Friction stir welding and grinding machining method and robot integrated equipment

Abstract

The invention discloses a friction stir welding and grinding machining method and robot integrated equipment, and belongs to the technical field of equipment machining. The robot integrated equipment comprises robot equipment, a control system and an electric spindle system, the control system is used for determining a friction stir welding track and a grinding track based on a butt joint face track, and the electric spindle system is used for controlling the friction stir welding track and the grinding track based on application scene data of a to-be-machined workpiece. Friction stir welding process parameters and polishing process parameters are obtained by combining the friction stir welding neural network model and the polishing neural network model, and based on the friction stir welding process parameters, the friction stir welding track, the polishing process parameters and the polishing track, the polishing process parameters are obtained. The robot equipment and the electric spindle system are driven to conduct friction stir welding and grinding on the workpiece to be machined, the robot equipment and the electric spindle system are driven to adjust the friction stir welding and grinding process, and integration of friction stir welding and grinding based on the robot integrated equipment is achieved. And the friction stir welding and grinding quality is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of equipment processing, and particularly relates to a processing method and a robot integrated device for friction stir welding and grinding. Background Art

[0002] Friction stir welding refers to a solid-phase welding method in which materials are heated and plasticized by the rotation and friction of a stirring head, and metallurgical bonding is achieved. It is widely used in new energy vehicles, aviation, aerospace and other fields. Moreover, after friction stir welding, the weld needs to be ground to ensure the flatness and smoothness of the weld surface. For friction stir welding or grinding equipment, it is mainly divided into two types: gantry type and robot type. Among them, due to its advantages such as good accessibility of spatial pose, high production line integration ability, small volume, easy handling, and easy realization of automation and intelligence, the robot type equipment has become the key development direction of friction stir welding equipment in the future.

[0003] In the current processing of friction stir welding and grinding, usually through repeated experiments and attempts, a relatively suitable process parameter is found for welding and grinding. However, since such process parameters are obtained through multiple attempts, the efficiency is low and the accuracy is poor, resulting in low accuracy of friction stir welding and grinding; moreover, due to the great differences in the structure, size, clamping and heat dissipation adjustment of the workpiece to be processed, especially in the case of two-dimensional and three-dimensional trajectory welding, it is more obvious. And simply performing quantitative welding and grinding through process parameters is difficult to adapt to the working condition changes at different processing positions, resulting in low accuracy of friction stir welding and grinding. In addition, the existing equipment only has a single function of friction stir welding or grinding. If it is necessary to realize the two functions of welding and grinding, different equipment needs to be used, increasing the equipment investment and site occupation. Therefore, there is an urgent need for a processing method and a robot integrated device for friction stir welding and grinding to solve the defects of the existing technology. Summary of the Invention

[0004] The present invention aims to provide a processing method and a robot integrated device for friction stir welding and grinding to solve the above technical problems. By calculating the friction stir welding trajectory and the grinding trajectory, as well as the friction stir welding neural network model and the grinding neural network model, the friction stir welding process parameters and the grinding process parameters are determined, realizing the adjustment of friction stir welding and grinding, and improving the processing accuracy of robot friction stir welding and grinding.

[0005] To solve the above technical problems, an embodiment of the present invention provides a processing method for friction stir welding and grinding, including:

[0006] Obtaining a workpiece to be processed, a friction stir welding neural network model and a grinding neural network model;

[0007] Based on the docking surface trajectory of the workpiece to be processed, obtain the friction stir welding trajectory and grinding trajectory of the workpiece to be processed;

[0008] Based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model to perform data deduction on the preset process parameter database, and obtain the friction stir welding process parameters and grinding process parameters of the workpiece to be processed;

[0009] According to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, perform friction stir welding on the workpiece to be processed, and collect real-time friction stir welding data, and then adjust the friction stir welding process of the workpiece to be processed based on the real-time friction stir welding data;

[0010] After completing the friction stir welding of the workpiece to be processed, based on the grinding process parameters and the grinding trajectory of the workpiece to be processed, grind the workpiece to be processed, and collect real-time grinding data, and then adjust the grinding process of the workpiece to be processed according to the real-time grinding data to complete the grinding of the workpiece to be processed.

[0011] It can be understood that, compared with the prior art, the present invention determines the friction stir welding trajectory and the grinding trajectory through the docking surface trajectory of the workpiece to be processed, and then based on the application scenario data of the workpiece to be processed, combines the friction stir welding neural network model and the grinding neural network model to obtain the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed; then performs friction stir welding on the workpiece to be processed through the friction stir welding process parameters and the friction stir welding trajectory, and performs real-time adjustment of the friction stir welding based on the collected real-time friction stir welding data. After completing the friction stir welding, grind the workpiece to be processed through the grinding process parameters and the grinding trajectory, and perform real-time adjustment of the grinding based on the collected real-time grinding data, realizing precise friction stir welding and grinding of the workpiece to be processed, and improving the quality of friction stir welding and grinding. The present invention generates the friction stir welding trajectory and the grinding trajectory through the docking surface trajectory of the workpiece to be processed, ensuring the matching of the friction stir welding trajectory and the grinding trajectory, so that it is possible to perform friction stir welding and grinding simultaneously by a robot device; through the friction stir welding neural network model and the grinding neural network model, it is possible to realize the intelligent decision-making of the friction stir welding process parameters and the grinding process parameters, avoiding the problem of the decrease in the accuracy of the process parameters caused by repeated trial and error commonly used in traditional processes, and also by performing real-time adjustment of the friction stir welding and grinding processes, it is possible to cope with complex working conditions and sudden disturbances during the processing, avoiding processing defects, thereby improving the accuracy of friction stir welding and grinding.

[0012] As a preferred solution, the steps of obtaining the workpiece to be processed, the friction stir welding neural network model, and the grinding neural network model specifically include: obtaining an initial neural network model and the workpiece to be processed, and obtaining a friction stir welding process database and a grinding process database based on the workpiece to be processed; obtaining friction stir welding process feature data based on the friction stir welding process database, and training the neural network structure, loss function, and optimizer of the initial neural network model based on the friction stir welding process feature data to obtain a friction stir welding neural network model; obtaining grinding process feature data based on the grinding process database, and training the neural network structure, loss function, and optimizer of the initial neural network model based on the grinding process feature data to obtain a grinding neural network model.

[0013] As a preferred solution, the steps of obtaining the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed based on the docking surface trajectory of the workpiece to be processed specifically include: determining the welding discrete point data of the workpiece to be processed based on the docking surface trajectory of the workpiece to be processed, and determining the welding coordinates and welding attitude values of the workpiece to be processed based on the welding discrete point data of the workpiece to be processed; determining the friction stir welding trajectory of the workpiece to be processed based on the welding coordinates and welding attitude values of the workpiece to be processed; determining the grinding trajectory of the workpiece to be processed based on the friction stir welding trajectory of the workpiece to be processed.

[0014] As a preferred solution, the steps of determining the welding discrete point data of the workpiece to be processed based on the docking surface trajectory of the workpiece to be processed, and determining the welding coordinates and welding attitude values of the workpiece to be processed based on the welding discrete point data of the workpiece to be processed specifically include: determining the welding discrete point data of the workpiece to be processed based on the docking surface trajectory of the workpiece to be processed, where the welding discrete point data includes: three-dimensional coordinate data of the welding points; performing two-dimensional coordinate projection on the welding discrete point data of the workpiece to be processed based on the three-dimensional coordinate data of the welding points to determine the rotation angle value of the welding points of the workpiece to be processed; calculating the rotation attitude value of the workpiece to be processed based on the rotation angle value of the welding points of the workpiece to be processed; determining the welding attitude value of the workpiece to be processed based on the rotation attitude value of the workpiece to be processed; determining the welding coordinates of the workpiece to be processed based on the three-dimensional coordinate data of the welding points of the workpiece to be processed.

[0015] As a preferred solution, based on the application scenario data of the workpiece to be processed, combining the friction stir welding neural network model and the grinding neural network model, data deduction is performed on the preset process parameter database to obtain the friction stir welding process parameters and grinding process parameters of the workpiece to be processed, specifically including: based on the application scenario data of the workpiece to be processed, weight distribution is performed on the tensile strength, yield strength, and elastic modulus of the workpiece to be processed to obtain the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus; based on the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus of the workpiece to be processed, a quality evaluation function is constructed; the candidate friction stir welding process parameters in the preset process parameter database are input into the friction stir welding neural network model, and data deduction is performed in combination with the quality evaluation function to obtain the friction stir welding process parameters of the workpiece to be processed; the candidate grinding process parameters in the preset process parameter database are input into the friction stir welding neural network model, and data deduction is performed in combination with the quality evaluation function to obtain the grinding process parameters of the workpiece to be processed.

[0016] As a preferred solution, according to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, friction stir welding is performed on the workpiece to be processed, and real-time friction stir welding data is collected. Furthermore, based on the real-time friction stir welding data, the friction stir welding process of the workpiece to be processed is adjusted, specifically including: according to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, friction stir welding is performed on the workpiece to be processed; the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed during the friction stir welding process are collected to determine the real-time friction stir welding data of the workpiece to be processed; based on the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed, the friction stir welding process of the workpiece to be processed is adjusted.

