cleaning, and calculate the amount e to be removed along the Z-axis. z =Ze-Zt;

[0060] If the amount to be removed in the Z-axis direction is e z When the value is ≥5, the Z-axis cleaning path is planned when fz = 2mm / s;

[0061] If the amount to be removed in the Z-axis direction is ez When <5, the Z-axis cleaning path is planned using the PID algorithm formula (2):

[0062] fz=Kp*e z (t)+Ki*∑e z (t)+K d *(e z (t)-e z (t-1)) (2)

[0063] Where fz is the feed amount in the Z-axis direction; t is the contact time between the cleaning tool and the workpiece for each feed. The time t depends on the form of the part to be cleaned. For large and thick gates and risers, t is generally selected as 2ms. For ordinary flash or burrs, t is generally selected as 1ms; Kp is the scaling factor, which is generally selected as 0.5 to 2; Ki is the integral factor, which is generally selected as 0.2 to 3; and kd is the differential factor, which is generally selected as 0 to 2.

[0064] Step 4: The host computer controls a six-axis industrial robot to grab the plasma air planer through a quick-change disc and perform preliminary air planing cuts on the part of the casting workpiece to be cleaned according to the planned path of the cleaning trajectory. Specifically, after the preliminary cut, the residual amount of the part to be cleaned is controlled to be about 3mm.

[0065] Step 5: After the initial air gouging is completed, the host computer controls the six-axis robot to place the plasma air gouging machine back onto the tool holder via a quick-change disc and re-grab the 3D vision camera to scan and identify the area to be cleaned on the casting workpiece after the initial air gouging. The robot obtains point cloud data of the area after the initial air gouging and uses this point cloud data to determine the location of feature points on the area after the initial air gouging. Based on these feature point locations, the robot also determines the residual amount of the area after the initial air gouging. It then determines whether the residual amount is within a first set threshold range. If not, the cleaning trajectory is replanned, and the host computer controls the six-axis robot to grab the plasma air gouging machine for a second air gouging. If yes, step 6 is executed. After the second air gouging, the residual amount of the area to be cleaned is controlled to be approximately 1mm.

[0066] Step 6: The host computer controls the six-axis industrial robot to switch tools via the quick-change disc. The six-axis industrial robot grabs the telescopic file to clean the rounded corners of the casting workpiece. After the rounded corners are cleaned, step 7 is executed. The six-axis industrial robot grabs the telescopic file to clean the rounded corners of the casting workpiece, which can ensure the curvature requirements of the rounded corners and the smooth surface.

[0067] Step 7: The host computer controls the six-axis industrial robot to switch tools through the quick-change disc. The six-axis industrial robot grabs the high-power grinder to grind the residual material after air gouging and completes the cleaning of the casting workpiece. After grinding, the residual part is controlled within 0.3mm, thus completing the intelligent cleaning of the casting workpiece.

[0068] In this embodiment, different tools are grasped by a quick-change disc to realize processing steps such as visual measurement, path planning, cutting and grinding. The integration is high, and the robot can replace manual cleaning and grinding of the casting surface, reducing the harm to the workers' bodies during the cleaning process, improving processing efficiency and enhancing processing accuracy.

[0069] Furthermore, during the initial horizontal path planning for air gouging, the deviation between the highest point and the three-dimensional digital model of the cast workpiece sample is set as (Δx). p Δy p Δz p -L1);

[0070] The coordinate system is calibrated using formula (3), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the preliminary air gouging cutting horizontal cleaning planning path:

[0071]

[0072] Where (Rx, Ry, Rz) is the starting target position of the robot's tool path, and L1 is the first set threshold, which is generally taken as L1 = 3.

