Multi-specification nut rapid locking method and device based on 3D vision

Abstract

The invention provides a multi-specification nut rapid locking method and device based on 3D vision, and relates to the technical field of automatic assembly, and the method comprises the steps: obtaining the three-dimensional point cloud data of a target nut through the scanning of a 3D vision module; determining a target nut locking strategy according to the nut code identifier; inputting the simulation result into a nut simulation model to carry out locking simulation, and carrying out quality analysis on the simulation result; if the target locking index does not meet the preset standard, optimizing a target nut locking strategy; and performing automatic locking on the target nut based on the target locking optimal strategy. The technical problem that in the prior art, due to the fact that nuts of different materials or specifications have different requirements for torque in the locking process, unified torque control and precision are difficult to achieve, and consequently the nut locking precision is low can be solved, the locking strategy is determined through the three-dimensional point cloud data, the locking strategy is optimized in combination with analogue simulation, and the nut locking precision is improved. And the nut locking precision and efficiency are improved.

Description

Technical Field

[0001] The present application relates to the field of automated assembly technology, and in particular to a method and device for quickly locking nuts of multiple specifications based on 3D vision. Background Art

[0002] Nut locking is a common fixing method in mechanical assembly and is widely used in the automotive, aerospace, home appliance and other industries. The locking operation of the nut usually relies on torque control, and the size of the torque directly determines the locking effect of the nut and the firmness of the connector. The locking requirements of different nuts are not the same, especially the torque requirements of the nut materials (such as steel, aluminum, plastic, etc.) vary greatly. Existing methods are often difficult to quickly and accurately adapt to the locking requirements of all types of nuts. The torque requirements vary depending on the material and size of the nut, and a unified torque control strategy is often difficult to meet the requirements of all nut types. Excessive or insufficient tightening of nuts still occurs from time to time, affecting the locking accuracy and reliability of the connector.

[0003] In summary, the prior art has a technical problem in which nuts of different materials or specifications have different torque requirements during the locking process, and unified torque control and accuracy are difficult to achieve, resulting in low nut locking accuracy. Summary of the invention

[0004] The purpose of this application is to provide a method and device for quickly locking nuts of multiple specifications based on 3D vision, so as to solve the technical problem in the prior art that nuts of different materials or specifications have different torque requirements during the locking process, and unified torque control and accuracy are difficult to achieve, resulting in low nut locking accuracy.

[0005] In view of the above problems, the present application provides a method and device for quickly locking nuts of multiple specifications based on 3D vision.

[0006] In a first aspect, the present application provides a method for quickly locking multi-specification nuts based on 3D vision, and the method for quickly locking multi-specification nuts based on 3D vision is implemented by a device for quickly locking multi-specification nuts based on 3D vision, wherein the method for quickly locking multi-specification nuts based on 3D vision includes: scanning and acquiring three-dimensional point cloud data of a target nut through a 3D vision module built into a multi-axis robot; a nut type classifier determines a target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identifier; interactively acquiring a set of historical nut locking schemes, and traversing the historical nut locking schemes according to the nut coding identifier to obtain a target nut assembly station, a target nut locking tool, and a target nut locking parameter, and forming a target nut locking strategy; a nut simulation model performs locking simulation through the target nut locking strategy, and performs quality analysis on the simulation results to obtain a target locking index; if the target locking index does not meet a preset standard, the target nut locking strategy is optimized to obtain an optimal target locking strategy; the multi-axis robot automatically locks the target nut based on the optimal target locking strategy.

[0007] In the second aspect, the present application also provides a multi-specification nut quick locking device based on 3D vision, which is used to execute the multi-specification nut quick locking method based on 3D vision as described in the first aspect, wherein the multi-specification nut quick locking device based on 3D vision includes: a point cloud data acquisition module, the point cloud data acquisition module is used to scan and obtain the three-dimensional point cloud data of the target nut through the built-in 3D vision module of the multi-axis robot; a nut type determination module, the nut type determination module is used for the nut type classifier to determine the target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identification; a locking strategy determination module, the locking strategy determination module is used to interactively obtain historical nut locking strategies A case set, according to the nut coding identifier, traverses the historical nut locking solution set to obtain the target nut assembly station, target nut locking tool and target nut locking parameters, and constitutes a target nut locking strategy; a locking simulation module, the locking simulation module is used for the nut simulation model to perform locking simulation through the target nut locking strategy, and perform quality analysis on the simulation result to obtain a target locking index; a locking strategy optimization module, the locking strategy optimization module is used to optimize the target nut locking strategy if the target locking index does not meet the preset standard to obtain the target locking optimal strategy; an automatic locking module, the automatic locking module is used for the multi-axis robot to automatically lock the target nut based on the target locking optimal strategy.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: The three-dimensional point cloud data of the target nut is obtained by scanning the built-in 3D vision module of the multi-axis robot; the nut type classifier determines the target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identification; the historical nut locking scheme set is interactively obtained, and the target nut assembly station, target nut locking tool and target nut locking parameter are traversed in the historical nut locking scheme set according to the nut coding identification, and the target nut locking strategy is formed; the nut simulation model performs locking simulation through the target nut locking strategy, and performs quality analysis on the simulation results to obtain the target locking index; if the target locking index does not meet the preset standard, the target nut locking strategy is optimized to obtain the target locking optimal strategy; the multi-axis robot automatically locks the target nut based on the target locking optimal strategy. In other words, the specification type of the nut is identified by the 3D vision module, the corresponding nut locking strategy is selected from the historical nut locking scheme set, the target nut locking strategy is optimized through simulation, and the target nut is automatically locked, thereby improving the accuracy and efficiency of nut locking.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0011] Figure 1 This is a flow chart of a method for fast locking nuts of multiple specifications based on 3D vision in this application; Figure 2 This is a schematic diagram of the structure of a multi-specification nut quick locking device based on 3D vision in this application.

[0012] Explanation of the reference numerals: point cloud data acquisition module 11 , nut type determination module 12 , locking strategy determination module 13 , locking simulation module 14 , locking strategy optimization module 15 , automatic locking module 16 . DETAILED DESCRIPTION

[0013] This application provides a method and device for fast locking nuts of multiple specifications based on 3D vision, which solves the technical problem in the prior art that nuts of different materials or specifications have different torque requirements during the locking process, and it is difficult to achieve unified torque control and accuracy, resulting in low nut locking accuracy. The specification type of the nut is identified through a 3D vision module, and the corresponding nut locking strategy is selected from the historical nut locking scheme set. The target nut locking strategy is optimized through simulation, and the target nut is automatically locked, thereby improving the accuracy and efficiency of nut locking.

[0014] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0015] For example, please refer to the attached Figure 1 The present application provides a method for fast locking of multi-specification nuts based on 3D vision, wherein the method for fast locking of multi-specification nuts based on 3D vision is applied to a device for fast locking of multi-specification nuts based on 3D vision, and the method for fast locking of multi-specification nuts based on 3D vision specifically comprises the following steps: S100: The 3D point cloud data of the target nut is obtained by scanning with the built-in 3D vision module of the multi-axis robot.