[0017] As a preferred solution, based on the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed, the friction stir welding process of the workpiece to be processed is adjusted, specifically including: the friction stir welding process parameters of the workpiece to be processed include: the optimal welding pressure, the optimal welding torque, and the optimal welding resistance; the welding pressure threshold, welding torque threshold, and welding resistance threshold of the workpiece to be processed are obtained;

[0018] Compare the real-time welding pressure, the optimal welding pressure, and the welding pressure threshold to obtain a welding pressure comparison result, and adjust the welding penetration depth of the friction stir welding process of the workpiece to be processed according to the welding pressure comparison result; compare the real-time welding torque, the optimal welding torque, and the welding torque threshold to obtain a welding torque comparison result, and adjust the welding rotation speed of the friction stir welding process of the workpiece to be processed according to the welding torque comparison result; compare the real-time welding resistance, the optimal welding resistance, and the welding resistance threshold to obtain a welding resistance comparison result, and adjust the welding speed of the friction stir welding process of the workpiece to be processed according to the welding resistance comparison result.

[0019] As a preferred solution, based on the grinding process parameters and grinding trajectory of the workpiece to be processed, grind the workpiece to be processed, collect real-time grinding data, and then adjust the grinding process of the workpiece to be processed according to the real-time grinding data to complete the grinding of the workpiece to be processed, specifically including: grinding the workpiece to be processed according to the grinding process parameters and grinding trajectory of the workpiece to be processed; collecting the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance during the grinding process of the workpiece to be processed to determine the real-time grinding data of the workpiece to be processed; adjusting the grinding process of the workpiece to be processed based on the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed to complete the grinding of the workpiece to be processed.

[0020] As a preferred solution, adjusting the grinding process of the workpiece to be processed based on the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed specifically includes: the grinding process parameters of the workpiece to be processed include: optimal grinding pressure, optimal grinding torque, and optimal grinding resistance; obtaining the grinding pressure threshold, grinding torque threshold, and grinding resistance threshold of the workpiece to be processed; comparing the real-time grinding pressure, optimal grinding pressure, and grinding pressure threshold to obtain a grinding pressure comparison result, and adjusting the grinding penetration depth of the grinding process of the workpiece to be processed according to the grinding pressure comparison result; comparing the real-time grinding torque, optimal grinding torque, and grinding torque threshold to obtain a grinding torque comparison result, and adjusting the grinding rotation speed of the grinding process of the workpiece to be processed according to the grinding torque comparison result; comparing the real-time grinding resistance, optimal grinding resistance, and grinding resistance threshold to obtain a grinding resistance comparison result, and adjusting the grinding speed of the grinding process of the workpiece to be processed according to the grinding resistance comparison result.

[0021] Correspondingly, an embodiment of the present invention provides a robot integrated device for friction stir welding and grinding, including: a robot device, a control system, and an electric spindle system; a drive control output end of the control system is electrically connected to a drive control input end of the robot device; a drive control output end of the control system is electrically connected to a drive control input end of the electric spindle system; a real-time processing data input end of the control system is wirelessly connected to a real-time processing data output end of the electric spindle system; the robot device is connected to the electric spindle system; the control system is configured to obtain a workpiece to be processed, a friction stir welding neural network model, and a grinding neural network model; based on the butt joint surface trajectory of the workpiece to be processed, obtain the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed; based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model, perform data deduction on a preset process parameter database, and obtain the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed; the control system is further configured to drive the robot device and the electric spindle system to perform friction stir welding on the workpiece to be processed based on the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed; the control system is further configured to, after completing the friction stir welding of the workpiece to be processed, drive the robot device and the electric spindle system to perform grinding on the workpiece to be processed based on the grinding process parameters and the grinding trajectory of the workpiece to be processed; the electric spindle system is configured to collect real-time friction stir welding data, so that the control system adjusts the friction stir welding process of the workpiece to be processed by driving the robot device and the electric spindle system based on the real-time friction stir welding data; the electric spindle system is further configured to collect real-time grinding data, so that the control system adjusts the grinding process of the workpiece to be processed by driving the robot device and the electric spindle system based on the real-time grinding data.

[0022] It is understandable that, compared with the prior art, the present device determines the friction stir welding trajectory and the grinding trajectory through the docking surface trajectory of the workpiece to be processed, and then, based on the application scenario data of the workpiece to be processed, combines the friction stir welding neural network model and the grinding neural network model to obtain the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed; then, the workpiece to be processed is subjected to friction stir welding through the friction stir welding process parameters and the friction stir welding trajectory, and the friction stir welding is adjusted in real time based on the collected real-time friction stir welding data. After the friction stir welding is completed, the workpiece to be processed is ground through the grinding process parameters and the grinding trajectory, and the grinding is adjusted in real time based on the collected real-time grinding data, realizing precise friction stir welding and grinding of the workpiece to be processed, and improving the quality of friction stir welding and grinding. The present device generates the friction stir welding trajectory and the grinding trajectory through the docking surface trajectory of the workpiece to be processed, ensuring the matching of the friction stir welding trajectory and the grinding trajectory, so that the friction stir welding and grinding can be carried out simultaneously by the robot device; through the friction stir welding neural network model and the grinding neural network model, the intelligent decision-making of the friction stir welding process parameters and the grinding process parameters can be realized, avoiding the problem of the decrease in the accuracy of the process parameters caused by repeated trial and error commonly used in the traditional process. By adjusting the friction stir welding and the grinding process in real time, it can cope with the complex working conditions and sudden interferences in the processing process, avoid processing defects, and thus improve the accuracy of friction stir welding and grinding. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flowchart of the steps of a processing method for friction stir welding and grinding provided by an embodiment of the present invention;

[0024] Figure 2 It is a schematic diagram of the equipment connection of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention;

[0025] Figure 3 It is a welding schematic diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention;

[0026] Figure 4 It is a tool changing schematic diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention;

[0027] Figure 5 It is a grinding schematic diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention;

[0028] Figure 6 It is a structural schematic diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts belong to the scope of protection of the present invention.

[0030] Embodiment 1

[0031] Please refer to Figure 1 , which is a step flowchart of a processing method for friction stir welding and grinding provided by an embodiment of the present invention, including steps S101 to S105.

[0032] Step S101: Obtain a workpiece to be processed, a friction stir welding neural network model, and a grinding neural network model.

[0033] Step S102: Based on the butt joint surface trajectory of the workpiece to be processed, obtain the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed.

[0034] Step S103: Based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model to perform data deduction on a preset process parameter database, and obtain the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed.

[0035] Step S104: According to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, perform friction stir welding on the workpiece to be processed, and collect real-time friction stir welding data, and then adjust the friction stir welding process of the workpiece to be processed based on the real-time friction stir welding data.

[0036] Step S105: After completing the friction stir welding of the workpiece to be processed, based on the grinding process parameters and the grinding trajectory of the workpiece to be processed, perform grinding on the workpiece to be processed, and collect real-time grinding data, and then adjust the grinding process of the workpiece to be processed according to the real-time grinding data to complete the grinding of the workpiece to be processed.

[0037] In this embodiment, the friction stir welding trajectory and the grinding trajectory are determined based on the docking surface trajectory of the workpiece to be processed. Then, based on the application scenario data of the workpiece to be processed and in combination with the friction stir welding neural network model and the grinding neural network model, the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed are obtained. Subsequently, the workpiece to be processed is subjected to friction stir welding using the friction stir welding process parameters and the friction stir welding trajectory, and the friction stir welding is adjusted in real time based on the collected real-time friction stir welding data. After the friction stir welding is completed, the workpiece to be processed is ground using the grinding process parameters and the grinding trajectory, and the grinding is adjusted in real time based on the collected real-time grinding data, realizing precise friction stir welding and grinding of the workpiece to be processed and improving the quality of friction stir welding and grinding. In this embodiment, the friction stir welding trajectory and the grinding trajectory are generated based on the docking surface trajectory of the workpiece to be processed, ensuring the matching of the friction stir welding trajectory and the grinding trajectory, so that the friction stir welding and grinding can be simultaneously carried out by a robot device. Through the friction stir welding neural network model and the grinding neural network model, intelligent decision-making of the friction stir welding process parameters and the grinding process parameters can be realized, avoiding the problem of the decrease in the accuracy of the process parameters caused by repeated trial and error commonly used in traditional processes. By adjusting the friction stir welding and grinding processes in real time, complex working conditions and sudden interferences during the processing can be handled, avoiding processing defects, and thus improving the accuracy of friction stir welding and grinding.

