An assembly detection method for automobile tie rod structure

Abstract

The invention discloses an assembly detection method for an automobile tie rod structure, comprising step S1, obtaining a standard assembly image before riveting, processing to obtain a standard ball pin coordinate, step S2, calculating to obtain a comprehensive hardness parameter and a comprehensive elasticity parameter; step S4, obtaining a dynamic assembly video, a rotation speed and a riveting pressure value during riveting, and obtaining a ball pin processing coordinate according to the dynamic assembly video processing; step S4, obtaining a real-time offset angle according to the ball pin standard coordinate and the ball pin processing coordinate processing, step S5, inputting the real-time offset angle, each riveting pressure value, the rotation speed, the comprehensive hardness parameter, and the comprehensive elasticity parameter into a ball pin comprehensive deformation prediction model, predicting to obtain a ball pin comprehensive deformation score, step S6, generating an assembly alarm instruction when the ball pin comprehensive deformation score is greater than a deformation threshold, and adjusting the riveting output state of the riveting machine. The invention improves the safety of the ball rivet assembly during the riveting assembly process.

Description

Technical Field

[0001] The invention relates to the technical field of automobile tie rods, and in particular to an assembly detection method for an automobile tie rod structure. Background Art

[0002] The automobile tie rod structure is a mechanical structure used to connect and support various components of the automobile. It is usually used in the suspension system, steering system, and the connection between the body and chassis. The main function of the tie rod is to transmit force and movement to ensure the stability and controllability of the vehicle during driving.

[0003] During the assembly process of the automobile tie rod structure, the ball pin assembly needs to be accurately installed into the ball seat structure of the automobile tie rod, and then the ball seat is fixed by riveting with a riveting machine. During the riveting assembly process of the ball pin assembly, the originally assembled ball head pin is prone to deflection, causing the ball head pin to be separated from the center of the ball seat. At the same time, the pressing speed and pressure during the riveting process are not appropriate, which will cause the internal ball pin assembly to deform, affecting the life of the ball pin assembly. In the prior art, problems that occur during the assembly process are generally detected through manual observation. This method has extremely low detection efficiency and cannot promptly detect problems and adjust the state of the riveting machine during the riveting process, resulting in the ball pin assembly being easily damaged during the riveting assembly process. Summary of the invention

[0004] In view of the deficiencies in the prior art, an object of the present invention is to provide an assembly detection method for an automobile tie rod structure, so as to improve the safety of a ball rivet assembly during a rotary riveting assembly process.

[0005] To achieve the above object, the present invention provides the following technical solutions: an automobile tie rod structure, comprising a tie rod assembly, wherein the tie rod assembly comprises an integrally formed tie rod portion and two ball seat portions, wherein the two ball seat portions are located at two ends of the tie rod portion, wherein an assembled ball pin assembly is arranged in the ball seat portion, wherein the ball pin assembly comprises a ball stud, an oil nozzle, a first ball bowl, a second ball bowl, a sleeve, a rubber ring, and an end cover;

[0006] The ball stud pin includes a ball head portion and a pin rod portion, the end cover covers the upper end of the ball head portion, an oil inlet hole is also opened in the top center of the ball head portion, the oil nozzle passes through the end cover from the upper end of the end cover and is connected to the oil inlet hole, the first ball bowl is sleeved on the top outer side of the ball head portion, the second ball bowl is sleeved on the middle outer side of the ball head portion, the sleeve is sleeved on the top outer side of the pin rod portion, and the rubber ring is sleeved on the outside of the sleeve.

[0007] Furthermore, the ball and pin assembly also includes a dust cover and a gasket, the dust cover is sleeved on the outer side of the middle part of the pin rod, the upper end surface of the dust cover is fitted and abutted against the lower end surface of the sleeve, and the gasket is sleeved on the lower end of the top edge of the dust cover.

[0008] An assembly detection method, applied to the automobile tie rod structure as described above, comprises:

[0009] Step S1, before the riveting machine performs riveting processing on the ball pin assembly, a standard assembly image of the ball pin assembly is photographed and hardness parameters and elastic parameters of various parts of the ball pin assembly are input simultaneously, and the standard coordinates of the ball pin of the ball pin assembly in a three-dimensional coordinate system are obtained according to the standard assembly image processing;

[0010] Step S2, inputting the hardness parameter and the elasticity parameter of each part into a preset comprehensive hardness calculation formula and a preset comprehensive elasticity calculation formula, and calculating the comprehensive hardness parameter and the comprehensive elasticity parameter respectively;