[0016] Specifically, the multi-axis robot scans the target nut through the built-in 3D vision module. The 3D vision module uses a 3D camera to obtain the three-dimensional point cloud data of the target nut. The built-in 3D vision module is activated to scan the nut. During the scanning process, the module emits light (such as laser or structured light) and captures the reflected light. The position of each point in the three-dimensional space is calculated through the principle of triangulation and converted into a point cloud format, that is, a three-dimensional model of the nut. The 3D vision module is a device that uses a depth sensor (such as lidar, structured light, stereo vision, etc.) to obtain three-dimensional information of an object. Through a high-resolution camera or sensor, it can not only capture the plane image of the object, but also obtain the depth information of the object (that is, the distance between the object and the sensor). The three-dimensional point cloud data provides detailed geometric information about the target nut, including the diameter, height, thread size of the nut and its position in space.

[0017] Through the 3D vision module, the robot scans different types of nuts (such as M8 and M10 nuts). The point cloud data of each nut contains thousands of points, describing its appearance and geometric information. If the surface of a nut contains debris, the point cloud collected by the 3D vision module may contain unnecessary data. Use the filtering function in PCL (Point Cloud Library) to remove these noise data and obtain a more accurate surface shape of the nut. Through the 3D vision module, the robot can quickly and accurately obtain the geometric shape and size of the nut, with higher accuracy than traditional manual measurement methods.

[0018] S200: A nut type classifier determines a target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identification.

[0019] Specifically, the nut type classifier is a tool used to identify and classify different types of nuts. It classifies nuts into different types by inputting feature data (such as 3D point cloud data, image data, etc.), analyzing the shape, size, thread type and other features of the nut to automatically identify the type of nut. The nut type classifier classifies nuts into different types (such as M6, M8, M10, etc.) by analyzing the geometric features of the nuts (such as outer diameter, thickness, thread type, etc.). After the classification is completed, each nut will be assigned a unique nut code identification.

[0020] Two-dimensional image datasets of different types of nuts. The image datasets are obtained by photographing or scanning the nuts at different angles, including front view, top view and side view. Each view can provide different appearance information of the nut. The front view shows the front of the nut, the top view shows the top view of the nut, and the side view shows the side of the nut. The geometric features of the nut can be fully understood through different viewpoints. After collecting a large number of nut images, data preprocessing is required to ensure the consistency and quality of the data. All images are unified to the same size for easy input into the neural network. In order to increase the diversity of training data, data enhancement techniques (such as rotation, flipping, translation, scaling, etc.) are used to expand the dataset. The pixel values ​​of the image are normalized to make the range of the input data uniform.

[0021] In order to realize the automatic classification of nut types, a suitable deep learning model is first selected. Usually, a convolutional neural network (CNN) or other models based on deep neural networks are used to process point cloud data. CNN is a deep learning model that is good at processing spatial data (such as images, point clouds, etc.), and is particularly suitable for extracting geometric features and spatial structure information from point cloud data. The convolutional neural network architecture includes convolutional layers, pooling layers, and fully connected layers. The convolutional layer extracts image features, the pooling layer reduces the feature dimension, and the fully connected layer performs classification. The convolutional neural network architecture is trained with the processed two-dimensional image dataset. The labeled nut types (such as M6, M8, etc.) are used as labels. The back propagation algorithm is used to adjust the weights and biases in the network to optimize the network performance. The training process includes forward propagation, loss calculation, back propagation, and parameter update. After training, the nut type classifier is used in practical applications. When a new nut sample is input, the nut type is classified in real time and the corresponding nut type is output.

[0022] The three-dimensional point cloud data is converted into a two-dimensional image format, and the target front view image, target top view image and target side view image of the target nut are obtained through multi-visual projection, that is, the two-dimensional image set of the target nut. The converted two-dimensional image set is standardized, such as normalization, cropping and scaling, to make the image size consistent. The converted two-dimensional image set is input into the nut type classifier for prediction, and the type of the target nut is output.

[0023] After the classification is completed, the nut type classifier will assign a unique coding mark to each type of nut, which can not only distinguish different types of nuts, but also match the relevant locking schemes, tool selection, torque settings and other parameters to ensure the accuracy of the locking process. Through the nut type classifier, different types of nuts can be accurately identified. The automatic generation of nut coding marks enables the entire nut assembly process to be completed without manual intervention, and automatically calls the locking parameters that match the nut type (such as tool selection, torque setting), thereby improving the locking accuracy, reducing the possibility of errors, and improving the automation level of the production line.

[0024] S300: interactively obtain a historical nut locking solution set, traverse the historical nut locking solution set according to the nut coding identifier to obtain a target nut assembly station, a target nut locking tool and a target nut locking parameter, and form a target nut locking strategy.

[0025] Specifically, the historical nut locking scheme set is obtained through the cloud storage system, which contains the historical locking processes of different nut types (such as M6, M8, M10, etc.), including workstations, tool selection, locking parameters and other information. According to the coding identification of the target nut, the historical nut locking scheme set is traversed to find the target nut assembly workstation, target nut locking tool and target nut locking parameters that match the target nut type. Different types of nuts have unique identification codes, which refer to the specific nut type, and also include the corresponding specific location or workbench for nut locking operation. Different nut types need to be locked at different workstations, for example, there are special automated equipment or workbenches for assembling M8 nuts. The target nut locking tool refers to the tool used to perform the nut locking operation, including nut wrench, torque tool, electric screwdriver, etc. Different tools are selected according to the nut type and locking requirements. The target nut locking parameters refer to the specific parameters related to nut locking, such as torque value, locking speed, locking angle, etc., which affect the locking effect of the nut and ensure that the nut is not too tight or too loose.

[0026] According to the specific assembly stations, tools and locking parameters that match the target nut type, a detailed operation strategy is formulated for the target nut, and automatic selection is performed according to the actual configuration of the production line (such as station number, tool model). By combining the nut coding identification, target station, target tool and locking parameters, a complete target nut locking strategy is generated to ensure that the nut can be locked at the correct station with the correct tools and parameters. By automatically querying historical locking schemes and generating target locking strategies, the intelligent and automated nut locking process is realized. According to the locking strategy optimized by the historical scheme set, each nut is ensured to be locked at the appropriate station, tool and parameter, which improves the locking accuracy and consistency.

[0027] S400: The nut simulation model performs locking simulation according to the target nut locking strategy, and performs quality analysis on the simulation result to obtain a target locking index.

[0028] Specifically, according to the target nut type, the corresponding target nut simulation channel is determined in the nut simulation model, and the target nut locking strategy (including torque, locking tool, work station and other parameters) needs to be input into the nut simulation model. The simulation parameters are updated through the corresponding target nut simulation channel to simulate the performance of the nut during the locking process. Multiple test scenario parameters are obtained from the storage module, and multiple test scenario parameters are used to perform locking simulation on the target nut simulation channel. The simulation will be performed under different test scenarios, such as standard working conditions, extreme working conditions and abnormal working conditions, to simulate different working conditions. Each simulation will produce a result, which will be collected and quality analyzed, including the torque application process, thread engagement, stress distribution, nut deformation, etc., to finally obtain the target locking index.