[0038] Embodiment 2

[0039] Please refer to Figure 2 , which is a schematic diagram of the equipment connection of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention; please refer to Figure 3 , which is a welding schematic diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention; please refer to Figure 4 , which is a tool changing schematic diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention; please refer to Figure 5 , which is a grinding schematic diagram of a robot integrated device provided by an embodiment of the present invention for improvement; Figure 6 , which is a structural schematic diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention;

[0040] As Figures 2 to 6 shown, a processing method for friction stir welding and grinding in Embodiment 1 of the present invention is applied to as Figures 2 to 6A robotic integrated device for friction stir welding and grinding. In a robotic integrated device for friction stir welding and grinding (hereinafter simply referred to as the robotic integrated device), it includes a robotic device 1, an electric spindle system 2, and a control system 3. Among them, the drive control output end of the control system 3 is electrically connected to the drive control input end of the robotic device 1; the drive control output end of the control system 3 is electrically connected to the drive control input end of the electric spindle system 2; the real-time processing data input end of the control system 3 is wirelessly connected to the real-time processing data output end of the electric spindle system 2; the robotic device is connected to the electric spindle system. The electric spindle system 2 includes: an electric spindle 21, a multi-dimensional force sensor 22, a stirring head 23, and a grinding head 24. Among them, the electric spindle 21 and the multi-dimensional force sensor 22 are coaxially connected to the end shaft of the robotic device 1 in sequence; the stirring head 23 is detachably coaxially connected to the lower end of the electric spindle 21; the grinding head 24 is detachably coaxially connected to the lower end of the electric spindle 21; the multi-dimensional force sensor 22 is used to collect process data during friction stir welding and grinding; the electric spindle 21 is used for the detachable coaxial connection of the stirring head 23 or the grinding head 24; the data output end of the multi-dimensional force sensor serves as the real-time processing data output end of the electric spindle system 2. The control system 3 includes: a host computer 31, a motion control module 32, an auxiliary function control module 34, and a spindle control module 33. Among them, the drive instruction output end of the host computer 31 is electrically connected to the drive instruction input end of the motion control module 32; the drive instruction output end of the host computer 31 is electrically connected to the drive instruction input end of the auxiliary function control module 34; the drive instruction output end of the host computer 31 is electrically connected to the drive instruction input end of the spindle control module 33; the data input end of the host computer 31 is wirelessly communicatively connected to the data output end of the multi-dimensional force sensor 22. The drive instruction output ends of the motion control module 32 and the spindle control module 33 both serve as the drive control output ends of the control system 3; the drive input end of the electric spindle 21 is electrically connected to the drive instruction output end of the spindle control module 33; the drive input end of the robotic device 1 is electrically connected to the drive instruction output end of the motion control module 32; the data input end of the host computer 31 is used to receive the process data output by the multi-dimensional force sensor 22; the motion control module 32 is used to receive the motion control instructions from the host computer 31, so as to control the pose and motion of the robotic device 1, and further realize the friction stir welding trajectory, welding pose, welding penetration depth dw, welding speed v w , grinding trajectory, grinding pose, grinding penetration depth dp, grinding speed v pControl; the spindle control module 33 is used to receive the spindle control instruction from the host computer 31, so as to control the rotation speed of the electric spindle 21 and tool change, and further realize the welding rotation speed ω of friction stir welding w and the grinding rotation speed ω p , and control the function switching between welding and grinding; the auxiliary function control module 34 is used to receive the auxiliary function control instruction from the host computer and perform auxiliary function regulation on the robot integrated device. The host computer 31 is provided with process big data software, which is used to set the friction stir welding process parameters and grinding process parameters.

[0041] It should be noted that the friction stir welding and grinding processing method described in the embodiments of the present invention is not limited to the robot integrated device as Figure 2 shown. Other devices with similar structures can also apply the friction stir welding and grinding processing method described in the embodiments of the present invention. The following will specifically expand and explain the friction stir welding and grinding processing method described in the embodiments of the present invention in conjunction with the robot integrated device as Figure 2 shown.

[0042] As Figure 1 shown, step S101 is to obtain the workpiece to be processed, the friction stir welding neural network model and the grinding neural network model.

[0043] In this embodiment, the obtaining of the workpiece to be processed, the friction stir welding neural network model and the grinding neural network model specifically includes: obtaining the initial neural network model and the workpiece to be processed, and obtaining the friction stir welding process database and the grinding process database based on the workpiece to be processed; obtaining the friction stir welding process feature data based on the friction stir welding process database, and training the neural network structure, loss function and optimizer of the initial neural network model based on the friction stir welding process feature data to obtain the friction stir welding neural network model; obtaining the grinding process feature data based on the grinding process database, and training the neural network structure, loss function and optimizer of the initial neural network model based on the grinding process feature data to obtain the grinding neural network model.

[0044] In an alternative embodiment, the initial neural network model is obtained. The initial neural network model in this embodiment is a BP neural network; the friction stir welding process database is obtained based on the workpiece to be processed, and then the friction stir welding process feature data is obtained based on the friction stir welding process database. The friction stir welding process feature data includes the input information and output information of the friction stir welding process database; specifically, the input information of the friction stir welding process database includes, but is not limited to, the welding process parameters of different material types and different material thicknesses, such as: welding rotation speed ω w, welding speed v w , welding penetration depth d w , welding pressure P w , welding torque N w , welding resistance N w etc.; the output information includes welding quality information, i.e., the tensile strength σ of the joint y , the yield strength σ of the joint u , the elastic modulus E of the joint, etc.; then the friction stir welding process characteristic data is divided into a training set and a test set; by training the neural network structure, loss function, optimizer and other model parameters of the initial neural network model with the training set, and evaluating the final model performance with the test set, the neural network model of friction stir process parameters - mechanical parameters - welding quality is determined, that is, the friction stir welding neural network model is determined.

[0045] In an alternative embodiment, an initial neural network model is obtained. The initial neural network model in this embodiment is a BP neural network; a grinding process database is obtained based on the workpiece to be processed, and then the grinding process characteristic data is obtained based on the grinding process database. The grinding process characteristic data includes the input information and output information of the grinding process database; specifically, the input information of the grinding process database includes, but is not limited to, the grinding process parameters of different material types, i.e., the grinding speed ω p , grinding speed v p , grinding penetration depth d p , grinding pressure P p , grinding torque N p , grinding resistance F p etc., and the output information includes the weld surface roughness R, weld surface finish etc.; then the grinding process characteristic data is divided into a training set and a test set; by training the neural network structure, loss function, optimizer and other model parameters of the initial neural network model with the training set, and evaluating the final model performance with the test set, the neural network model of grinding process - mechanical parameters - grinding quality is determined, that is, the grinding neural network model is determined. It should be noted that the BP neural network (BackProgation Network) is a multi-layer feedforward neural network model. The core of the BP neural network lies in using the error backpropagation algorithm, which enables the network to learn complex input-output mapping relationships, thus completing various tasks such as function approximation, pattern recognition, classification, prediction, etc. In particular, the initial neural network model in this embodiment is not limited to the BP neural network. Neural network models such as CNN models and LSTM models with classification, prediction and other functions can all be applied to this embodiment after corresponding modifications.

[0046] In this embodiment, the friction stir welding process feature data and grinding process feature data of the workpiece to be processed are obtained through the friction stir welding process database and grinding process database of the workpiece to be processed, and then the initial neural network model is trained respectively to obtain the friction stir welding neural network model and the grinding neural network model, so that the friction stir welding neural network model and the grinding neural network model can fully explore the processing characteristics of the workpiece to be processed, and thus can realize the intelligent decision-making of the friction stir welding process parameters and grinding process parameters, avoiding the problem of the decrease in the accuracy of the process parameters caused by repeated trial and error commonly used in the traditional process, thereby improving the accuracy of friction stir welding and grinding.

[0047] Step S102 is to obtain the friction stir welding trajectory and grinding trajectory of the workpiece to be processed based on the butt joint surface trajectory of the workpiece to be processed.

[0048] In this embodiment, the obtaining of the friction stir welding trajectory and grinding trajectory of the workpiece to be processed based on the butt joint surface trajectory of the workpiece to be processed specifically includes: determining the welding discrete point data of the workpiece to be processed based on the butt joint surface trajectory of the workpiece to be processed, and determining the welding coordinates and welding attitude values of the workpiece to be processed based on the welding discrete point data of the workpiece to be processed; determining the friction stir welding trajectory of the workpiece to be processed based on the welding coordinates and welding attitude values of the workpiece to be processed; and determining the grinding trajectory of the workpiece to be processed based on the friction stir welding trajectory of the workpiece to be processed.

[0049] In this embodiment, the welding discrete point data is directly derived from the butt joint surface trajectory, ensuring that the welding coordinates and welding attitude values can accurately meet the actual friction stir welding requirements of the workpiece to be processed. Through the collaborative calculation of the welding coordinates and welding attitude values, it can automatically adapt to the complex processing process of the workpiece to be processed, thereby improving the accuracy of the friction stir welding trajectory. Then, the grinding trajectory is determined based on the friction stir welding trajectory to realize the quality collaborative control of different processes, further ensuring the accuracy of the grinding trajectory, and thus improving the accuracy of friction stir welding and grinding.