[0011] Step S3, when the riveting machine performs riveting processing on the ball pin assembly, a dynamic assembly video of the ball pin assembly is recorded, and at the same time, a rotation speed of a rotation output end of the riveting machine is detected, and riveting pressure values ​​at each riveting pressing point between the riveting machine and the ball seat are detected, and a plurality of riveting processing images are obtained according to the splitting of the dynamic assembly video, and then the ball pin processing coordinates of the ball pin assembly in a three-dimensional coordinate system are obtained according to the riveting processing images;

[0012] Step S4, obtaining the real-time offset angle of the ball pin assembly according to the standard ball pin coordinates and the processing ball pin coordinates in the same three-dimensional coordinate system;

[0013] Step S5, pre-training a ball pin comprehensive deformation prediction model, inputting the real-time offset angle, each of the rivet pressure values, the rotation speed, the comprehensive hardness parameter, and the comprehensive elasticity parameter into the ball pin comprehensive deformation prediction model, and predicting the ball pin comprehensive deformation score at the current moment;

[0014] Step S6, when the comprehensive deformation score of the ball pin is greater than a preset deformation threshold, an assembly alarm instruction is generated, and according to the assembly alarm instruction, an external party is warned of an assembly error of the ball pin assembly and the riveting output state of the riveting machine is adjusted.

[0015] Furthermore, the comprehensive hardness calculation formula is configured as:

[0016]

[0017] Among them, C hc Used to express the comprehensive hardness parameter, B s Used to indicate the hardness of the ball stud, B b1 Used to indicate the hardness of the first bowl, B b2 Used to indicate the hardness of the second bowl, S lUsed to indicate sleeve hardness, R r Used to indicate the hardness of rubber ring, E c Used to indicate the hardness of the end cap, ε 1 , ε 2 , ε 3 , ε 4 , ε 5 , ε 6 are used to represent a preset first hardness conversion coefficient, a second hardness conversion coefficient, a third hardness conversion coefficient, a fourth hardness conversion coefficient, a fifth hardness conversion coefficient and a sixth hardness conversion coefficient respectively;

[0018] The comprehensive elasticity calculation formula is configured as:

[0019]

[0020] Among them, C ec Used to express the comprehensive elastic parameter, B he Used to indicate the elasticity of the ball stud, B be1 Used to indicate the elasticity of the first bowl, B be2 Used to indicate the elasticity of the second bowl, S e Used to indicate sleeve elasticity, R re Used to indicate the elasticity of the rubber ring, E ce Used to represent the elasticity of the end cap, σ 1 , σ 2 , σ 3 , σ 4 , σ 5 , σ 6 They are respectively used to represent the preset first elastic conversion coefficient, second elastic conversion coefficient, third elastic conversion coefficient, fourth elastic conversion coefficient, fifth elastic conversion coefficient and sixth elastic conversion coefficient.

[0021] Furthermore, the step S5 comprises:

[0022] Step S51, storing training parameters of several historical moments in a preset ball pin riveting processing database, wherein the training parameters include historical offset angles, historical riveting pressure values, historical rotation speeds, historical comprehensive hardness parameters, historical comprehensive elastic parameters, and historical ball pin comprehensive deformation scores, and dividing the training parameters into a training set, a test set, and a verification set according to a proportion;

[0023] Step S52, introducing an initial model, taking the historical offset angle, the historical riveting pressure value, the historical rotation speed, the historical comprehensive hardness parameter and the historical comprehensive elasticity parameter in the training set as input, taking the historical ball pin comprehensive deformation score in the training set as output, and training to obtain an initial prediction model;

[0024] Step S53, using the training parameters in the validation set to adjust the parameters of the initial prediction model, using the training parameters in the test set to test the accuracy of the initial prediction model, and finally outputting it as the ball pin comprehensive deformation prediction model after the accuracy of the initial prediction model is greater than a preset accuracy threshold.

[0025] Furthermore, the initial model is an XLSTM model.

[0026] Furthermore, the data ratio in the training set, validation set, and test set is 3:1:1.

[0027] Furthermore, in the step S5, it also includes real-time detection of environmental parameters, wherein the environmental parameters include environmental temperature, environmental humidity, ultraviolet intensity and antioxidant concentration;

[0028] The weight parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ambient temperature and the ambient humidity to obtain optimized weight parameters, and the bias parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ultraviolet intensity and the antioxidant concentration to obtain optimized bias parameters, so as to obtain the optimized ball pin comprehensive deformation prediction model, and the real-time offset angle, each of the rivet pressure values, the rotation speed, the comprehensive hardness parameter, and the comprehensive elasticity parameter are input into the optimized ball pin comprehensive deformation prediction model to predict the ball pin comprehensive deformation score.