[0029] The simulation results are analyzed for quality, mainly including the evaluation of torque fluctuation, thread engagement, stress distribution and other aspects. Torque analysis is used to check whether the torque value is stable and whether the expected locking accuracy is achieved; thread engagement analysis is used to evaluate the ratio of thread contact area to preset contact area to ensure complete thread engagement; stress analysis is used to evaluate the stress distribution between nuts and bolts to avoid excessive local stress leading to material fatigue or thread damage. According to the results of quality analysis, the target locking index is calculated. The target locking index combines the analysis results of torque control accuracy, thread engagement quality and stress distribution, and ultimately reflects the overall quality of the locking process. Usually, the analysis results of torque control accuracy, thread engagement quality and stress distribution under each specific working condition are combined to obtain the locking index of the working condition, that is, the effect of the target nut locking strategy under the working condition.

[0030] The locking indexes obtained under the three working conditions are normalized and standardized to avoid errors caused by dimensionality problems, and then the final target locking index is calculated based on the predetermined ratio. Through simulation and quality analysis, potential problems in the locking process are predicted before actual operation, thereby optimizing the locking strategy, improving the overall locking quality, identifying the key factors affecting the locking quality, and then adjusting the locking parameters (such as torque, locking tools, workstations, etc.) to improve process accuracy.

[0031] S500: If the target locking index does not meet a preset standard, the target nut locking strategy is optimized to obtain an optimal target locking strategy.

[0032] Specifically, the preset standard is a reference value set according to production requirements, process specifications or quality control objectives, which is usually a locking quality requirement verified by industry standards or internal control processes of the enterprise. If the target locking index does not meet the preset standard, the locking strategy of the target nut is optimized to obtain the optimal strategy for the target locking. The optimization process usually constructs a locking parameter optimization space based on the historical nut locking scheme set, and takes maximizing locking accuracy, maximizing thread engagement area and minimizing stress concentration as the multivariate optimization objectives of the target nut. Multi-objective optimization is performed in the locking parameter optimization space to determine multiple locking optimization strategies that meet the multivariate optimization objectives.

[0033] The importance of each optimization target in the multivariate optimization target is evaluated, the evaluation coefficient is determined, and the weights are assigned to the multivariate optimization targets to obtain multiple comprehensive evaluation results of multiple locking optimization strategies. The locking optimization strategy with the highest score in the comprehensive evaluation results is selected as the optimal locking strategy. The target locking optimal strategy is the best locking solution obtained after multiple optimizations, which can maximize the locking requirements of the nuts and ensure that the preset quality standards can be met under different working conditions. By optimizing the target locking strategy, the locking quality of the nuts under different working conditions is more stable, thereby improving the overall assembly quality. The optimized strategy can control the torque more accurately to avoid excessive or insufficient tightening. The optimized locking strategy can reduce rework and readjustment caused by locking quality problems, thereby improving production efficiency and reducing costs.

[0034] S600: The multi-axis robot automatically locks the target nut based on the target locking optimal strategy.

[0035] Specifically, a multi-axis robot refers to an industrial robot with multiple degrees of freedom, which can usually operate in multiple directions. Each degree of freedom can control the position, posture and motion trajectory of the robot's end effector, making it highly flexible and accurate in complex tasks. The target locking optimal strategy is the best locking solution based on the aforementioned multivariate optimization process. It usually includes the best locking parameters, as well as the assembly station and locking tools corresponding to the target nut to ensure the best nut locking quality and meet production requirements. The multi-axis robot automatically locks the target nut according to the final determined target locking optimal strategy, performs tasks according to the preset control system and optimization strategy, and achieves efficient and accurate nut locking.

[0036] The multi-axis robot obtains all necessary operating parameters according to the optimal strategy for target locking, including torque, locking tools (such as electric wrenches or pneumatic tools), assembly stations, operation sequences, etc. The multi-axis robot locates the precise position of the target nut through the built-in vision system (such as a 3D vision module) or sensor, and controls its end effector (usually a gripper or a specific tool) to grab the target nut. The target nut is transferred to the designated assembly station and is ready to perform the locking operation. The configuration and operation sequence of the assembly station are also designed based on the optimal strategy. According to the optimal strategy for target locking, the multi-axis robot applies the specified torque value through an electric tool (such as an electric wrench), and monitors the torque in real time through a built-in torque sensor or a closed-loop control system to ensure that it meets the set standards.

[0037] During the locking process, the robot will monitor in real time, check the locking status of the nut and the thread engagement, and ensure that there is no over- or under-locking. The locking quality is verified by visual modules or other sensors to ensure that the nut is firmly fixed and meets the design requirements. Once the locking is completed, the robot will release the nut and prepare to grab the next target nut. All operations are adjusted based on real-time feedback to ensure the efficiency and accuracy of the entire process. Through automatic locking, production efficiency is greatly improved. The multi-axis robot can continuously and quickly complete the process of grabbing, locking and testing nuts, greatly shortening the production cycle and improving the throughput of the assembly line.

[0038] Further, the present application S400 includes: Based on the target nut type, the nut simulation model is traversed to determine the corresponding target nut simulation channel; based on the target nut locking strategy, the parameters of the target nut simulation channel are updated; multiple test scenario parameters are obtained, and locking simulation is performed on the target nut simulation channel in turn through the multiple test scenario parameters to obtain multiple simulation results; quality analysis is performed on the multiple simulation results to obtain the target locking index.

[0039] Specifically, according to the determined target nut type, the corresponding target nut simulation channel is determined in the nut simulation model. The target nut simulation channel refers to the simulation process or calculation path related to a specific nut type in the nut simulation model. Each nut type has a corresponding simulation channel, which contains the geometry, material, locking conditions and other characteristics of the nut type, and is used to simulate its locking effect under specific working conditions.

[0040] According to the target nut locking strategy (such as the selected locking position, tool, torque parameters, etc.), the relevant parameters in the simulation channel are updated, including setting torque, locking angle, tool selection and other factors. In other words, the locking strategy parameters are input into the corresponding target nut simulation channel in the simulation software, and the boundary conditions and initial conditions in the simulation model are updated.

[0041] The simulation software can configure multiple scenario parameters and simulate them in sequence to simulate the locking process under different working conditions. Using the configured simulation channel, simulate the performance of the target nut type under different working conditions, including standard working conditions, extreme working conditions, and abnormal working conditions. Perform quality analysis on the simulation results under different working conditions, evaluate the torque distribution, thread engagement, stress distribution, etc. during the locking process, which helps to ensure the locking quality of the nut under different working conditions. Simulate the distribution of torque during the locking process to check whether it is applied evenly to avoid excessive stress on the nut or thread. Simulate the engagement of the nut and bolt threads to check for problems such as local slippage or excessive contact pressure. Analyze the stress distribution of the nut under different working conditions to identify possible failure areas, especially stress concentration in the threaded area.

[0042] The standard working condition locking index, extreme working condition locking index and abnormal working condition locking index obtained by torque analysis, thread engagement analysis and stress analysis are normalized and standardized to avoid errors caused by inconsistent dimensions. Finally, the target locking index is obtained by weighted calculation based on a predetermined ratio. Through simulation analysis under multiple working conditions, it is possible to predict possible problems in the locking process and optimize the locking strategy to ensure the locking quality of the nut.