[0050] In this embodiment, based on the docking surface trajectory of the workpiece to be processed, the welding discrete point data of the workpiece to be processed is determined, and based on the welding discrete point data of the workpiece to be processed, the welding coordinates and welding attitude values of the workpiece to be processed are determined, which specifically includes: based on the docking surface trajectory of the workpiece to be processed, determining the welding discrete point data of the workpiece to be processed, and the welding discrete point data includes: three-dimensional coordinate data of welding points; based on the three-dimensional coordinate data of the welding points, performing two-dimensional coordinate projection on the welding discrete point data of the workpiece to be processed to determine the rotation angle value of the welding points of the workpiece to be processed; calculating the rotation attitude value of the workpiece to be processed based on the rotation angle value of the welding points of the workpiece to be processed; determining the welding attitude value of the workpiece to be processed based on the rotation attitude value of the workpiece to be processed; and determining the welding coordinates of the workpiece to be processed based on the three-dimensional coordinate data of the welding points of the workpiece to be processed.

[0051] In this embodiment, by obtaining the three-dimensional coordinate data of the welding points, it is ensured that the welding path strictly matches the actual geometric features of the workpiece to be processed. Then, through two-dimensional coordinate projection, the attitude calculation process of complex space curves can be simplified, thereby reducing the computational complexity, avoiding errors caused by overly complex calculations, and further improving the accuracy of the welding coordinates and welding attitude values, thus improving the accuracy of the friction stir welding trajectory and the grinding trajectory, and further improving the precision of friction stir welding and grinding.

[0052] It should be noted that the process inclination angle is a unique process characteristic of friction stir welding and is also one of the key factors determining the weld quality. Since the process inclination angle needs to be kept consistent during welding, that is, friction stir welding requires real-time dynamic compensation for the offset of the linear coordinates of the tool center point caused by the inclination movement of the stirring head, and complex three-dimensional curve welding trajectories cannot be programmed using common circular arc motion and motion orientation control instructions. The welding trajectory coordinate values are composed of the spatial coordinate values XYZ, the attitude values ABC, and the additional XY direction inclination component values BC after adding the process. Therefore, in this embodiment, by performing two-dimensional coordinate projection on the three-dimensional coordinate data of the welding points, the maintenance of the process inclination angle is realized, and thus the accuracy of the friction stir welding trajectory and the grinding trajectory is ensured.

[0053] Specifically, based on the docking surface trajectory of the workpiece to be processed, the welding discrete point data of the workpiece to be processed is determined, and the welding discrete point data includes: three-dimensional coordinate data of welding points, and for each welding point P i and the coordinate point data of the next welding point P i+1 , the calculation can be specifically as follows:

[0054]

[0055] After that, calculate the rotation angle value A i of the welding point projected on the XY plane as:

[0056]

[0057] Therefore, the rotation attitude value in the Z-axis direction is equal to the rotation angle value A of the welding point i ; then, the additional value b of the post-welding process inclination angle is calculated according to the following formula i and c i , where θ is the post-welding process inclination angle (the value range is 0 to 3 degrees);

[0058]

[0059] Then, based on the additional value b of the post-welding process inclination angle i and c i , the rotation attitude values B i and C i in the X-axis and Y-axis directions are calculated, so as to obtain the welding coordinates and welding attitude values Pi(X i , Y i , Z i , A i , B i , C i ) of each welding point; thus, the friction stir welding trajectory of the workpiece to be processed is obtained; and the friction stir welding trajectory of the workpiece to be processed is used as the grinding trajectory of the workpiece to be processed.

[0060] It should be noted that the compensation calculation process of the process inclination angle described in the above text embodiment belongs to a conventional offline programming calculation method. Experimental personnel can also select other process inclination angle calculation algorithms according to actual needs to calculate, so as to construct the friction stir welding trajectory and the grinding trajectory.

[0061] Step S103 is to perform data deduction on the preset process parameter database based on the application scenario data of the workpiece to be processed, in combination with the friction stir welding neural network model and the grinding neural network model, so as to obtain the friction stir welding process parameters and grinding process parameters of the workpiece to be processed.

[0062] In this embodiment, based on the application scenario data of the workpiece to be processed, combining the friction stir welding neural network model and the grinding neural network model, data deduction is performed on the preset process parameter database to obtain the friction stir welding process parameters and grinding process parameters of the workpiece to be processed, specifically including: based on the application scenario data of the workpiece to be processed, weight distribution is performed on the tensile strength, yield strength, and elastic modulus of the workpiece to be processed to obtain the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus; based on the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus of the workpiece to be processed, a quality evaluation function is constructed; the candidate friction stir welding process parameters in the preset process parameter database are input into the friction stir welding neural network model, and data deduction is performed in combination with the quality evaluation function to obtain the friction stir welding process parameters of the workpiece to be processed; the candidate grinding process parameters in the preset process parameter database are input into the friction stir welding neural network model, and data deduction is performed in combination with the quality evaluation function to obtain the grinding process parameters of the workpiece to be processed.

[0063] In this embodiment, weight distribution is performed on the tensile strength, yield strength, and elastic modulus through the application scenario data of the workpiece to be processed, so as to ensure that the optimization direction of the friction stir welding process parameters and the grinding process parameters is strictly aligned with the actual working condition requirements. By performing data deduction on the preset process parameter database through the friction stir welding neural network model and the grinding neural network model, intelligent decision-making of the friction stir welding process parameters and the grinding process parameters can be realized, avoiding the problem of reduced accuracy of process parameters caused by repeated trial and error commonly used in traditional processes, thereby improving the accuracy of friction stir welding and grinding.

[0064] In an alternative embodiment, since the application scenarios of the workpiece to be processed are different, for example, the priority of attention to tensile strength is high in the quality characteristics of the automotive field, and the priority of attention to yield strength is high in the quality characteristics of the high-speed rail field. Therefore, in this embodiment, weight distribution is performed on the tensile strength, yield strength, and elastic modulus through the application scenario data of the workpiece to be processed, so as to make the friction stir welding and grinding of the workpiece to be processed more in line with the actual working condition requirements. Specifically, through the maximum-minimum normalization method, the data of the tensile strength, yield strength, and elastic modulus of the workpiece to be processed are mapped into the interval [0,1]. The normalized values of the tensile strength, yield strength, and elastic modulus of the workpiece to be processed are set as Snorm, Ynorm, and Enorm respectively; their respective weights are w Snorm 、w Ynorm and w Enorm ; their weight coefficients satisfy the following priority: w Snorm >w Ynorm ,w Snorm >w Enorm; Therefore, the weight distribution can be carried out based on the following formula:

[0065] w Snorm +w Ynorm +w Enorm = 1;

[0066] w Snorm ≥ 2w Ynorm ;

[0067] w Snorm ≥ 2w Enorm ;

[0068] Specifically, in this embodiment, its weight is set to w Snorm = 0.5, w Ynorm = 0.25, w Enorm = 0.25;

[0069] Therefore, based on the weight values corresponding to the tensile strength, yield strength, and elastic modulus of the workpiece to be processed, a quality evaluation function is constructed, specifically:

[0070] Q = w Snorm ·Snorm + w Ynorm ·Ynorm + w Enorm ·Enorm;

[0071] The preset process parameter database described in this embodiment includes the grinding process database and the friction stir welding process database described above; by inputting the candidate friction stir welding process parameters in the preset process parameter database into the friction stir welding neural network model, with the maximization of the value of the quality evaluation function as the data deduction target, the optimal welding process parameters are selected from the friction stir welding process database to obtain the friction stir welding process parameters of the workpiece to be processed. The friction stir welding process parameters include: the optimal welding rotation speed ω w0 、the optimal welding speed v w0 、the optimal welding penetration depth d w0 、the optimal welding pressure P w0 、the optimal welding torque N w0 and the optimal welding resistance F w0 ; by inputting the candidate grinding process parameters in the preset process parameter database into the friction stir welding neural network model, with the maximization of the value of the quality evaluation function as the data deduction target, the optimal welding process parameters are selected from the grinding process database to obtain the grinding process parameters of the workpiece to be processed. The grinding process parameters include: the optimal grinding rotation speed ω p0 、the optimal grinding speed v p0 、the optimal grinding penetration depth d p0 、the optimal grinding pressure P p0 、the optimal grinding torque N p0 、the optimal grinding resistance Fp0 。

[0072] Step S104 is to perform friction stir welding on the workpiece to be processed according to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, collect real-time friction stir welding data, and then adjust the friction stir welding process of the workpiece to be processed based on the real-time friction stir welding data.

[0073] In this embodiment, the performing friction stir welding on the workpiece to be processed according to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, collecting real-time friction stir welding data, and then adjusting the friction stir welding process of the workpiece to be processed based on the real-time friction stir welding data specifically includes: performing friction stir welding on the workpiece to be processed according to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed; collecting the real-time welding pressure, real-time welding torque, and real-time welding resistance during the friction stir welding process of the workpiece to be processed to determine the real-time friction stir welding data of the workpiece to be processed; and adjusting the friction stir welding process of the workpiece to be processed based on the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed.