[0073] In a specific embodiment, during the horizontal path planning process for secondary air gouging, the deviation between the highest point and the three-dimensional digital model of the casting workpiece sample is set as (Δx). p Δy p Δz p -L2);

[0074] The coordinate system is calibrated using formula (4), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the secondary air gouging cutting horizontal cleaning planning path:

[0075]

[0076] Where (Rx, Ry, Rz) is the starting target position of the robot's tool path, and L2 is the residual amount of the part to be cleaned after the secondary air gouging cut, which is generally taken as L2 = 1.

[0077] In a specific embodiment, during the horizontal path planning process of the high-intensity grinding machine, the deviation between the highest point and the three-dimensional digital model of the cast workpiece sample is set as (Δx). p Δy p Δz p -L3);

[0078] The coordinate system is calibrated using formula (5), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the secondary air gouging cutting horizontal cleaning planning path:

[0079]

[0080] Where (Rx, Ry, Rz) is the starting target position of the robot's tool path, and L3 is the residual amount of the part to be cleaned after grinding by the high-intensity grinder, which is generally taken as L3 = 0.3.

[0081] In this application, during the path planning process, L1 is set as the first set threshold, L2 is the residual amount of the part to be cleaned after secondary air gouging, and L3 is the residual amount of the part to be cleaned after high-intensity grinding. This is to better utilize the processing capabilities and effects of the air gouging and grinding tools and improve the processing efficiency of the entire casting residue.

[0082] Furthermore, during the cleaning process of the six-axis industrial robot gripping and cleaning the part of the casting workpiece to be cleaned, a force control device collects force signal data during the cleaning process through a force sensor installed between the quick-change disc and the end effector of the six-axis industrial robot. This force signal data is then transmitted to a host computer via Ethernet communication. The host computer uses the force signal data to determine whether the force on the end effector of the six-axis robot is within a second preset threshold range. If not, the host computer sends a stop command to the intelligent equipment control system for casting cleaning. This step is used to protect the six-axis industrial robot from overload. By setting up a force control device, the force between the end effector of the six-axis robot and the workpiece to be cleaned can be obtained based on the force signal collected by the force control device. This allows the control system to monitor the entire cleaning process and prevent excessive force between the cleaning tool and the workpiece from damaging the tool or the robot.

[0083] Furthermore, the force control device also includes collecting force signal data when the cleaning tool first contacts the part to be cleaned via a force sensor, and transmitting this force signal data to a host computer. The host computer uses this force signal data to determine whether the force applied when the cleaning tool first contacts the part to be cleaned is less than a third preset threshold. If so, it issues a command to replace the cleaning tool. This step is used for signal feedback upon completion of each cleaning process. By collecting the contact force between the tool and the workpiece through the force sensor of the force control device, if, for example, the contact force of the tool decreases by 40% when cleaning begins at the same starting point, the tool is considered to have reached the end of its service life and needs to be replaced. At the same time, a prompt indicating that the tool needs to be replaced is displayed on the touch screen of the control system, so as to prevent excessive wear of the cleaning tool from affecting the overall cleaning time and cleaning effect, and to replace it in advance to avoid reducing cleaning efficiency or damaging castings. For example, the lifespan of the grinding wheel of a high-intensity grinding machine can be predicted based on its usage time, and a prompt indicating that the tool needs to be replaced can be displayed on the touch screen of the control system, so that the staff can know the tool's lifespan and status in advance and replace it in advance to avoid reducing cleaning efficiency or damaging the coupler castings.

[0084] In a specific embodiment, during steps 3 to 7, the intelligent equipment control system for casting cleaning detects whether the safety door of the intelligent equipment for casting cleaning is opened. If so, the intelligent equipment control system for casting cleaning cuts off the power supply to the equipment inside the opened safety door. Specifically, the safety door mainly includes a sliding door, a safety protection door, a light curtain, and a safety door lock. When the operator enters the corresponding workplace, the corresponding safety door is opened. At this time, the power supply to the equipment inside the safety door is cut off, which plays a role in protecting the safety of the operator.