[0043] Furthermore, the present application also includes the following steps: According to the multiple test scenario parameters, standard test scenario parameters, extreme test scenario parameters and abnormal test scenario parameters are extracted; standard test results corresponding to the standard test scenario parameters are extracted from the multiple simulation results; torque analysis, thread engagement analysis and stress analysis are performed on the standard test results to obtain a standard operating condition locking index; based on the extreme test scenario parameters, an extreme operating condition locking index is obtained; based on the abnormal test scenario parameters, an abnormal operating condition locking index is obtained; a predetermined ratio is obtained, and the target locking index is obtained by combining the standard operating condition locking index, the extreme operating condition locking index and the abnormal operating condition locking index.

[0044] Specifically, based on the multiple test scenario parameters obtained, standard test scenario parameters, extreme test scenario parameters and abnormal test scenario parameters are extracted and locking simulation is performed through the target nut simulation channel. After completing the simulation of different test scenarios, the simulation results corresponding to the standard test scenario are extracted respectively, including the behavior of the nut under standard working conditions, extreme working conditions and abnormal working conditions. Standard test scenario parameters refer to test scenario parameters used under normal or standard working conditions, representing common working conditions. Extreme test scenario parameters refer to test parameters used when approaching or reaching the capacity limit of the equipment or process, and are usually used to evaluate the performance of the locking strategy under extreme conditions. Abnormal test scenario parameters refer to test scenario parameters used under uncommon or abnormal working conditions, and are usually used to detect fault tolerance and emergency performance.

[0045] The standard test results corresponding to the standard test scenario parameters are subjected to torque analysis, thread engagement analysis and stress analysis to obtain the standard working condition locking index. The locking index is an indicator obtained based on a series of analysis results (such as torque analysis, thread engagement analysis and stress analysis). It measures the quality of the nut locking process and reflects whether the nut locking is smooth and whether there are potential problems. Analyze whether the torque required for the nut under standard working conditions meets the specifications and whether the torque distribution is uniform. Analyze the engagement of the nut and bolt threads to check for poor engagement or slippage. Analyze the stress distribution of various parts of the nut to confirm whether there are stress concentrations or areas that may cause failure.

[0046] Under extreme working conditions, the same analysis process is repeated to obtain simulation results under extreme working conditions, which reflects the locking performance of the nut under large torque or harsh environmental conditions. Under abnormal working conditions, torque analysis, thread engagement analysis and stress analysis are also performed to evaluate the locking performance of the nut under non-ideal conditions. The locking index of standard working conditions, the locking index of extreme working conditions and the locking index of abnormal working conditions are combined.

[0047] The lock-up index of standard working condition, lock-up index of extreme working condition and lock-up index of abnormal working condition are normalized and standardized. Normalization is to convert the data into a unified standard range through certain mathematical operations, usually between 0 and 1, which helps to eliminate the influence of dimension and make various indicators comparable under the same standard. Through normalization, the dimension difference of the lock-up index under each working condition is eliminated to ensure that it is within the same dimension range, which is convenient for subsequent weighted calculation. Standardization is to process the data so that the data is converted into a standard normal distribution with a mean of 0 and a variance of 1, which helps to eliminate the unit and scale differences of the data and make different variables comparable under the same standard. Through standardization, the scale difference of the lock-up index of different working conditions is eliminated, making the index of each working condition more comparable, and its distribution is a standard normal distribution.

[0048] The standardized or normalized values ​​are combined according to a predetermined ratio for weighted calculation to obtain the final target locking index. Usually, the standard working condition locking index accounts for the highest proportion. A simple weighted average method is used to combine the results of each index to obtain the target locking index. By comprehensively considering the results of standard, extreme and abnormal working conditions, the locking quality of the nut is comprehensively evaluated. According to the analysis results under different working conditions, potential locking problems are discovered in advance and optimized to ensure the stability and reliability of the locking under various working conditions.

[0049] Furthermore, the present application also includes the following steps: An initial ratio is determined according to the occurrence frequency and duration of the standard test scenario parameters, the extreme test scenario parameters and the abnormal test scenario parameters; a deviation ratio is calculated based on the standard operating condition lock index, the extreme operating condition lock index and the abnormal operating condition lock index; if the deviation ratio does not meet the predetermined ratio constraint, the initial ratio is adjusted based on the deviation ratio to obtain the predetermined ratio.

[0050] Specifically, the frequency and duration of standard working conditions, extreme working conditions and abnormal working conditions are collected and counted. The frequency of occurrence refers to the frequency of a specific working condition, which is usually used to measure the frequency of a certain working condition in the entire production process; the duration refers to the length of time the working condition lasts, indicating the duration of the impact of the working condition on the whole. According to the frequency of occurrence and duration of the working condition, the initial ratio is determined to indicate the relative importance or impact of different working conditions in the total lock-in process.

[0051] By calculating the average values ​​of the lock index of standard working condition, the lock index of extreme working condition and the lock index of abnormal working condition, and then calculating the deviation ratio of each working condition according to the deviation ratio formula, that is, the absolute value of the difference between each working condition and the mean, divided by the mean, the deviation ratio of each working condition is obtained. The deviation ratio is used to measure the degree of deviation between the lock performance of each working condition and the overall performance. The calculation method of the deviation ratio is to compare the score of a single working condition with the mean of all working condition scores to obtain the relative deviation between the working condition and the average value.

[0052] The predetermined ratio is an adjusted ratio that takes into account the correction of the initial ratio by the deviation ratio. By calculating the deviation ratio and adjusting the initial ratio when the predetermined standard is not met, the most suitable ratio is finally obtained. Determine whether the deviation ratio meets the predetermined ratio constraint. For example, the predetermined deviation ratio threshold is 0.15. If the deviation ratio exceeds this value, the initial ratio needs to be adjusted. Adjust the ratio of the working condition by comparing the initial ratio with the actual deviation ratio. Combined with the adjustment result of the deviation ratio, calculate and obtain the predetermined ratio. Ensure that the adjusted ratio meets the actual needs of the locking task, ensure the stability of the locking quality, and avoid excessive deviation from a certain working condition. By calculating and adjusting the deviation ratio, ensure that the locking quality of standard working conditions, extreme working conditions, and abnormal working conditions can be balanced, avoiding overall quality instability due to excessive emphasis on a certain working condition.

[0053] Furthermore, the present application also includes the following steps: According to the standard test results, standard torque test results, standard thread engagement test results and standard stress distribution test results are extracted; based on the standard torque test results, a standard torque fluctuation curve is drawn; the standard torque fluctuation curve is segmented according to a sliding time window to obtain a plurality of standard torque fluctuation intervals; a differential calculation is performed on the plurality of standard torque fluctuation intervals to obtain a plurality of standard variations, and the mean of the plurality of standard variations is calculated to obtain a standard torque index; the ratio of the thread contact area in the standard thread engagement test results to the preset contact area is calculated to obtain a standard thread engagement index; based on the ratio of the maximum stress value to the average stress value in the standard stress distribution test results, a standard stress distribution index is obtained; a weighted calculation is performed on the standard torque index, the standard thread engagement index and the standard stress distribution index to obtain the standard working condition locking index.