[0074] In this embodiment, the adjusting the friction stir welding process of the workpiece to be processed based on the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed specifically includes: the friction stir welding process parameters of the workpiece to be processed include: the optimal welding pressure, the optimal welding torque, and the optimal welding resistance; obtaining the welding pressure threshold, welding torque threshold, and welding resistance threshold of the workpiece to be processed; comparing the real-time welding pressure, the optimal welding pressure, and the welding pressure threshold to obtain a welding pressure comparison result, and adjusting the welding penetration depth of the friction stir welding process of the workpiece to be processed according to the welding pressure comparison result; comparing the real-time welding torque, the optimal welding torque, and the welding torque threshold to obtain a welding torque comparison result, and adjusting the welding rotation speed of the friction stir welding process of the workpiece to be processed according to the welding torque comparison result; comparing the real-time welding resistance, the optimal welding resistance, and the welding resistance threshold to obtain a welding resistance comparison result, and adjusting the welding speed of the friction stir welding process of the workpiece to be processed according to the welding resistance comparison result.

[0075] In this embodiment, by obtaining the real-time welding pressure, real-time welding torque, and real-time welding resistance, and comparing them with the optimal welding pressure, optimal welding torque, and optimal welding resistance respectively, it is possible to adjust the welding rotation speed, welding speed, and welding penetration depth of the workpiece to be processed during the friction stir welding process, so as to adapt to the working condition changes at different processing positions, thereby improving the accuracy of friction stir welding.

[0076] In an optional embodiment, the real-time welding pressure P is collected by the multi-dimensional mechanical sensor 22 w1 , the real-time welding torque N w1 and the real-time welding resistance F w1 ; the welding pressure threshold ΔP of the workpiece to be processed w is set to 1%P w0 ~15%P w0 ; the welding torque threshold ΔN w is set to 1%N w0 ~15%N w0 ; the welding resistance threshold ΔF w is set to 1%F w0 ~15%F w0 ;

[0077] After that, the real-time welding pressure P w1 is compared with the optimal welding pressure P w0 . When the absolute value of the difference between the two is greater than the welding pressure threshold, that is, |P w1 -P w0 |>ΔP w , the welding penetration depth is adjusted. Specifically, when P w1 is greater than P w0 , the real-time welding penetration depth d w1 is reduced at a certain rate until |P w1 -P w0 |≤ΔP w ; when P w1 is less than P w0 , the real-time welding penetration depth d w1 is increased at a certain rate until |P w1 -P w0 |≤ΔP w . In particular, a motion control instruction is sent from the process big data software of the host computer 31 to the motion control module 32, so that the motion control module 32 controls the motion of the robot device 1, thereby realizing the control of the real-time welding penetration depth d w1 . Among them, the control of the real-time welding penetration depth d w1 can be completed by methods such as classical PID control and fuzzy PID control;

[0078] After that, the real-time welding torque N w1 is compared with the optimal welding torque N w0 . When the absolute value of the difference between the two is greater than the welding torque threshold ΔN w , that is, |N w1 -N w0 |>ΔN w , the welding speed is adjusted. Specifically, when N w1 is greater than Nw0 When it is time, increase the real-time welding rotation speed ω at a certain rate w1 , until |N w1 -N w0 |≤ΔN w ; When N w1 is less than N w0 , decrease the real-time welding rotation speed ω at a certain rate w1 , until |N w1 -N w0 |≤ΔN w . In particular, the motion control instruction is sent from the process big data software of the host computer 31 to the spindle control module 33, so that the spindle control module 33 controls the rotation of the motorized spindle 21, and then realizes the control of the real-time welding rotation speed ω w1 ; The control of the real-time welding penetration depth ω w1 can be completed by methods such as classical PID control and fuzzy PID control;

[0079] After that, compare the real-time welding resistance F w1 with the optimal welding resistance F w0 . When the absolute value of the difference between the two is greater than the welding torque threshold ΔF w , that is, |F w1 -F w0 |>ΔF w , start to adjust the welding speed. Specifically, when F w1 is greater than F w0 , decrease the real-time welding speed v w1 at a certain rate, until |F w1 -F w0 |≤ΔF w ; When F w1 is less than F w0 , increase the real-time welding speed v w1 at a certain rate, until |F w1 -F w0 |≤ΔF w . In particular, the motion control instruction is sent from the process big data software of the host computer 31 to the motion control module 32, so that the motion control module 32 controls the motion of the robot device 1, and then realizes the control of the real-time welding speed v w1 ; The control of the real-time welding speed v w1 can be completed by methods such as classical PID control and fuzzy PID control.

[0080] It should be noted that for the above comparisons of the real-time welding pressure, real-time welding torque and real-time welding resistance, those skilled in the art can make combined selections according to actual needs.

[0081] Step S105 is to perform grinding on the workpiece to be processed based on the grinding process parameters and grinding trajectory of the workpiece to be processed after the friction stir welding of the workpiece to be processed is completed, collect real-time grinding data, and then adjust the grinding process of the workpiece to be processed according to the real-time grinding data to complete the grinding of the workpiece to be processed.

[0082] In this embodiment, the performing grinding on the workpiece to be processed based on the grinding process parameters and grinding trajectory of the workpiece to be processed, collecting real-time grinding data, and then adjusting the grinding process of the workpiece to be processed according to the real-time grinding data to complete the grinding of the workpiece to be processed specifically includes: performing grinding on the workpiece to be processed according to the grinding process parameters and grinding trajectory of the workpiece to be processed; collecting the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed during the grinding process, and determining the real-time grinding data of the workpiece to be processed; adjusting the grinding process of the workpiece to be processed based on the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed to complete the grinding of the workpiece to be processed.

[0083] In this embodiment, the adjusting the grinding process of the workpiece to be processed based on the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed specifically includes:

[0084] The grinding process parameters of the workpiece to be processed include: optimal grinding pressure, optimal grinding torque, and optimal grinding resistance; obtaining the grinding pressure threshold, grinding torque threshold, and grinding resistance threshold of the workpiece to be processed; comparing the real-time grinding pressure, optimal grinding pressure, and grinding pressure threshold to obtain a grinding pressure comparison result, and adjusting the grinding depth of the grinding process of the workpiece to be processed according to the grinding pressure comparison result; comparing the real-time grinding torque, optimal grinding torque, and grinding torque threshold to obtain a grinding torque comparison result, and adjusting the grinding speed of the grinding process of the workpiece to be processed according to the grinding torque comparison result; comparing the real-time grinding resistance, optimal grinding resistance, and grinding resistance threshold to obtain a grinding resistance comparison result, and adjusting the grinding speed of the grinding process of the workpiece to be processed according to the grinding resistance comparison result.

[0085] In this embodiment, by obtaining the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance, and respectively comparing them with the optimal grinding pressure, optimal grinding torque, and optimal grinding resistance, the grinding speed, grinding speed, and grinding depth of the workpiece to be processed during the grinding process can be adjusted, so as to adapt to the working condition changes at different processing positions, thereby improving the grinding accuracy.

[0086] In an alternative embodiment, such as Figure 3As shown, the robot integrated device performs friction stir welding on the workpiece 4 to be processed through the stirring head 23. After completing the friction stir welding of the workpiece to be processed, as Figure 4 shown, the host computer 31 issues a spindle control instruction to the spindle control module 33, so that the spindle control module 33 drives the electric spindle system 2 to replace the stirring head 23 at the end of the electric spindle 21 with a grinding head 24. After that, as Figure 5 shown, the robot integrated device grinds the workpiece 4 to be processed through the grinding head 24.

[0087] In an optional embodiment, the real-time grinding pressure P is collected through the multi-dimensional force sensor 22 p1 , the real-time grinding torque N p1 and the real-time grinding resistance F p1 ; the grinding pressure threshold ΔP of the workpiece to be processed p is set to 1%P p0 ~15%P p0 ; the grinding torque threshold ΔN p is set to 1%N p0 ~15%N p0 ; the grinding resistance threshold ΔF p is set to 1%F p0 ~15%F p0 ;

[0088] After that, the real-time grinding pressure P p1 is compared with the optimal grinding pressure P p0 . When the absolute value of the difference between the two is greater than the grinding pressure threshold, that is, |P p1 -P p0 |>ΔP p , the regulation of the grinding depth is started. Specifically, when P p1 is greater than P p0 , the real-time grinding depth d p1 is reduced at a certain rate until |P p1 -P p0 |≤ΔP p ; when P p1 is less than P p0 , the real-time grinding depth d p1 is increased at a certain rate until |P p1 -P p0 |≤ΔP p . In particular, the motion control instruction is sent from the process big data software of the host computer 31 to the motion control module 32, so that the motion control module 32 controls the motion of the robot device 1, and further realizes the control of the real-time grinding depth d p1 ; the control of the real-time welding depth d p1 can be completed by methods such as classical PID control and fuzzy PID control;