[0029] Furthermore, in step S5, the ambient temperature and the ambient humidity are input into a preset weight parameter optimization formula to calculate the optimized weight parameter, and the weight parameter optimization formula is configured as:

[0030]

[0031] Wherein, W' is used to represent the optimization weight parameter, W 0 It is used to represent the weight parameter before optimization, a t Used to indicate the ambient temperature, a h Used to represent the ambient humidity, χ 1 , χ 2 They are respectively used to represent the preset first temperature conversion coefficient and the first humidity conversion coefficient.

[0032] Further, in the step S5, the ultraviolet intensity and the antioxidant concentration are input into a preset bias parameter optimization formula to calculate the optimized bias parameter, and the bias parameter optimization formula is configured as:

[0033]

[0034] Wherein, B' is used to represent the optimized bias parameter, B 0 It is used to represent the bias parameters before optimization, U i Used to indicate the intensity of ultraviolet light, a c Used to express the antioxidant concentration, δ 1 , δ 2 They are respectively used to represent the preset first ultraviolet intensity conversion coefficient and the first antioxidant concentration conversion coefficient.

[0035] Beneficial effects of the present invention:

[0036] The present invention dynamically predicts the ball pin comprehensive deformation score of the ball pin assembly according to the calculated and detected real-time offset angle, various rivet pressure values, rotation speed, comprehensive hardness parameter and comprehensive elastic parameter during the assembly process of the ball pin assembly to reflect the deformation degree of the ball pin assembly. At the same time, the output state of the rivet machine is adjusted in time when the ball pin comprehensive deformation score is greater than the deformation threshold value, which can avoid the ball pin assembly from being greatly deformed during the rivet processing, thereby significantly improving the safety of the ball rivet assembly during the rivet assembly process. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the overall structure of the automobile tie rod structure of the present invention;

[0038] Figure 2 It is a structural schematic diagram of the ball pin assembly in the present invention;

[0039] Figure 3 It is a flow chart of the steps of the assembly detection method of the present invention;

[0040] Figure 4 It is a step flow chart of step S5 in the present invention.

[0041] Figure numerals: 1. tie rod assembly; 2. tie rod portion; 3. ball seat portion; 4. ball pin assembly; 41. ball stud; 42. oil nozzle; 43. first ball bowl; 44. second ball bowl; 45. sleeve; 46. rubber ring; 47. end cover; 48. dust cover; 49. gasket. DETAILED DESCRIPTION

[0042] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. The same parts are represented by the same reference numerals. It should be noted that the words "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to directions in the accompanying drawings, and the words "bottom surface" and "top surface", "inner" and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0043] Example 1, reference Figure 1 to Figure 2, which is the first embodiment of the present invention, and provides an automobile tie rod structure, which can improve the integration of the ball pin assembly 4. The automobile tie rod structure includes a tie rod assembly 1, the tie rod assembly 1 includes an integrally formed tie rod portion 2 and two ball seat portions 3, the two ball seat portions 3 are located at both ends of the tie rod portion 2, and the ball seat portion 3 is provided with an assembled ball pin assembly 4, the ball pin assembly 4 includes a ball head pin 41, an oil nozzle 42, a first ball bowl 43, a second ball bowl 44, a sleeve 45, a rubber ring 46, and an end cover 47;

[0044] The ball pin 41 includes a ball head and a pin rod, an end cover 47 covers the upper end of the ball head, an oil inlet hole is also opened in the top center of the ball head, an oil nozzle 42 passes through the end cover 47 from the upper end and is connected to the oil inlet hole, a first ball bowl 43 is sleeved on the top outer side of the ball head, a second ball bowl 44 is sleeved on the middle outer side of the ball head, a sleeve 45 is sleeved on the top outer side of the pin rod, and a rubber ring 46 is sleeved on the outer side of the sleeve 45.

[0045] Working principle of embodiment 1:

[0046] In this embodiment, the oil nozzle 42, the end cover 47, and the first ball bowl 43 are integrated on the top of the ball stud 41, the second ball bowl 44 is sleeved on the outside of the ball head part, the sleeve 45 is sleeved on the outside of the pin rod part, and finally the rubber ring 46 is sleeved on the outside of the sleeve 45, thereby achieving a high degree of integration of the ball and pin assembly 4, which is beneficial to reducing the overall volume of the ball and pin assembly 4.

[0047] Preferably, the ball pin assembly 4 also includes a dust cover 48 and a washer 49. The dust cover 48 is sleeved on the outer side of the middle part of the pin rod. The upper end surface of the dust cover 48 fits against the lower end surface of the sleeve 45. The washer 49 is sleeved on the lower end of the top edge of the dust cover 48.