[0054] Specifically, the standard torque test results refer to the torque data measured when the nut is locked under standard working conditions. It usually includes torque values ​​at different time points, reflecting the amount of force applied during the nut locking process. The torque data at different time points are extracted from the standard torque test results, and a curve of torque change over time is drawn to reflect the fluctuations in the torque application process. The standard torque fluctuation curve is divided according to the sliding time window, and the data in each time window represents a fluctuation interval. The sliding time window is a common data analysis method. By setting a time window of a fixed length, the window is gradually slid for data analysis, the data trend over time is observed, and the data in each window is processed.

[0055] The standard torque fluctuation interval refers to the periodic fluctuation interval of torque changes in the standard torque fluctuation curve, which is usually divided by the sliding time window method, and the torque data in each window is regarded as a fluctuation interval. The sliding step size refers to the length of time the window moves each time it slides. The size of the step size will also affect the results of the analysis. If the step size is too short, many overlapping windows may be generated, resulting in over-analysis of the data; if the step size is too long, important fluctuation features may be missed. Usually the step size is set to be equal to the window size so that there is no overlap between each window. Apply the sliding time window to the torque fluctuation curve, starting from the beginning of the data, and sliding one step each time until the end of the data. Each slide will generate a new subsequence, that is, a standard torque fluctuation interval.

[0056] Perform differential calculation on each standard torque fluctuation interval to obtain the change. The differential calculation is the change amplitude of the torque in each fluctuation interval. By calculating the change of multiple standard fluctuation intervals, the average is obtained, and finally the standard torque index is obtained. The average of the standard change is the standard torque index. The standard torque index reflects the stability of the torque change. The smaller the value, the smaller the torque fluctuation and the more stable the locking process.

[0057] The standard thread engagement test result refers to the meshing state test result between the nut and the bolt under standard working conditions. It mainly measures the size of the thread contact area and the integrity of the meshing, and is used to evaluate whether the nut is properly meshed to avoid loosening or falling off due to poor meshing. According to the standard thread engagement test results, the ratio of the thread contact area to the preset contact area is calculated to obtain the thread engagement index, which reflects the meshing quality between the nut and the bolt. The closer the value is to 1, the more perfect the meshing is.

[0058] The standard stress distribution test results refer to the stress distribution data generated during the nut locking process under standard working conditions. The stress at different positions of the nut is usually measured to evaluate the stress concentration area and the overall stress distribution to ensure that the nut will not be damaged due to excessive stress during the locking process. The ratio of the maximum stress to the average stress in the standard stress distribution test results is calculated to obtain the standard stress distribution index, which reflects the uniformity of the stress distribution. The smaller the standard stress distribution index, the more uniform the stress distribution and the more stable the nut locking process.

[0059] According to actual needs, determine the weights of torque, thread engagement, and stress. Usually, torque accounts for the largest proportion. The standard torque index, standard thread engagement index, and standard stress distribution index are combined through weighted calculation to obtain the standard working condition locking index. The standard working condition locking index is the sum of the weighted results. By weighted calculation of different indexes, the locking performance under standard working conditions is comprehensively evaluated from multiple aspects, including torque control, thread engagement quality, and uniformity of stress distribution. Through the comprehensive analysis of the torque fluctuation curve, thread engagement index, and stress distribution index, the performance of the nut locking process under standard working conditions is accurately evaluated, and torque fluctuations are timely discovered and optimized, unnecessary fluctuations are reduced, and the stability and quality of the locking process are improved.

[0060] Furthermore, the present application also includes the following steps: According to multiple nut types, multiple nut locking tools and multiple nut assembly stations are obtained from the historical nut locking scheme set; a first nut type is extracted from the multiple nut types; a first nut simulation channel is configured according to the first nut type and the corresponding first nut locking tool and the first nut assembly station; and so on, multiple nut simulation channels are configured based on the multiple nut types; the first nut simulation channel and the multiple nut simulation channels are connected in parallel to complete the construction of the nut simulation model.

[0061] Specifically, multiple nut types refer to nuts of different specifications, sizes and materials used in production. Each nut type may have different geometric shapes, thread characteristics, locking requirements, etc. Multiple nut locking tools and multiple nut assembly stations corresponding to multiple nut types are extracted from the historical nut locking scheme set. Each nut type has its specific tool and station requirements. A nut type is randomly selected, and the first nut simulation channel is configured according to its corresponding first nut locking tool and first nut assembly station. Use simulation software (such as ANSYS, ABAQUS, etc.) to configure the simulation channel related to the first nut type, perform meshing, and set simulation parameters, including material properties, boundary conditions, etc. Considering factors such as the geometry and torque requirements of the first nut type, the behavior of the nut under different conditions, including torque, stress, thread engagement, etc., is predicted.

[0062] According to the same steps as above, different simulation channels are configured for different types of nuts according to their characteristics such as material, geometric dimensions and thread structure. Each simulation channel involves different locking tools, workstations and parameter settings. Once the simulation channels are configured for different nut types, multiple nut simulation channels are connected in parallel to form a complete nut simulation model. Ensure that each nut simulation channel can run independently, but can be processed in parallel in the overall simulation model. Through parallel computing, simulation results of multiple nut types can be obtained at the same time, saving time and improving efficiency. By connecting multiple nut simulation channels in parallel, multiple nut types can be processed at the same time, and simulations of multiple nut types are carried out simultaneously, quickly identifying problems of different types of nuts in the locking process, optimizing in advance, and improving the locking quality.

[0063] Furthermore, the present application S500 includes: Based on the target nut type, a historical target locking data set is obtained from the historical nut locking scheme set to construct a locking parameter optimization space; maximizing locking accuracy, maximizing thread engagement area and minimizing stress concentration are used as the multivariate optimization goals of the target nut; based on the multivariate optimization goals, multi-objective optimization is performed in the locking parameter optimization space to determine multiple locking optimization strategies; the importance of the multivariate optimization goals is evaluated to obtain multivariate goal importance evaluation coefficients; weights are assigned to the multivariate optimization goals according to the multivariate goal importance evaluation coefficients, and multiple comprehensive evaluation results are obtained by combining the multiple locking optimization strategies; based on the multiple comprehensive evaluation results, the optimal locking strategy is selected from the multiple locking optimization strategies.

[0064] Specifically, all historical target locking data sets related to the target nut are extracted from the historical nut locking scheme set, including parameters such as locking tools, torque settings, and assembly stations. Based on this, the locking parameter optimization space is constructed, including all possible locking parameter combinations, such as torque, locking tools, assembly stations, locking sequences, and other parameters. Through the optimization algorithm, the best locking strategy can be found in this space. Multivariate optimization objectives refer to considering multiple objectives in the optimization process, such as locking accuracy, thread engagement area, and stress concentration. These objectives are usually interrelated, but there may be certain conflicts, so it is necessary to find a balance between multiple objectives.