[0089] After that, the real-time grinding torque N p1 is compared with the optimal grinding torque N p0 When the absolute value of the difference between the two is greater than the grinding torque threshold ΔN p , that is, |N p1 -N p0 |>ΔN p , the regulation of the grinding speed starts. Specifically, when N p1 is greater than N p0 , the real-time grinding speed ω p1 is increased at a certain rate until |N p1 -N p0 |≤ΔN p ; when N p1 is less than N p0 , the real-time grinding speed ω p1 is decreased at a certain rate until |N p1 -N p0 |≤ΔN p . Particularly, the motion control instruction is sent from the process big data software of the host computer 31 to the spindle control module 33, so that the spindle control module 33 controls the rotation of the electric spindle 21, and further realizes the control of the real-time grinding speed ω p1 ; the control of the real-time grinding speed ω p1 can be completed by methods such as classical PID control and fuzzy PID control;

[0090] After that, the real-time grinding resistance F p1 is compared with the optimal grinding resistance F p0 When the absolute value of the difference between the two is greater than the grinding resistance threshold ΔF p , that is, |F p1 -F p0 |>ΔF p , the regulation of the grinding speed starts. Specifically, when F p1 is greater than F p0 , the real-time grinding speed v p1 is decreased at a certain rate until |F p1 -F p0 |≤ΔF p ; when F p1 is less than F p0 , the real-time grinding speed v p1 is increased at a certain rate until |F p1 -F p0 |≤ΔF p . Particularly, the motion control instruction is sent from the process big data software of the host computer 31 to the motion control module 32, so that the motion control module 32 controls the motion of the robot device 1, and further realizes the control of the real-time grinding speed vp1 Control; for the real-time grinding speed v p1 The control can be completed by methods such as classical PID control and fuzzy PID control.

[0091] It should be specifically noted that for the above comparisons of grinding welding pressure, grinding welding torque, and real-time grinding resistance, those skilled in the art can make combined selections according to actual needs.

[0092] A friction stir welding and grinding processing method described in Embodiment 2 of the present invention trains an initial neural network model through a friction stir welding process database and a grinding process database respectively to obtain a friction stir welding neural network model and a grinding neural network model; then, according to the butt joint surface trajectory of the workpiece to be processed, two-dimensional coordinate projection is performed to calculate the welding coordinates and welding attitude values of the workpiece to be processed, and the friction stir welding trajectory and grinding trajectory of the workpiece to be processed are determined; then, based on the application scenario data of the workpiece to be processed, in combination with the friction stir welding neural network model and the grinding neural network model, data deduction is performed through a quality evaluation function to obtain friction stir welding process parameters and grinding process parameters; then, the workpiece to be processed is sequentially subjected to friction stir welding and grinding according to the friction stir welding process parameters and the grinding process parameters, and real-time adjustment is performed during the friction stir welding and grinding processes, realizing accurate friction stir welding and grinding of the workpiece to be processed. The present invention generates a friction stir welding trajectory and a grinding trajectory through the butt joint surface trajectory of the workpiece to be processed, ensuring the matching of the friction stir welding trajectory and the grinding trajectory, so that friction stir welding and grinding can be simultaneously performed by a robot device; through the friction stir welding neural network model and the grinding neural network model, intelligent decision-making of the friction stir welding process parameters and the grinding process parameters can be realized, avoiding the problem of decreased accuracy of the process parameters caused by repeated trial and error commonly used in traditional processes, and also through real-time adjustment of the friction stir welding and grinding processes, complex working conditions and sudden interferences during the processing can be dealt with, avoiding processing defects, thereby improving the accuracy of friction stir welding and grinding.

[0093] Embodiment 3

[0094] Please refer to Figure 6, which is a schematic structural diagram of a robot integrated device for friction stir welding and grinding provided by an embodiment of the present invention, includes: a robot device 1, a control system 3, and an electric spindle system 2; the drive control output end of the control system 3 is electrically connected to the drive control input end of the robot device 1; the drive control output end of the control system 3 is electrically connected to the drive control input end of the electric spindle system 2; the real-time processing data input end of the control system 3 is wirelessly connected to the real-time processing data output end of the electric spindle system 2; the robot device is connected to the electric spindle system; the control system 3 is used to obtain a workpiece to be processed, a friction stir welding neural network model, and a grinding neural network model; based on the docking surface trajectory of the workpiece to be processed, obtain the friction stir welding trajectory and grinding trajectory of the workpiece to be processed; based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model, perform data deduction on a preset process parameter database, and obtain the friction stir welding process parameters and grinding process parameters of the workpiece to be processed; the control system 3 is further used to drive the robot device and the electric spindle system to perform friction stir welding on the workpiece to be processed based on the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed; the control system 3 is further used to drive the robot device and the electric spindle system to perform grinding on the workpiece to be processed based on the grinding process parameters and the grinding trajectory of the workpiece to be processed after completing the friction stir welding of the workpiece to be processed; the electric spindle system 2 is used to collect real-time friction stir welding data, so that the control system adjusts the friction stir welding process of the workpiece to be processed by driving the robot device and the electric spindle system based on the real-time friction stir welding data; the electric spindle system 2 is further used to collect real-time grinding data, so that the control system adjusts the grinding process of the workpiece to be processed by driving the robot device and the electric spindle system based on the real-time grinding data.

[0095] In this embodiment, the electric spindle system 2 includes: an electric spindle 21, a multi-dimensional force sensor 22, a stirring head 23, and a grinding head 24. Among them, the electric spindle 21 and the multi-dimensional force sensor 22 are coaxially connected to the end shaft of the robot device 1 in sequence; the stirring head 23 is detachably coaxially connected to the lower end of the electric spindle 21; the grinding head 24 is detachably coaxially connected to the lower end of the electric spindle 21; the multi-dimensional force sensor 22 is used to collect process data during friction stir welding and grinding; the electric spindle 21 is used for detachably coaxially connecting the stirring head 23 or the grinding head 24; the drive input end of the electric spindle 21 serves as the drive control input end of the electric spindle system 2.

[0096] In this embodiment, the control system 3 includes: a host computer 31, a motion control module 32, an auxiliary function control module 34, and a spindle control module 33; wherein, the drive instruction output end of the host computer 31 is electrically connected to the drive instruction input end of the motion control module 32; the drive instruction output end of the host computer 31 is electrically connected to the drive instruction input end of the auxiliary function control module 34; the drive instruction output end of the host computer 31 is electrically connected to the drive instruction input end of the spindle control module 33; the data input end of the host computer 31 is wirelessly communicatively connected to the data output end of the multi-dimensional force sensor 22, and the drive instruction output ends of the motion control module 32 and the spindle control module 33 both serve as the drive control output ends of the control system 3; the drive input end of the electric spindle 21 is electrically connected to the drive instruction output end of the spindle control module 33; the drive input end of the robot device 1 is electrically connected to the drive instruction output end of the motion control module 32;

[0097] The data input end of the host computer 31 is used to receive the process data output by the multi-dimensional force sensor 22; the motion control module 32 is used to receive the motion control instructions of the host computer 31, so as to control the pose and motion of the robot device 1, and thus realize the control of the friction stir welding trajectory, welding pose, welding penetration depth dw, welding speed v w , grinding trajectory, grinding pose, grinding penetration depth dp, grinding speed v p ; the spindle control module 33 is used to receive the spindle control instructions of the host computer 31, so as to control the rotation speed of the electric spindle 21 and tool change, and thus realize the control of the welding rotation speed ω w and grinding rotation speed ω p , as well as the control of the function switching between welding and grinding; the auxiliary function control module 34 is used to receive the auxiliary function control instructions of the host computer to perform auxiliary function regulation on the robot integrated device. A process big data software is set in the host computer 31, which is used to set the friction stir welding process parameters and grinding process parameters.