[0048] Specifically, in the present embodiment, dust protection for the bottom of the ball pin assembly 4 is achieved by providing a dust cover 48 on the outside of the pin rod portion, and a gasket 49 is provided at the lower end of the top edge of the dust cover 48, so that the dust cover 48 can be protected during the riveting and extrusion process of the ball seat portion 3, thereby avoiding damage to the dust cover 48 due to extrusion, thereby improving the durability of the dust cover 48.

[0049] Example 2, reference Figure 3 , which is the second embodiment of the present invention. Different from the previous embodiment, this embodiment provides an assembly detection method, which is applied to the automobile tie rod structure as described above, and can improve the safety of the ball rivet assembly during the spin riveting assembly process, and includes:

[0050] Step S1, before the riveting machine performs riveting processing on the ball pin assembly 4, a standard assembly image of the ball pin assembly 4 is photographed and hardness parameters and elastic parameters of various parts of the ball pin assembly 4 are recorded, and the standard coordinates of the ball pin of the ball pin assembly 4 in the three-dimensional coordinate system are obtained according to the standard assembly image processing;

[0051] Step S2, inputting the hardness parameters and elasticity parameters of each part into a preset comprehensive hardness calculation formula and a comprehensive elasticity calculation formula, and calculating the comprehensive hardness parameters and the comprehensive elasticity parameters respectively;

[0052] Step S3, when the riveting machine performs riveting processing on the ball pin assembly 4, a dynamic assembly video of the ball pin assembly 4 is recorded, and at the same time, a rotation speed of the rotation output end of the riveting machine is detected, and the riveting pressure value at each riveting pressing point between the riveting machine and the ball seat part 3 is detected, and a plurality of riveting processing images are obtained according to the splitting of the dynamic assembly video, and then the ball pin processing coordinates of the ball pin assembly 4 in the three-dimensional coordinate system are obtained according to the riveting processing image processing;

[0053] Step S4, obtaining the real-time offset angle of the ball pin assembly 4 according to the ball pin standard coordinates and the ball pin processing coordinates in the same three-dimensional coordinate system;

[0054] Step S5, pre-training a ball pin comprehensive deformation prediction model, inputting the real-time offset angle, each rivet pressure value, rotation speed, comprehensive hardness parameter, and comprehensive elasticity parameter into the ball pin comprehensive deformation prediction model, and predicting the ball pin comprehensive deformation score at the current moment;

[0055] Step S6, when the ball pin comprehensive deformation score is greater than the preset deformation threshold, an assembly alarm instruction is generated, and according to the assembly alarm instruction, an external ball pin assembly 4 is warned of assembly error and the riveting output state of the riveting machine is adjusted.

[0056] Working principle of embodiment 2:

[0057] In this embodiment, an intelligent processing terminal is configured, and an image processing program, a parameter entry program, a comprehensive calculation program, a deformation prediction program, and an alarm adjustment program are configured on the intelligent processing terminal. In step S1, a CCD industrial camera is used to shoot the standard assembly image of the ball pin assembly 4. The hardness parameters and elasticity parameters of each part of the ball pin assembly 4 can be entered manually or through a parameter entry program. The image processing program has a built-in convolutional neural network, which can automatically identify and learn features, build a three-dimensional model in a three-dimensional coordinate system according to the drawing information, and generate three-dimensional coordinates. The riveting processing image is input into the image processing program to generate the ball pin processing coordinates of the ball pin assembly 4 in the three-dimensional coordinate system. In step S2, the comprehensive hardness calculation formula and the comprehensive elasticity calculation formula are pre-stored in the comprehensive calculation program, and the hardness parameters and elasticity parameters of each part are input into the comprehensive calculation program, and the built-in comprehensive hardness calculation formula and the comprehensive elasticity calculation formula are used to calculate the comprehensive hardness parameter and the comprehensive elasticity parameter respectively. Step S3, during the riveting process of the ball pin assembly 4 by the riveting machine, a CCD industrial camera is used to record the dynamic assembly video of the ball pin assembly 4. The CCD industrial camera records 30 frames of video image frames per second, so the dynamic assembly video is divided into 30 riveting processing images per second, and then the first video image frame in the dynamic assembly video is input into the image processing program. The image processing program repeats the image processing operation in step S1, and the video image frame is processed to obtain the ball pin processing coordinates of the ball pin assembly 4 in the three-dimensional coordinate system. At the same time, an encoder is configured at the riveting output end of the riveting machine to detect the rotation speed. Several riveting pressing points are configured between the riveting machine and the ball seat part 3, and pressure sensors are provided at each riveting pressing point to detect the riveting pressure value at each riveting pressing point. In step S4, the standard coordinates of the ball pin and the processing coordinates of the ball pin are calculated using the angle calculation formula in solid geometry to obtain the real-time offset angle of the ball pin assembly 4 in the same three-dimensional coordinate system. In step S5, the deformation prediction program is pre-configured with a trained pin comprehensive deformation prediction model, and the real-time offset angle, each rivet pressure value, rotation speed, comprehensive hardness parameter, and comprehensive elastic parameter obtained and calculated in the above steps are input into the ball pin comprehensive deformation prediction model to predict the ball pin comprehensive deformation score at the current moment. In step S6, the deformation threshold can be 30%. When the ball pin comprehensive deformation score exceeds 30%, it indicates that the current deformation of the ball pin assembly 4 is too large, and an alarm is required to alert external personnel. At the same time, the rotation speed of the rivet machine and the rivet pressure on the ball seat 3 are immediately reduced to avoid further damage to the ball pin assembly 4.