[0065] According to the target nut type, set the multivariate optimization objectives of locking, including maximizing locking accuracy, maximizing thread engagement area and minimizing stress concentration. Maximizing locking accuracy ensures accurate torque control and avoids excessive or insufficient tightening; maximizing thread engagement area is to improve the thread contact quality and ensure the strength and stability after locking; minimizing stress concentration avoids excessive stress at the contact point between the nut and the screw and reduces the risk of thread damage. Use multi-objective optimization algorithms (such as genetic algorithms, particle swarm algorithms, etc.) to search the optimization space of locking parameters and find the optimal combination of locking strategies. Multi-objective optimization will consider the relationship between torque accuracy, thread engagement area and stress concentration.

[0066] The importance of each optimization target is evaluated and the weights between different targets are determined. Usually, locking accuracy and thread engagement area are considered more important than stress concentration. Weights are assigned to multiple locking optimization strategies according to the importance evaluation coefficient of multiple targets, and then a comprehensive evaluation is performed based on the simulation results of multiple strategies to calculate the comprehensive score of each strategy. Based on the comprehensive evaluation results of multiple optimization strategies, the strategy with the highest score is selected as the optimal strategy for target locking. By comprehensively considering multiple optimization targets, the quality of locking is significantly improved, including improving torque control accuracy, thread engagement quality and reducing stress concentration, ensuring that the nut can be stably locked under different working conditions, and improving the reliability and consistency of nut locking during the production process.

[0067] Furthermore, the present application also includes the following steps: According to the mapping relationship between the multiple locking optimization strategies and the multiple comprehensive evaluation results, a Gaussian process regression model is constructed; based on the multiple locking optimization strategies, uniform random sampling is performed, and multiple candidate strategies are screened through the multivariate optimization objectives to generate multiple candidate strategies; combining the multiple locking optimization strategies and the multiple candidate strategies, multiple evaluation values ​​and multiple variance values ​​are obtained through the Gaussian process regression model; a preset evaluation function is introduced, and multiple comprehensive scores are obtained by combining the multiple evaluation values ​​and the multiple variance values; the locking optimization strategy corresponding to the maximum comprehensive score among the multiple comprehensive scores is taken as the optimal locking strategy.

[0068] Specifically, a mapping relationship is established between multiple locking optimization strategies and their corresponding multiple comprehensive evaluation results, and the locking strategies and their corresponding evaluation results (such as torque accuracy, thread engagement, etc.) are used as training data to construct a Gaussian process regression model. Gaussian process regression is a probability-based non-parametric regression method that can model complex functional relationships and is widely used in prediction and optimization problems. The effect of the locking strategy is predicted based on the input features of multiple locking optimization strategies (such as torque, tool type, etc.), and the evaluation value and uncertainty (variance) of each strategy are provided.

[0069] Through the known multiple locking optimization strategies and comprehensive evaluation results, a mapping relationship is established to predict the performance of other untested strategies, thereby helping to select the optimal strategy. Through the Gaussian process regression model, the parameters of each locking optimization strategy (such as torque, tool type, assembly station, etc.) are input, and the comprehensive evaluation score corresponding to the strategy (such as locking accuracy, thread engagement accuracy, stress distribution, etc.) is output. The Gaussian process regression model continuously learns the relationship between each input parameter (such as torque, tool type, station, etc.) and the output result (such as locking accuracy, thread engagement, etc.). The training process involves selecting a suitable kernel function (such as radial basis function, RBF), which is used to calculate the similarity between different sample points to predict the output value of new input data. The Gaussian process regression model also calculates the variance to represent the uncertainty of the prediction result. Use part of the data set as a validation set to test the accuracy of the model. The validation set does not participate in the training process, but is used to evaluate the performance of the model on unknown data. By comparing with the actual test results, the parameters of the model, such as the kernel function type, length scale, etc., can be adjusted to improve the accuracy and robustness of the model.

[0070] Based on the multiple locking optimization strategies that have been optimized, each strategy is uniformly randomly sampled through the Gaussian process regression model to generate multiple candidate strategies. Uniform random sampling is a process of randomly extracting samples from a specified range to ensure that the probability of occurrence of all samples is equal, and is used to explore a wide range of solution spaces. The purpose of uniform random sampling is to select a set of sample strategies from all possible locking strategy combinations to ensure that a variety of candidate strategies are obtained from the entire strategy space, avoid bias in selecting strategies, and make the final optimization process more global. These candidate strategies are screened using multivariate optimization objectives (such as maximizing locking accuracy, maximizing thread engagement area, and minimizing stress concentration). Through multivariate optimization objective screening, a strategy combination that performs well under all optimization objectives is selected, which involves weighing multiple performance indicators (such as locking accuracy, thread engagement, stress analysis, etc.) and screening out strategies that meet the multivariate objectives.

[0071] By further comprehensively evaluating the generated candidate strategies, the most suitable strategy is selected as the final lock-in solution. Using the Gaussian process regression model, the evaluation value and variance value of each strategy are obtained by predicting multiple lock-in optimization strategies and multiple candidate strategies. The evaluation value indicates the effect of the strategy under each optimization goal, and the variance value indicates the reliability of the prediction result. By comparing the evaluation values ​​and variance values ​​of different strategies, the most promising strategies are screened out. Generally speaking, strategies with higher evaluation values ​​and smaller variances are more likely to be the optimal solution.

[0072] The preset evaluation function is a predefined mathematical function or algorithm, which is used to comprehensively evaluate different strategies or solutions according to different evaluation criteria (such as evaluation value, variance value, etc.) to obtain a comprehensive score. By combining multiple factors together, a comprehensive score is provided for each strategy to compare the advantages and disadvantages of different solutions. Combined with the preset evaluation function, multiple evaluation values ​​and variance values ​​are input into the evaluation function to obtain the comprehensive score of each candidate strategy. The preset evaluation function normalizes the variance value, smoothes it by 1, and then calculates it with the evaluation value. That is to say, after smoothing the normalized variance value by adding 1, the ratio of the evaluation value to the smoothed variance value is obtained to obtain the comprehensive score corresponding to each strategy. By evaluating the comprehensive scores of all candidate strategies, the strategy with the highest score is selected as the final optimal locking strategy, which is the most outstanding in the comprehensive performance of multiple objectives, and the credibility of its prediction results is also high. Through the prediction of the Gaussian process regression model and the screening of multivariate optimization objectives, the locking accuracy is significantly improved, ensuring that the locking of each nut can achieve the predetermined accuracy and stability.

[0073] In summary, the multi-specification nut quick locking method based on 3D vision provided by the present application has the following technical effects: The three-dimensional point cloud data of the target nut is obtained by scanning the built-in 3D vision module of the multi-axis robot; the nut type classifier determines the target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identification; the historical nut locking scheme set is interactively obtained, and the target nut assembly station, target nut locking tool and target nut locking parameter are traversed in the historical nut locking scheme set according to the nut coding identification, and the target nut locking strategy is formed; the nut simulation model performs locking simulation through the target nut locking strategy, and performs quality analysis on the simulation results to obtain the target locking index; if the target locking index does not meet the preset standard, the target nut locking strategy is optimized to obtain the target locking optimal strategy; the multi-axis robot automatically locks the target nut based on the target locking optimal strategy. In other words, the specification type of the nut is identified by the 3D vision module, the corresponding nut locking strategy is selected from the historical nut locking scheme set, the target nut locking strategy is optimized through simulation, and the target nut is automatically locked, thereby improving the accuracy and efficiency of nut locking.