[0085] In a specific embodiment, a dust removal device is also included. When the system starts working, the dust removal device is turned on, and the high-power dust removal device sucks away the smoke and dust generated during the cleaning process, maintaining the cleanliness of the working environment inside the safe room.

[0086] Specifically, the host computer interacts with the robot controller, PLC, force-controlled digital amplifier, and 3D vision camera via Ethernet communication. The host computer communicates with the robot controller, force-controlled digital amplifier, and 3D vision camera using the TCP / IP protocol, while the host computer communicates with the PLC using the Modbus-Tcp protocol.

[0087] In this embodiment, the control system PLC is responsible for controlling the power supply of the entire robot's casting cleaning intelligent equipment, and controlling the actions of safety relays, circuit breakers, contactors and servo drives, as well as the signal input and output of the entire system. It is also responsible for communicating with the host computer, robot controller and force control digital amplifier. In addition, the control system is responsible for monitoring the service life of the tools and displaying prompts for tool replacement on the control system touch screen, so that the staff can know the tool's life and status in advance and replace it in time to avoid reducing cleaning efficiency or damaging the castings.

[0088] Furthermore, step 3 also includes the host computer inputting the image information of the casting workpiece to be cleaned into a convolutional neural network for deep learning to obtain a sample model of the casting workpiece, and storing the sample model of the casting workpiece in a sample library; the host computer comparing the point cloud data of the casting workpiece to be cleaned with all casting workpiece samples in the sample library to obtain the feature point positions of the parts to be cleaned in the comparison between the casting workpiece to be cleaned and the casting workpiece sample with the largest overlap, as well as the offset of the feature point positions from the reference path, and performing cleaning trajectory planning based on the offset of the feature point positions from the reference path to obtain the cleaning trajectory planning path.