[0058] In this embodiment, during the assembly process of the ball pin assembly 4, the ball pin comprehensive deformation score of the ball pin assembly 4 is dynamically predicted based on the calculated and detected real-time offset angle, each rivet pressure value, rotation speed, comprehensive hardness parameter, and comprehensive elastic parameter to reflect the deformation degree of the ball pin assembly 4. At the same time, when the ball pin comprehensive deformation score is greater than the deformation threshold, the output state of the rivet machine is adjusted in time, which can avoid large deformation of the ball pin assembly 4 during the rivet processing, thereby significantly improving the safety of the ball rivet assembly during the rivet assembly process.

[0059] Preferably, the comprehensive hardness calculation formula is configured as:

[0060]

[0061] Among them, C hc Used to indicate comprehensive hardness parameter, B s Used to indicate the hardness of the ball stud, B b1 Used to indicate the hardness of the first bowl, B b2 Used to indicate the hardness of the second bowl, S l Used to indicate sleeve hardness, R r Used to indicate the hardness of rubber ring, E c Used to indicate the hardness of the end cap, ε 1 , ε 2 , ε 3 , ε 4 , ε 5 , ε 6 are used to represent a preset first hardness conversion coefficient, a second hardness conversion coefficient, a third hardness conversion coefficient, a fourth hardness conversion coefficient, a fifth hardness conversion coefficient and a sixth hardness conversion coefficient respectively;

[0062] The comprehensive elasticity calculation formula is configured as:

[0063]

[0064] Among them, C ec Used to represent the comprehensive elastic parameter, B he Used to indicate the elasticity of the ball stud, B be1 Used to indicate the elasticity of the first bowl, B be2 Used to indicate the elasticity of the second bowl, S e Used to indicate sleeve elasticity, R re Used to indicate the elasticity of the rubber ring, E ce Used to represent the elasticity of the end cap, σ 1 , σ 2 , σ 3 , σ 4 , σ 5 , σ 6They are respectively used to represent the preset first elastic conversion coefficient, second elastic conversion coefficient, third elastic conversion coefficient, fourth elastic conversion coefficient, fifth elastic conversion coefficient and sixth elastic conversion coefficient.

[0065] Specifically, in this embodiment, through the comprehensive hardness calculation formula, the ball pin hardness is multiplied by the first hardness conversion coefficient to obtain the influence parameter of the ball pin hardness on the comprehensive hardness; the first ball bowl hardness is multiplied by the second hardness conversion coefficient to obtain the influence parameter of the first ball bowl hardness on the comprehensive hardness; the second ball bowl hardness is multiplied by the third hardness conversion coefficient to obtain the influence parameter of the second ball bowl hardness on the comprehensive hardness; the sleeve hardness is multiplied by the fourth hardness conversion coefficient to obtain the influence parameter of the sleeve hardness on the comprehensive hardness; the rubber ring hardness is multiplied by the fifth hardness conversion coefficient to obtain the influence parameter of the rubber ring hardness on the comprehensive hardness; the end cover hardness is multiplied by the sixth hardness conversion coefficient to obtain the influence parameter of the end cover hardness on the comprehensive hardness, and finally summarized into the comprehensive hardness calculation formula to calculate the comprehensive hardness parameter;

[0066] Through the comprehensive elasticity calculation formula, the ball pin elasticity is multiplied by the first elastic conversion coefficient to obtain the influence parameter of the ball pin elasticity on the comprehensive elasticity; the first ball bowl elasticity is multiplied by the second elastic conversion coefficient to obtain the influence parameter of the first ball bowl elasticity on the comprehensive elasticity; the second ball bowl elasticity is multiplied by the third elastic conversion coefficient to obtain the influence parameter of the second ball bowl elasticity on the comprehensive elasticity; the sleeve elasticity is multiplied by the fourth elastic conversion coefficient to obtain the influence parameter of the sleeve elasticity on the comprehensive elasticity; the rubber ring elasticity is multiplied by the fifth elastic conversion coefficient to obtain the influence parameter of the rubber ring elasticity on the comprehensive elasticity; the end cover elasticity is multiplied by the sixth elastic conversion coefficient to obtain the influence parameter of the end cover elasticity on the comprehensive elasticity, and finally summarized into the comprehensive elasticity calculation formula to calculate the comprehensive elasticity parameters.