[0098] In this embodiment, the control system 3 is used to obtain the workpiece to be processed, the friction stir welding neural network model, and the grinding neural network model; based on the butt joint surface trajectory of the workpiece to be processed, obtain the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed; based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model, and perform data deduction on the preset process parameter database to obtain the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed; wherein, the obtaining of the workpiece to be processed, the friction stir welding neural network model, and the grinding neural network model specifically includes: obtaining an initial neural network model and the workpiece to be processed, and based on the workpiece to be processed, obtaining a friction stir welding process database and a grinding process database; obtaining friction stir welding process feature data based on the friction stir welding process database, and training the neural network structure, loss function, and optimizer of the initial neural network model based on the friction stir welding process feature data to obtain a friction stir welding neural network model; obtaining grinding process feature data based on the grinding process database, and training the neural network structure, loss function, and optimizer of the initial neural network model based on the grinding process feature data to obtain a grinding neural network model. The obtaining of the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed based on the butt joint surface trajectory of the workpiece to be processed specifically includes: based on the butt joint surface trajectory of the workpiece to be processed, determining the welding discrete point data of the workpiece to be processed, and based on the welding discrete point data of the workpiece to be processed, determining the welding coordinates and welding attitude values of the workpiece to be processed; based on the welding coordinates and welding attitude values of the workpiece to be processed, determining the friction stir welding trajectory of the workpiece to be processed; based on the friction stir welding trajectory of the workpiece to be processed, determining the grinding trajectory of the workpiece to be processed. The determining of the welding discrete point data of the workpiece to be processed based on the butt joint surface trajectory of the workpiece to be processed, and based on the welding discrete point data of the workpiece to be processed, determining the welding coordinates and welding attitude values of the workpiece to be processed specifically includes: based on the butt joint surface trajectory of the workpiece to be processed, determining the welding discrete point data of the workpiece to be processed, and the welding discrete point data includes: three-dimensional coordinate data of the welding points; based on the three-dimensional coordinate data of the welding points, performing two-dimensional coordinate projection on the welding discrete point data of the workpiece to be processed to determine the rotation angle value of the welding points of the workpiece to be processed; calculating the rotation attitude value of the workpiece to be processed based on the rotation angle value of the welding points of the workpiece to be processed; determining the welding attitude value of the workpiece to be processed based on the rotation attitude value of the workpiece to be processed; determining the welding coordinates of the workpiece to be processed based on the three-dimensional coordinate data of the welding points of the workpiece to be processed.Based on the application scenario data of the workpiece to be processed, combining the friction stir welding neural network model and the grinding neural network model, data deduction is performed on the preset process parameter database to obtain the friction stir welding process parameters and grinding process parameters of the workpiece to be processed, specifically including: based on the application scenario data of the workpiece to be processed, weight distribution is performed on the tensile strength, yield strength, and elastic modulus of the workpiece to be processed to obtain the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus; based on the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus of the workpiece to be processed, a quality evaluation function is constructed; the candidate friction stir welding process parameters in the preset process parameter database are input into the friction stir welding neural network model, and data deduction is performed in combination with the quality evaluation function to obtain the friction stir welding process parameters of the workpiece to be processed; the candidate grinding process parameters in the preset process parameter database are input into the friction stir welding neural network model, and data deduction is performed in combination with the quality evaluation function to obtain the grinding process parameters of the workpiece to be processed.

[0099] In this embodiment, the electric spindle system 2 is used to collect real-time friction stir welding data, so that the control system 3 adjusts the friction stir welding process of the workpiece to be processed based on the real-time friction stir welding data by driving the robot device and the electric spindle system; wherein, the multi-dimensional mechanical sensor 22 is used to collect the real-time welding pressure, real-time welding torque and real-time welding resistance of the workpiece to be processed during the friction stir welding process, and determine the real-time friction stir welding data of the workpiece to be processed; so that the control system 3 adjusts the friction stir welding process of the workpiece to be processed by driving the robot device 1 and the electric spindle system 2 based on the real-time welding pressure, real-time welding torque and real-time welding resistance of the workpiece to be processed; wherein, the friction stir welding process parameters of the workpiece to be processed include: optimal welding pressure, optimal welding torque and optimal welding resistance; the host computer 31 is used to obtain the welding pressure threshold, welding torque threshold and welding resistance threshold of the workpiece to be processed; the host computer 31 is also used to compare the real-time welding pressure, optimal welding pressure and welding pressure threshold to obtain a welding pressure comparison result, and generate a motion control instruction according to the welding pressure comparison result, so that the motion control module 32 controls the motion of the robot device 1 based on the motion control instruction, and then adjusts the welding penetration depth of the friction stir welding process of the workpiece to be processed; the host computer 31 is also used to compare the real-time welding torque, optimal welding torque and welding torque threshold to obtain a welding torque comparison result, and generate a motion control instruction according to the welding torque comparison result, so that the spindle control module 33 controls the rotation of the electric spindle 21 based on the motion control instruction, and then adjusts the welding rotation speed of the friction stir welding process of the workpiece to be processed; the host computer 31 is also used to compare the real-time welding resistance, optimal welding resistance and welding resistance threshold to obtain a welding resistance comparison result, and generate a motion control instruction according to the welding torque comparison result, so that the motion control module 32 controls the motion of the robot device 1 based on the motion control instruction, and then adjusts the welding speed of the friction stir welding process of the workpiece to be processed.

[0100] In this embodiment, the electric spindle system 2 is further configured to collect real-time grinding data, so that the control system 3 adjusts the grinding process of the workpiece to be machined based on the real-time grinding data by driving the robot device 1 and the electric spindle system 2; wherein, the multi-dimensional force sensor 22 is used to collect the real-time grinding pressure, real-time grinding torque and real-time grinding resistance of the workpiece to be machined during the grinding process, and determine the real-time grinding data of the workpiece to be machined; so that the control system 3 adjusts the grinding process of the workpiece to be machined based on the real-time grinding pressure, real-time grinding torque and real-time grinding resistance of the workpiece to be machined to complete the grinding of the workpiece to be machined; wherein, the grinding process parameters of the workpiece to be machined include: optimal grinding pressure, optimal grinding torque and optimal grinding resistance; the host computer 31 is used to obtain the grinding pressure threshold, grinding torque threshold and grinding resistance threshold of the workpiece to be machined; the host computer 31 is further used to compare the real-time grinding pressure, the optimal grinding pressure and the grinding pressure threshold to obtain a grinding pressure comparison result, and generate a motion control command according to the grinding pressure comparison result, so that the motion control module 32 controls the motion of the robot device 1 based on the motion control command, and further adjusts the grinding depth of the grinding process of the workpiece to be machined; the host computer 31 is further used to compare the real-time grinding torque, the optimal grinding torque and the grinding torque threshold to obtain a grinding torque comparison result, and generate a motion control command according to the grinding torque comparison result, so that the spindle control module 33 controls the rotation of the electric spindle 21 based on the motion control command, and further adjusts the grinding speed of the grinding process of the workpiece to be machined; the host computer 31 is further used to compare the real-time grinding resistance, the optimal grinding resistance and the grinding resistance threshold to obtain a grinding resistance comparison result, and generate a motion control command according to the grinding resistance comparison result, so that the motion control module 32 controls the motion of the robot device 1 based on the motion control command, and further adjusts the grinding speed of the grinding process of the workpiece to be machined.

[0101] In this embodiment, the friction stir welding trajectory and the grinding trajectory are determined based on the docking surface trajectory of the workpiece to be processed. Then, based on the application scenario data of the workpiece to be processed, the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed are obtained by combining the friction stir welding neural network model and the grinding neural network model. Subsequently, the workpiece to be processed is subjected to friction stir welding using the friction stir welding process parameters and the friction stir welding trajectory, and the friction stir welding is adjusted in real time based on the collected real-time friction stir welding data. After the friction stir welding is completed, the workpiece to be processed is ground using the grinding process parameters and the grinding trajectory, and the grinding is adjusted in real time based on the collected real-time grinding data, realizing precise friction stir welding and grinding of the workpiece to be processed and improving the quality of friction stir welding and grinding. In this embodiment, the friction stir welding trajectory and the grinding trajectory are generated based on the docking surface trajectory of the workpiece to be processed, ensuring the matching of the friction stir welding trajectory and the grinding trajectory, so that the friction stir welding and grinding can be simultaneously carried out by a robotic device. Through the friction stir welding neural network model and the grinding neural network model, intelligent decision-making of the friction stir welding process parameters and the grinding process parameters can be realized, avoiding the problem of reduced accuracy of the process parameters caused by repeated trial and error commonly used in traditional processes. By adjusting the friction stir welding and grinding processes in real time, complex working conditions and sudden interferences during the processing can be handled, avoiding processing defects, and thus improving the precision of friction stir welding and grinding.

[0102] In summary, in the embodiment of the present invention, the friction stir welding trajectory and the grinding trajectory are determined through the docking surface trajectory of the workpiece to be processed. Then, based on the application scenario data of the workpiece to be processed, the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed are obtained by combining the friction stir welding neural network model and the grinding neural network model. Then, the workpiece to be processed is subjected to friction stir welding by the friction stir welding process parameters and the friction stir welding trajectory, and the friction stir welding is adjusted in real time based on the collected real-time friction stir welding data. After the friction stir welding is completed, the workpiece to be processed is ground by the grinding process parameters and the grinding trajectory, and the grinding is adjusted in real time based on the collected real-time grinding data, realizing precise friction stir welding and grinding of the workpiece to be processed and improving the quality of friction stir welding and grinding. In the embodiment of the present invention, the friction stir welding trajectory and the grinding trajectory are generated through the docking surface trajectory of the workpiece to be processed, ensuring the matching of the friction stir welding trajectory and the grinding trajectory, so that the friction stir welding and grinding can be realized simultaneously by a robot device. Through the friction stir welding neural network model and the grinding neural network model, the intelligent decision-making of the friction stir welding process parameters and the grinding process parameters can be realized, avoiding the problem of the decrease in the accuracy of the process parameters caused by repeated trial and error commonly used in the traditional process. By adjusting the friction stir welding and grinding processes in real time, the complex working conditions and sudden interferences in the processing process can be handled, avoiding processing defects, and thus improving the accuracy of friction stir welding and grinding.