[0089] Specifically, a convolutional neural network is also installed in the host computer. After the 3D vision camera acquires image information of the casting to be cleaned and sends it to the host computer, the host computer inputs the image information of the casting to be cleaned into the convolutional neural network for deep learning. That is, the convolutional layer in the deep learning algorithm of the convolutional neural network extracts features from the images of different samples of the coupler casting acquired by the 3D vision camera to form a feature description matrix. The features mainly refer to the size and height of the gating riser and flash burrs. Then, the pooling layer is used to reduce the dimensionality and process overfitting of the feature description matrix. After processing through the fully connected layer, the deep learning sample model is output. A large number of output sample models are classified and detected to form a sample library. For each sample model in the sample library, a corresponding cleaning planning path is formed. The process involves several steps. First, when cleaning the next casting workpiece, the host computer compares the point cloud data of the next workpiece with the sample models of each casting workpiece in the sample library to determine the overlap. It then obtains the casting workpiece sample with the highest overlap and compares its point cloud data with this sample to obtain the feature point positions of the part to be cleaned (the redundant part between the part to be cleaned and the sample with the highest overlap). The offset between these feature point positions is calculated and superimposed on the cleaning planning path of the sample with the highest overlap to obtain the cleaning planning path for the part to be cleaned. Based on this cleaning planning path, the six-axis industrial robot can grasp the corresponding cleaning tools for rapid cleaning. In other words, this application, for newly arrived coupler castings, after acquiring visual images, can quickly match the closest sample model and use the cleaning trajectory of this sample model as the basis for the planning path of the casting, adding a small offset to form the actual cleaning trajectory path of the casting. This improves the speed and accuracy of the cleaning planning path, thereby improving the efficiency and quality of the cleaning.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control method for a robot-based intelligent equipment for casting cleaning, characterized in that: Includes the following steps: Step 1: Open the sliding door of the safety room and place the casting workpiece to be cleaned onto the displacement fixture using the automatic feeding device. The intelligent equipment control system for casting cleaning controls the hydraulic press to drive the displacement fixture to press the casting workpiece to be cleaned. Step 2: Close the sliding door of the safety room, start the preparation command of the intelligent equipment for casting cleaning, and the control system of the intelligent equipment for casting cleaning will perform a program self-check. After the program self-check is completed, the operation will be indicated on the control panel to allow the work and start the casting cleaning operation. Step 3: The host computer controls a six-axis industrial robot to grab a 3D vision camera via a quick-change tray. The 3D vision camera scans and identifies the casting workpiece to be cleaned, acquiring image information of the casting workpiece and sending the image information to the host computer. The host computer processes the image information of the casting workpiece to be cleaned into point cloud data, and compares the overlap between the point cloud data and the 3D digital model of the casting workpiece sample to obtain the feature point positions of the parts to be cleaned. The offset between the feature point positions and the reference path is obtained, and the cleaning trajectory is planned based on the offset between the feature point positions and the reference path to obtain the cleaning trajectory planning path. Step 4: The host computer controls a six-axis industrial robot to grab the plasma air planer through a quick-change disc and perform preliminary air planing cuts on the part of the casting workpiece to be cleaned according to the planned path of the cleaning trajectory. Step 5: After the initial air gouging is completed, the host computer controls the six-axis robot to place the plasma air gouging machine back onto the tool holder via the quick-change disc and re-grab the 3D vision camera to scan and identify the part of the casting workpiece to be cleaned after the initial air gouging. The robot obtains the point cloud data of the part of the casting workpiece to be cleaned after the initial air gouging, and obtains the feature point positions of the part of the casting workpiece to be cleaned after the initial air gouging based on the feature point positions. The robot also obtains the residual amount of the part to be cleaned after the initial air gouging based on the feature point positions. The robot determines whether the residual amount of the part to be cleaned after the initial air gouging is within the first set threshold range. If not, the cleaning trajectory is re-planned and the host computer controls the six-axis robot to grab the plasma air gouging machine for a second air gouging. If yes, proceed to step 6. Step 6: The host computer controls the six-axis industrial robot to switch tools through the quick-change disc. The six-axis industrial robot grabs the telescopic file to clean the rounded corners of the casting workpiece to be cleaned. After the rounded corners are cleaned, proceed to step 7. Step 7: The host computer controls the six-axis industrial robot to switch tools via a quick-change disc. The six-axis industrial robot picks up the high-intensity grinder to grind the residual material after air gouging, completing the cleaning of the casting workpiece. The specific process of obtaining the cleaning trajectory planning path based on the offset between the feature point position and the reference path in step 3 is as follows: 1) Obtain the highest point of the part of the casting workpiece to be cleaned as the starting point of the planned path. The coordinates of the highest point are (Px, Py, Pz). 2) Calculate the deviation between the highest point and the three-dimensional digital model of the cast workpiece sample. , , ); 3) Use formula (1) to calibrate the coordinate system, transform the workpiece coordinate system into the tool coordinate system, and obtain the horizontal cleaning planning path: （1） Where (Rx, Ry, Rz) is the starting target position of the robot's tool path; 4) Obtain the coordinates Ze of the highest point of the area to be cleaned along the Z-axis and the coordinates Zt of the target position after cleaning, and calculate the amount e to be removed along the Z-axis. z =Ze-Zt; If the amount to be removed in the Z-axis direction is e z When the value is ≥5, the Z-axis cleaning path is planned using fz=2mm / s; If the amount to be removed in the Z-axis direction is e z When <5, the Z-axis cleaning path is planned using the PID algorithm formula (2): fz= Kp*e z (t)+Ki*∑e z (t) + K d *(e z (t)- e z (t-1))(2) Where fz is the feed amount in the Z-axis direction; t is the contact time between the cleaning tool and the workpiece for each feed. The time t depends on the form of the part to be cleaned. For large and thick gates and risers, t is generally selected as 2ms. For ordinary flash or burrs, t is generally selected as 1ms; Kp is the scaling factor, which is generally selected as 0.5 to 2; Ki is the integral factor, which is generally selected as 0.2 to 3; and kd is the differential factor, which is generally selected as 0 to 2.