[0067] Preferably, Figure 4 As shown, step S5 includes:

[0068] Step S51, a preset ball pin riveting processing database stores training parameters of several historical moments, the training parameters including historical offset angles, historical riveting pressure values, historical rotation speeds, historical comprehensive hardness parameters, historical comprehensive elastic parameters, and historical ball pin comprehensive deformation scores, and the training parameters are divided into a training set, a test set, and a validation set according to the proportion;

[0069] Step S52, introducing an initial model, taking the historical offset angle, historical riveting pressure value, historical rotation speed, historical comprehensive hardness parameter and historical comprehensive elasticity parameter in the training set as input, taking the historical ball pin comprehensive deformation score in the training set as output, and training to obtain an initial prediction model;

[0070] Step S53, using the training parameters in the validation set to adjust the parameters of the initial prediction model, using the training parameters in the test set to test the accuracy of the initial prediction model, and finally outputting it as a ball pin comprehensive deformation prediction model after the accuracy of the initial prediction model is greater than a preset accuracy threshold.

[0071] Specifically, in this embodiment, the historical offset angle, historical riveting pressure value, historical rotation speed, historical comprehensive hardness parameter, and historical comprehensive elasticity parameter in the training parameters are obtained by detection and calculation at historical moments. The historical ball pin comprehensive deformation score is a parameter score artificially calculated and set by the staff based on the historical offset angle, historical riveting pressure value, historical rotation speed, historical comprehensive hardness parameter, and historical comprehensive elasticity parameter.

[0072] Preferably, the initial model is an XLSTM model.

[0073] Specifically, in this embodiment, the XLSTM model is an extension of the traditional LSTM model. It improves the LSTM model by introducing a new gating mechanism and memory structure, aiming to improve the performance and scalability of the LSTM model when processing large-scale data.

[0074] Preferably, the data ratio in the training set, validation set and test set is 3:1:1.

[0075] Embodiment 3 is the third embodiment of the present invention. Different from the previous embodiment, this embodiment introduces environmental parameters, and can optimize and adjust the ball pin comprehensive deformation prediction model by using environmental parameters to improve the prediction accuracy of the ball pin comprehensive deformation prediction model. In step S5, it also includes real-time detection of environmental parameters, and the environmental parameters include ambient temperature, ambient humidity, ultraviolet intensity and antioxidant concentration;

[0076] The weight parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ambient temperature and ambient humidity to obtain the optimized weight parameters, and the bias parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ultraviolet intensity and the antioxidant concentration to obtain the optimized bias parameters, so as to obtain the optimized ball pin comprehensive deformation prediction model, and the real-time offset angle, each rivet pressure value, rotation speed, comprehensive hardness parameter, and comprehensive elasticity parameter are input into the optimized ball pin comprehensive deformation prediction model to predict the ball pin comprehensive deformation score.

[0077] Working principle of embodiment 3:

[0078] The ambient temperature, ambient humidity, ultraviolet intensity and antioxidant concentration in the environment will affect the deformation of the ball pin assembly, which will reduce the prediction accuracy of the ball pin comprehensive deformation score of the ball pin comprehensive deformation prediction model. The weight parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ambient temperature and ambient humidity to obtain the optimized weight parameters, and the bias parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ultraviolet intensity and antioxidant concentration to obtain the optimized bias parameters. The weight parameters and bias parameters of the ball pin comprehensive deformation prediction model are dynamically adjusted according to the ambient temperature, ambient humidity, ultraviolet intensity and antioxidant concentration, so that the ball pin comprehensive deformation score predicted by the optimized ball pin comprehensive deformation prediction model incorporates the influence of environmental parameters. Therefore, the ball pin comprehensive deformation score is closer to the actual deformation degree of the ball pin assembly, and the prediction accuracy of the ball pin comprehensive deformation prediction model is significantly improved.