[0074] Embodiment 2, based on the same inventive concept as the method for fast locking nuts of multiple specifications based on 3D vision in the aforementioned embodiment 1, the present application also provides a fast locking device for nuts of multiple specifications based on 3D vision, please refer to the attached Figure 2 The multi-specification nut quick locking device based on 3D vision includes: A point cloud data acquisition module 11, the point cloud data acquisition module 11 is used to scan and acquire the three-dimensional point cloud data of the target nut through the built-in 3D vision module of the multi-axis robot; a nut type determination module 12, the nut type determination module 12 is used for a nut type classifier to determine the target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identification; a locking strategy determination module 13, the locking strategy determination module 13 is used to interactively acquire a historical nut locking solution set, and traverse the historical nut locking solution set according to the nut coding identification to obtain the target nut assembly station, the target nut locking station, and the target nut locking station. The locking parameters of the target nut and the target nut are combined to form a target nut locking strategy; a locking simulation module 14, wherein the locking simulation module 14 is used for the nut simulation model to perform locking simulation through the target nut locking strategy, and perform quality analysis on the simulation result to obtain a target locking index; a locking strategy optimization module 15, wherein the locking strategy optimization module 15 is used to optimize the target nut locking strategy if the target locking index does not meet the preset standard to obtain the target locking optimal strategy; an automatic locking module 16, wherein the automatic locking module 16 is used for the multi-axis robot to automatically lock the target nut based on the target locking optimal strategy.

[0075] Furthermore, the locking simulation module 14 in the 3D vision-based multi-specification nut fast locking device is also used for: Based on the target nut type, the nut simulation model is traversed to determine the corresponding target nut simulation channel; based on the target nut locking strategy, the parameters of the target nut simulation channel are updated; multiple test scenario parameters are obtained, and locking simulation is performed on the target nut simulation channel in turn through the multiple test scenario parameters to obtain multiple simulation results; quality analysis is performed on the multiple simulation results to obtain the target locking index.

[0076] Furthermore, the locking simulation module 14 in the 3D vision-based multi-specification nut fast locking device is also used for: According to the multiple test scenario parameters, standard test scenario parameters, extreme test scenario parameters and abnormal test scenario parameters are extracted; standard test results corresponding to the standard test scenario parameters are extracted from the multiple simulation results; torque analysis, thread engagement analysis and stress analysis are performed on the standard test results to obtain a standard operating condition locking index; based on the extreme test scenario parameters, an extreme operating condition locking index is obtained; based on the abnormal test scenario parameters, an abnormal operating condition locking index is obtained; a predetermined ratio is obtained, and the target locking index is obtained by combining the standard operating condition locking index, the extreme operating condition locking index and the abnormal operating condition locking index.

[0077] Furthermore, the locking simulation module 14 in the 3D vision-based multi-specification nut fast locking device is also used for: An initial ratio is determined according to the occurrence frequency and duration of the standard test scenario parameters, the extreme test scenario parameters and the abnormal test scenario parameters; a deviation ratio is calculated based on the standard operating condition lock index, the extreme operating condition lock index and the abnormal operating condition lock index; if the deviation ratio does not meet the predetermined ratio constraint, the initial ratio is adjusted based on the deviation ratio to obtain the predetermined ratio.

[0078] Furthermore, the locking simulation module 14 in the 3D vision-based multi-specification nut fast locking device is also used for: According to the standard test results, standard torque test results, standard thread engagement test results and standard stress distribution test results are extracted; based on the standard torque test results, a standard torque fluctuation curve is drawn; the standard torque fluctuation curve is segmented according to a sliding time window to obtain a plurality of standard torque fluctuation intervals; a differential calculation is performed on the plurality of standard torque fluctuation intervals to obtain a plurality of standard variations, and the mean of the plurality of standard variations is calculated to obtain a standard torque index; the ratio of the thread contact area in the standard thread engagement test results to the preset contact area is calculated to obtain a standard thread engagement index; based on the ratio of the maximum stress value to the average stress value in the standard stress distribution test results, a standard stress distribution index is obtained; a weighted calculation is performed on the standard torque index, the standard thread engagement index and the standard stress distribution index to obtain the standard working condition locking index.

[0079] Furthermore, the locking simulation module 14 in the 3D vision-based multi-specification nut fast locking device is also used for: According to multiple nut types, multiple nut locking tools and multiple nut assembly stations are obtained from the historical nut locking scheme set; a first nut type is extracted from the multiple nut types; a first nut simulation channel is configured according to the first nut type and the corresponding first nut locking tool and the first nut assembly station; and so on, multiple nut simulation channels are configured based on the multiple nut types; the first nut simulation channel and the multiple nut simulation channels are connected in parallel to complete the construction of the nut simulation model.

[0080] Furthermore, the locking strategy optimization module 15 in the 3D vision-based multi-specification nut fast locking device is also used for: Based on the target nut type, a historical target locking data set is obtained from the historical nut locking scheme set to construct a locking parameter optimization space; maximizing locking accuracy, maximizing thread engagement area and minimizing stress concentration are used as the multivariate optimization goals of the target nut; based on the multivariate optimization goals, multi-objective optimization is performed in the locking parameter optimization space to determine multiple locking optimization strategies; the importance of the multivariate optimization goals is evaluated to obtain multivariate goal importance evaluation coefficients; weights are assigned to the multivariate optimization goals according to the multivariate goal importance evaluation coefficients, and multiple comprehensive evaluation results are obtained by combining the multiple locking optimization strategies; based on the multiple comprehensive evaluation results, the optimal locking strategy is selected from the multiple locking optimization strategies.

[0081] Furthermore, the locking strategy optimization module 15 in the 3D vision-based multi-specification nut fast locking device is also used for: According to the mapping relationship between the multiple locking optimization strategies and the multiple comprehensive evaluation results, a Gaussian process regression model is constructed; based on the multiple locking optimization strategies, uniform random sampling is performed, and multiple candidate strategies are screened through the multivariate optimization objectives to generate multiple candidate strategies; combining the multiple locking optimization strategies and the multiple candidate strategies, multiple evaluation values ​​and multiple variance values ​​are obtained through the Gaussian process regression model; a preset evaluation function is introduced, and multiple comprehensive scores are obtained by combining the multiple evaluation values ​​and the multiple variance values; the locking optimization strategy corresponding to the maximum comprehensive score among the multiple comprehensive scores is taken as the optimal locking strategy.

[0082] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The multi-specification nut quick locking method based on 3D vision and the specific examples in the first embodiment are also applicable to the multi-specification nut quick locking device based on 3D vision in this embodiment. Through the above detailed description of the multi-specification nut quick locking method based on 3D vision, those skilled in the art can clearly know the multi-specification nut quick locking device based on 3D vision in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0083] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0084] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A method for fast locking nuts of multiple specifications based on 3D vision, characterized in that: include: The built-in 3D vision module of the multi-axis robot is used to scan and obtain the 3D point cloud data of the target nut; A nut type classifier determines a target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identification; Interactively obtain a historical nut locking solution set, traverse the historical nut locking solution set according to the nut coding identifier to obtain a target nut assembly station, a target nut locking tool and a target nut locking parameter, and form a target nut locking strategy; The nut simulation model performs locking simulation through the target nut locking strategy, and performs quality analysis on the simulation results to obtain a target locking index; If the target locking index does not meet the preset standard, the target nut locking strategy is optimized to obtain the target locking optimal strategy; The multi-axis robot automatically locks the target nut based on the target locking optimal strategy.