[0103] In the specific embodiments described above, the purpose, technical solutions, and beneficial effects of the present invention have been further described in detail. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. In particular, it is pointed out that for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A processing method of friction stir welding and grinding, characterized in that, Including: Obtain the workpiece to be processed, the friction stir welding neural network model, and the grinding neural network model; Based on the butt joint surface trajectory of the workpiece to be processed, obtain the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed; Based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model, perform data deduction on the preset process parameter database, and obtain the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed; According to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, perform friction stir welding on the workpiece to be processed, collect real-time friction stir welding data, and then adjust the friction stir welding process of the workpiece to be processed based on the real-time friction stir welding data; After completing the friction stir welding of the workpiece to be processed, based on the grinding process parameters and the grinding trajectory of the workpiece to be processed, perform grinding on the workpiece to be processed, collect real-time grinding data, and then adjust the grinding process of the workpiece to be processed according to the real-time grinding data to complete the grinding of the workpiece to be processed.

2. The friction stir welding and grinding processing method according to claim 1, characterized in that, The obtaining of the workpiece to be processed, the friction stir welding neural network model, and the grinding neural network model specifically includes: Obtain the initial neural network model and the workpiece to be processed, and based on the workpiece to be processed, obtain the friction stir welding process database and the grinding process database; Based on the friction stir welding process database, obtain the friction stir welding process feature data, and based on the friction stir welding process feature data, train the neural network structure, loss function, and optimizer of the initial neural network model to obtain the friction stir welding neural network model; Based on the grinding process database, obtain the grinding process feature data, and based on the grinding process feature data, train the neural network structure, loss function, and optimizer of the initial neural network model to obtain the grinding neural network model.

3. A processing method of friction stir welding and grinding according to claim 1, characterized in that, The obtaining of the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed based on the butt joint surface trajectory of the workpiece to be processed specifically includes: Based on the butt joint surface trajectory of the workpiece to be processed, determine the welding discrete point data of the workpiece to be processed, and based on the welding discrete point data of the workpiece to be processed, determine the welding coordinates and welding attitude values of the workpiece to be processed; Based on the welding coordinates and welding attitude values of the workpiece to be processed, determine the friction stir welding trajectory of the workpiece to be processed; Based on the friction stir welding trajectory of the workpiece to be processed, determine the grinding trajectory of the workpiece to be processed.

4. A processing method of friction stir welding and grinding according to claim 3, characterized in that, The determining of the welding discrete point data of the workpiece to be processed based on the butt joint surface trajectory of the workpiece to be processed, and the determining of the welding coordinates and welding attitude values of the workpiece to be processed based on the welding discrete point data of the workpiece to be processed specifically includes: Based on the butt joint surface trajectory of the workpiece to be processed, determine the welding discrete point data of the workpiece to be processed, and the welding discrete point data includes: three-dimensional coordinate data of the welding points; Based on the three-dimensional coordinate data of the welding points, project the discrete welding point data of the workpiece to be processed onto a two-dimensional coordinate system to determine the rotation angle value of the welding points of the workpiece to be processed; Calculate the rotation attitude value of the workpiece to be processed based on the rotation angle value of the welding points of the workpiece to be processed; Determine the welding attitude value of the workpiece to be processed based on the rotation attitude value of the workpiece to be processed; Determine the welding coordinates of the workpiece to be processed based on the three-dimensional coordinate data of the welding points of the workpiece to be processed.

5. The friction stir welding and grinding processing method according to claim 1, characterized in that, Based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model to perform data deduction on the preset process parameter database to obtain the friction stir welding process parameters and grinding process parameters of the workpiece to be processed, specifically including: Based on the application scenario data of the workpiece to be processed, assign weights to the tensile strength, yield strength, and elastic modulus of the workpiece to be processed to obtain the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus; Construct a quality evaluation function based on the respective weight values corresponding to the tensile strength, yield strength, and elastic modulus of the workpiece to be processed; Input the candidate friction stir welding process parameters in the preset process parameter database into the friction stir welding neural network model, and perform data deduction in combination with the quality evaluation function to obtain the friction stir welding process parameters of the workpiece to be processed; Input the candidate grinding process parameters in the preset process parameter database into the friction stir welding neural network model, and perform data deduction in combination with the quality evaluation function to obtain the grinding process parameters of the workpiece to be processed.

6. A processing method of friction stir welding and grinding according to claim 1, characterized in that, According to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed, perform friction stir welding on the workpiece to be processed, and collect real-time friction stir welding data. Then, based on the real-time friction stir welding data, adjust the friction stir welding process of the workpiece to be processed, specifically including: Perform friction stir welding on the workpiece to be processed according to the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed; Collect the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed during the friction stir welding process to determine the real-time friction stir welding data of the workpiece to be processed; Adjust the friction stir welding process of the workpiece to be processed based on the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed.

7. The friction stir welding and grinding processing method according to claim 6, characterized in that, Based on the real-time welding pressure, real-time welding torque, and real-time welding resistance of the workpiece to be processed, adjust the friction stir welding process of the workpiece to be processed, specifically including: The friction stir welding process parameters of the workpiece to be processed include: the optimal welding pressure, the optimal welding torque, and the optimal welding resistance; Obtain the welding pressure threshold, welding torque threshold, and welding resistance threshold of the workpiece to be processed; Compare the real-time welding pressure, the optimal welding pressure, and the welding pressure threshold to obtain a welding pressure comparison result, and adjust the welding penetration depth of the friction stir welding process of the workpiece to be processed according to the welding pressure comparison result; Compare the real-time welding torque, the optimal welding torque, and the welding torque threshold to obtain a welding torque comparison result, and adjust the welding rotation speed of the friction stir welding process of the workpiece to be processed according to the welding torque comparison result; Compare the real-time welding resistance, the optimal welding resistance, and the welding resistance threshold to obtain a welding resistance comparison result, and adjust the welding speed of the friction stir welding process of the workpiece to be processed according to the welding resistance comparison result.

8. A processing method of friction stir welding and grinding according to claim 7, characterized in that, Based on the grinding process parameters and the grinding trajectory of the workpiece to be processed, grind the workpiece to be processed, collect real-time grinding data, and then adjust the grinding process of the workpiece to be processed according to the real-time grinding data to complete the grinding of the workpiece to be processed. Specifically, it includes: Grind the workpiece to be processed according to the grinding process parameters and the grinding trajectory of the workpiece to be processed; Collect the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed during the grinding process, and determine the real-time grinding data of the workpiece to be processed; Based on the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed, adjust the grinding process of the workpiece to be processed to complete the grinding of the workpiece to be processed.

9. A processing method of friction stir welding and grinding according to claim 8, characterized in that, The adjustment of the grinding process of the workpiece to be processed based on the real-time grinding pressure, real-time grinding torque, and real-time grinding resistance of the workpiece to be processed specifically includes: The grinding process parameters of the workpiece to be processed include: optimal grinding pressure, optimal grinding torque, and optimal grinding resistance; Obtain the grinding pressure threshold, grinding torque threshold, and grinding resistance threshold of the workpiece to be processed; Compare the real-time grinding pressure, the optimal grinding pressure, and the grinding pressure threshold to obtain a grinding pressure comparison result, and adjust the grinding depth of the grinding process of the workpiece to be processed according to the grinding pressure comparison result; Compare the real-time grinding torque, the optimal grinding torque, and the grinding torque threshold to obtain a grinding torque comparison result, and adjust the grinding rotation speed of the grinding process of the workpiece to be processed according to the grinding torque comparison result; Compare the real-time grinding resistance, the optimal grinding resistance, and the grinding resistance threshold to obtain a grinding resistance comparison result, and adjust the grinding speed of the grinding process of the workpiece to be processed according to the grinding resistance comparison result.

10. A robotic integrated device for friction stir welding and grinding, characterized in that, It includes: A robot device, a control system, and an electric spindle system; The drive control output end of the control system is electrically connected to the drive control input end of the robot device; The drive control output end of the control system is electrically connected to the drive control input end of the electric spindle system; the real-time processing data input end of the control system is wirelessly connected to the real-time processing data output end of the electric spindle system; the robot device is connected to the electric spindle system; The control system is used to obtain the workpiece to be processed, the friction stir welding neural network model, and the grinding neural network model; based on the docking surface trajectory of the workpiece to be processed, obtain the friction stir welding trajectory and the grinding trajectory of the workpiece to be processed; based on the application scenario data of the workpiece to be processed, combine the friction stir welding neural network model and the grinding neural network model, perform data deduction on the preset process parameter database, and obtain the friction stir welding process parameters and the grinding process parameters of the workpiece to be processed; The control system is further used to drive the robot device and the electric spindle system to perform friction stir welding on the workpiece to be processed based on the friction stir welding process parameters and the friction stir welding trajectory of the workpiece to be processed; The control system is further used to drive the robot device and the electric spindle system to grind the workpiece to be processed based on the grinding process parameters and the grinding trajectory of the workpiece to be processed after completing the friction stir welding of the workpiece to be processed; The electric spindle system is used to collect real-time friction stir welding data, so that the control system can drive the robot device and the electric spindle system to adjust the friction stir welding process of the workpiece to be processed based on the real-time friction stir welding data; The electric spindle system is further used to collect real-time grinding data, so that the control system can drive the robot device and the electric spindle system to adjust the grinding process of the workpiece to be processed based on the real-time grinding data.

Images