[0079] Preferably, in step S5, the ambient temperature and ambient humidity are input into a preset weight parameter optimization formula to calculate the optimized weight parameter, and the weight parameter optimization formula is configured as:

[0080]

[0081] Among them, W' is used to represent the optimization weight parameter, W 0 Used to represent the weight parameters before optimization, a t Used to indicate ambient temperature, a h Used to indicate ambient humidity, χ 1 , χ 2 They are respectively used to represent the preset first temperature conversion coefficient and the first humidity conversion coefficient.

[0082] Preferably, in step S5, the ultraviolet intensity and the antioxidant concentration are input into a preset bias parameter optimization formula to calculate the optimized bias parameter, and the bias parameter optimization formula is configured as:

[0083]

[0084] Among them, B' is used to represent the optimized bias parameter, B 0 Used to represent the bias parameters before optimization, U i Used to indicate the intensity of ultraviolet light, a c Used to express antioxidant concentration, δ 1 , δ 2 They are respectively used to represent the preset first ultraviolet intensity conversion coefficient and the first antioxidant concentration conversion coefficient.

[0085] The above are only preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. An assembly detection method for an automobile tie rod structure, applied to an automobile tie rod structure, characterized in that: The automobile tie rod structure comprises: a tie rod assembly (1), wherein the tie rod assembly (1) comprises an integrally formed tie rod portion (2) and two ball seat portions (3), wherein the two ball seat portions (3) are located at two ends of the tie rod portion (2), wherein an assembled ball pin assembly (4) is arranged in the ball seat portion (3), and wherein the ball pin assembly (4) comprises a ball stud (41), an oil nozzle (42), a first ball bowl (43), a second ball bowl (44), a sleeve (45), a rubber ring (46), and an end cover (47); The ball pin (41) comprises a ball head and a pin rod, the end cover (47) covers the upper end of the ball head, an oil inlet hole is also opened at the top center of the ball head, the oil nozzle (42) penetrates the end cover (47) from the upper end of the end cover (47) and is connected to the oil inlet hole, the first ball bowl (43) is sleeved on the outer side of the top of the ball head, the second ball bowl (44) is sleeved on the outer side of the middle of the ball head, the sleeve (45) is sleeved on the outer side of the top of the pin rod, and the rubber ring (46) is sleeved on the outer side of the sleeve (45); The ball pin assembly (4) further comprises a dust cover (48) and a washer (49), wherein the dust cover (48) is sleeved on the outer side of the middle portion of the pin rod, the upper end surface of the dust cover (48) is fitted against the lower end surface of the sleeve (45), and the washer (49) is sleeved on the lower end of the top edge of the dust cover (48); The assembly detection method of the automobile tie rod structure comprises: Step S1, before the riveting machine performs riveting processing on the ball pin assembly (4), a standard assembly image of the ball pin assembly (4) is photographed and hardness parameters and elasticity parameters of various parts of the ball pin assembly (4) are simultaneously recorded, and the standard coordinates of the ball pin of the ball pin assembly (4) in a three-dimensional coordinate system are obtained according to the standard assembly image processing; Step S2, inputting the hardness parameter and the elasticity parameter of each part into a preset comprehensive hardness calculation formula and a preset comprehensive elasticity calculation formula, and calculating the comprehensive hardness parameter and the comprehensive elasticity parameter respectively; Step S3, during the process of the riveting machine performing riveting processing on the ball pin assembly (4), a dynamic assembly video of the ball pin assembly (4) is recorded, and at the same time, a rotation speed of the rotation output end of the riveting machine is detected, as well as the riveting pressure values ​​at each riveting pressure point between the riveting machine and the ball seat portion (3) are detected, and a plurality of riveting processing images are obtained according to the splitting of the dynamic assembly video, and then the ball pin processing coordinates of the ball pin assembly (4) in a three-dimensional coordinate system are obtained according to the riveting processing images; Step S4, obtaining the real-time offset angle of the ball pin assembly (4) in the same three-dimensional coordinate system according to the standard ball pin coordinates and the processing ball pin coordinates; Step S5, pre-training a ball pin comprehensive deformation prediction model, inputting the real-time offset angle, each of the rivet pressure values, the rotation speed, the comprehensive hardness parameter, and the comprehensive elasticity parameter into the ball pin comprehensive deformation prediction model, and predicting the ball pin comprehensive deformation score at the current moment; Step S6, when the ball pin comprehensive deformation score is greater than a preset deformation threshold, an assembly alarm instruction is generated, and according to the assembly alarm instruction, an external warning is given that the ball pin assembly (4) is assembled incorrectly and the riveting output state of the riveting machine is adjusted.