2. The control method for the robot-based intelligent equipment for casting cleaning according to claim 1, characterized in that: In the process of preliminary air gouging horizontal path planning, the deviation between the highest point and the three-dimensional digital model of the casting workpiece sample is set as ( , , - L 1); The coordinate system is calibrated using formula (3), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the preliminary air gouging cutting horizontal cleaning planning path: （3） Where (Rx, Ry, Rz) represents the starting target position of the robot's tool path. L 1 is the first set threshold.

3. The control method for the robot-based intelligent equipment for casting cleaning according to claim 1, characterized in that: During the horizontal path planning process for secondary air gouging, the deviation between the highest point and the 3D digital model of the cast workpiece sample is set as ( , , - L 2); The coordinate system is calibrated using formula (4), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the secondary air gouging cutting horizontal cleaning planning path: （4） Where (Rx, Ry, Rz) represents the starting target position of the robot's tool path. L 2 represents the residual amount of the area to be cleaned after secondary air gouging.

4. The control method for the robot-based intelligent equipment for casting cleaning according to claim 1, characterized in that: During the horizontal path planning process for high-intensity grinding, the deviation between the highest point and the three-dimensional digital model of the cast workpiece sample is set as ( , , - L 3); The coordinate system is calibrated using formula (5), and the workpiece coordinate system is transformed into the tool coordinate system to obtain the secondary air gouging cutting horizontal cleaning planning path: （5） Where (Rx, Ry, Rz) represents the starting target position of the robot's tool path. L 3 represents the residual amount of the area to be cleaned after grinding by the high-intensity mill.

5. The control method for the robot-based intelligent casting cleaning equipment according to any one of claims 1 to 4, characterized in that: The process also includes the following: during the cleaning of the part of the casting workpiece to be cleaned by the six-axis industrial robot gripping and cleaning tool, the force control device collects force signal data during the cleaning process through a force sensor installed between the quick-change plate and the end effector of the six-axis industrial robot, and transmits the force signal data to the host computer via Ethernet communication. The host computer is used to determine whether the force on the end effector of the six-axis robot is within a second set threshold range based on the force signal data. If not, the host computer sends a stop command to the intelligent equipment control system for casting cleaning.

6. The control method for the robot-based intelligent casting cleaning equipment according to claim 5, characterized in that: The device also includes the force control device collecting force signal data when the cleaning tool and the part to be cleaned just come into contact through a force sensor, and transmitting the force signal data to the host computer. The host computer is used to determine whether the force applied when the cleaning tool and the part to be clean just come into contact is less than a third preset threshold based on the force signal data. If so, it issues a command to replace the cleaning tool.

7. The control method for the robot-based intelligent equipment for casting cleaning according to claim 1, characterized in that: The process also includes steps 3 to 7, during which the intelligent equipment control system for casting cleaning detects whether the safety door of the intelligent equipment for casting cleaning is opened. If so, the intelligent equipment control system for casting cleaning cuts off the power supply to the equipment inside the opened safety door.

8. The control method for the robot-based intelligent casting cleaning equipment according to claim 2, characterized in that: Step 3 further includes the host computer inputting the image information of the casting workpiece to be cleaned into a convolutional neural network for deep learning to obtain a sample model of the casting workpiece, and storing the sample model of the casting workpiece in a sample library; the host computer compares the overlap of the point cloud data of the casting workpiece to be cleaned with all the sample models of casting workpieces in the sample library to obtain the feature point positions of the parts to be cleaned in the casting workpiece to be cleaned compared with the casting workpiece sample model with the largest overlap, as well as the offset of the feature point positions from the reference path, and performs cleaning trajectory planning based on the offset of the feature point positions from the reference path to obtain the cleaning trajectory planning path.

Images