2. The method for fast locking nuts of multiple specifications based on 3D vision according to claim 1 is characterized in that: The nut simulation model performs locking simulation through the target nut locking strategy, and performs quality analysis on the simulation results to obtain the target locking index, including: Based on the target nut type, traverse the nut simulation model to determine the corresponding target nut simulation channel; Based on the target nut locking strategy, updating parameters of the target nut simulation channel; Acquire multiple test scenario parameters, and perform locking simulation on the target nut simulation channel in sequence according to the multiple test scenario parameters to obtain multiple simulation results; Performing quality analysis on the multiple simulation results to obtain the target lock index.

3. The method for fast locking nuts of multiple specifications based on 3D vision according to claim 2 is characterized in that: Performing quality analysis on the multiple simulation results to obtain the target lock index includes: Extracting standard test scenario parameters, extreme test scenario parameters and abnormal test scenario parameters according to the multiple test scenario parameters; Extracting the standard test results corresponding to the standard test scenario parameters from the multiple simulation results; Perform torque analysis, thread engagement analysis and stress analysis on the standard test results to obtain a standard working condition locking index; Based on the extreme test scenario parameters, an extreme working condition locking index is obtained; Based on the abnormal test scenario parameters, an abnormal working condition lock index is obtained; A predetermined ratio is obtained, and the target locking index is obtained by combining the standard operating condition locking index, the extreme operating condition locking index and the abnormal operating condition locking index.

4. The method for fast locking nuts of multiple specifications based on 3D vision according to claim 3 is characterized in that: The obtaining of the predetermined ratio comprises: Determining an initial ratio according to the occurrence frequency and duration length of the standard test scenario parameters, the extreme test scenario parameters, and the abnormal test scenario parameters; Calculating a deviation ratio based on the standard operating condition lock index, the extreme operating condition lock index, and the abnormal operating condition lock index; If the deviation ratio does not meet the predetermined ratio constraint, the initial ratio is adjusted based on the deviation ratio to obtain the predetermined ratio.

5. The method for fast locking nuts of multiple specifications based on 3D vision according to claim 3 is characterized in that: The standard test results are subjected to torque analysis, thread engagement analysis and stress analysis to obtain the standard working condition locking index, including: According to the standard test results, extracting standard torque test results, standard thread engagement test results and standard stress distribution test results; Based on the standard torque test results, drawing a standard torque fluctuation curve; Segmenting the standard torque fluctuation curve according to the sliding time window to obtain a plurality of standard torque fluctuation intervals; Performing differential calculation on the plurality of standard torque fluctuation intervals to obtain a plurality of standard variations, and calculating the average of the plurality of standard variations to obtain a standard torque index; Calculating the ratio of the thread contact area in the standard thread engagement test result to the preset contact area to obtain a standard thread engagement index; Obtaining a standard stress distribution index based on a ratio of a maximum stress value to an average stress value in the standard stress distribution test result; The standard torque index, the standard thread engagement index and the standard stress distribution index are weightedly calculated to obtain the standard working condition locking index.

6. The method for fast locking nuts of multiple specifications based on 3D vision according to claim 2 is characterized in that: Traversing the nut simulation model to determine the corresponding target nut simulation channel includes: According to the multiple nut types, multiple nut locking tools and multiple nut assembly stations are acquired in the historical nut locking solution set; extracting a first nut type from the plurality of nut types; According to the first nut type and the corresponding first nut locking tool and the first nut assembly station, a first nut simulation channel is configured; By analogy, multiple nut simulation channels are configured based on the multiple nut types; The first nut simulation channel and the multiple nut simulation channels are connected in parallel to complete the construction of the nut simulation model.

7. The method for fast locking nuts of multiple specifications based on 3D vision according to claim 1 is characterized in that: The target nut locking strategy is optimized to obtain the target locking optimal strategy, including: Based on the target nut type, a historical target locking data set is obtained in the historical nut locking solution set to construct a locking parameter optimization space; Taking maximizing locking accuracy, maximizing thread engagement area and minimizing stress concentration as the multivariate optimization objectives of the target nut; Based on the multi-objective optimization objectives, a multi-objective optimization search is performed in the locking parameter optimization space to determine a plurality of locking optimization strategies; Performing importance evaluation on the multivariate optimization objectives to obtain multivariate objective importance evaluation coefficients; The weights of the multiple optimization objectives are assigned according to the multiple objective importance evaluation coefficients, and multiple comprehensive evaluation results are obtained by combining the multiple lock-in optimization strategies; Based on the multiple comprehensive evaluation results, the optimal locking strategy is selected from the multiple locking optimization strategies.

8. The method for fast locking nuts of multiple specifications based on 3D vision according to claim 7 is characterized in that: Based on the multiple comprehensive evaluation results, selecting the optimal locking strategy from the multiple locking optimization strategies includes: Constructing a Gaussian process regression model according to the mapping relationship between the multiple lock optimization strategies and the multiple comprehensive evaluation results; Performing uniform random sampling based on the multiple lock-payment optimization strategies and screening through the multivariate optimization objectives to generate multiple candidate strategies; Combining the multiple lock-in optimization strategies and the multiple candidate strategies, and obtaining multiple evaluation values ​​and multiple variance values ​​through the Gaussian process regression model; Introducing a preset evaluation function, combining the multiple evaluation values ​​and the multiple variance values, to obtain multiple comprehensive scores; The lock-payment optimization strategy corresponding to the maximum comprehensive score among multiple comprehensive scores is taken as the optimal lock-payment strategy.

9. A multi-specification nut quick locking device based on 3D vision, characterized in that: The method for quickly locking nuts with multiple specifications based on 3D vision according to any one of claims 1 to 8 is used, wherein the device for quickly locking nuts with multiple specifications based on 3D vision comprises: A point cloud data acquisition module, wherein the point cloud data acquisition module is used to scan and acquire three-dimensional point cloud data of a target nut through a built-in 3D vision module of a multi-axis robot; A nut type determination module, wherein the nut type determination module is used for a nut type classifier to determine a target nut type of the target nut based on the three-dimensional point cloud data, wherein the target nut type has a nut coding identification; A locking strategy determination module, the locking strategy determination module is used to interactively obtain a historical nut locking solution set, traverse the historical nut locking solution set according to the nut coding identifier to obtain a target nut assembly station, a target nut locking tool and a target nut locking parameter, and form a target nut locking strategy; A locking simulation module, wherein the locking simulation module is used for the nut simulation model to perform locking simulation through the target nut locking strategy, and to perform quality analysis on the simulation result to obtain a target locking index; A locking strategy optimization module, wherein the locking strategy optimization module is used to optimize the target nut locking strategy to obtain the target locking optimal strategy if the target locking index does not meet the preset standard; An automatic locking module is used by the multi-axis robot to automatically lock the target nut based on the target locking optimal strategy.

Images