2. The assembly detection method of the automobile tie rod structure according to claim 1, characterized in that: The comprehensive hardness calculation formula is configured as: Among them, C hc Used to express the comprehensive hardness parameter, B s Used to indicate the hardness of the ball stud, B b1 Used to indicate the hardness of the first bowl, B b2 Used to indicate the hardness of the second bowl, S l Used to indicate sleeve hardness, R r Used to indicate the hardness of rubber ring, E c Used to indicate the hardness of the end cap, ε1, ε2, ε3, ε4, ε5, ε6 are used to indicate the preset first hardness conversion coefficient, second hardness conversion coefficient, third hardness conversion coefficient, fourth hardness conversion coefficient, fifth hardness conversion coefficient and sixth hardness conversion coefficient respectively; The comprehensive elasticity calculation formula is configured as: Among them, C ec Used to express the comprehensive elastic parameter, B he Used to indicate the elasticity of the ball stud, B be1 Used to indicate the elasticity of the first bowl, B be2 Used to indicate the elasticity of the second bowl, S e Used to indicate sleeve elasticity, R re Used to indicate the elasticity of the rubber ring, E ce Used to represent the elasticity of the end cover, σ1, σ2, σ3, σ4, σ5, σ6 are used to represent the preset first elastic conversion coefficient, second elastic conversion coefficient, third elastic conversion coefficient, fourth elastic conversion coefficient, fifth elastic conversion coefficient and sixth elastic conversion coefficient, respectively.

3. The assembly detection method of the automobile tie rod structure according to claim 1, characterized in that: The step S5 comprises: Step S51, storing training parameters of several historical moments in a preset ball pin riveting processing database, wherein the training parameters include historical offset angles, historical riveting pressure values, historical rotation speeds, historical comprehensive hardness parameters, historical comprehensive elastic parameters, and historical ball pin comprehensive deformation scores, and dividing the training parameters into a training set, a test set, and a verification set according to a proportion; Step S52, introducing an initial model, taking the historical offset angle, the historical riveting pressure value, the historical rotation speed, the historical comprehensive hardness parameter and the historical comprehensive elasticity parameter in the training set as input, taking the historical ball pin comprehensive deformation score in the training set as output, and training to obtain an initial prediction model; Step S53, using the training parameters in the validation set to adjust the parameters of the initial prediction model, using the training parameters in the test set to test the accuracy of the initial prediction model, and finally outputting it as the ball pin comprehensive deformation prediction model when the accuracy of the initial prediction model is greater than a preset accuracy threshold.

4. The assembly detection method of the automobile tie rod structure according to claim 3, characterized in that: The initial model is an XLSTM model.

5. The assembly detection method of the automobile tie rod structure according to claim 3, characterized in that: The data ratio in the training set, validation set, and test set is 3:1:

1.

6. The assembly detection method of the automobile tie rod structure according to claim 3, characterized in that: In the step S5, it also includes real-time detection of environmental parameters, wherein the environmental parameters include ambient temperature, ambient humidity, ultraviolet intensity and antioxidant concentration; The weight parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ambient temperature and the ambient humidity to obtain optimized weight parameters, and the bias parameters between the input layer and the hidden layer of the ball pin comprehensive deformation prediction model are adjusted according to the ultraviolet intensity and the antioxidant concentration to obtain optimized bias parameters, so as to obtain the optimized ball pin comprehensive deformation prediction model, and the real-time offset angle, each of the rivet pressure values, the rotation speed, the comprehensive hardness parameter, and the comprehensive elasticity parameter are input into the optimized ball pin comprehensive deformation prediction model to predict the ball pin comprehensive deformation score.

7. The assembly detection method of the automobile tie rod structure according to claim 6, characterized in that: In step S5, the ambient temperature and the ambient humidity are input into a preset weight parameter optimization formula to calculate the optimized weight parameter, and the weight parameter optimization formula is configured as: Wherein, W' is used to represent the optimized weight parameter, W0 is used to represent the weight parameter before optimization, and a t Used to indicate the ambient temperature, a h It is used to represent the ambient humidity, and χ1 and χ2 are used to represent the preset first temperature conversion coefficient and the first humidity conversion coefficient respectively.

8. The assembly detection method of the automobile tie rod structure according to claim 6, characterized in that: In the step S5, the ultraviolet intensity and the antioxidant concentration are input into a preset bias parameter optimization formula to calculate the optimized bias parameter, and the bias parameter optimization formula is configured as: Wherein, B' is used to represent the optimized bias parameter, B0 is used to represent the bias parameter before optimization, and U i Used to indicate the intensity of ultraviolet light, a c It is used to represent the antioxidant concentration, and δ1 and δ2 are used to represent a preset first ultraviolet intensity conversion coefficient and a first antioxidant concentration conversion coefficient, respectively.

Images