Component size grading method and production device

Abstract

Embodiments of the present application provide a component size grading method and a production device. The method comprises: acquiring a plurality of pieces of size data of a plurality of components (S101); on the basis of the plurality of pieces of size data, determining a plurality of grading schemes of the plurality of components (S102); on the basis of the plurality of grading schemes of the plurality of components, determining matching schemes of the plurality of components (S103); calculating the fitness score of each matching scheme (S104); and on the basis of the fitness score of each matching scheme, determining a matching scheme of a preset assembly (S105). The embodiments of the present application implement automatic optimization of the component size matching scheme of a preset assembly, thereby effectively improving the assembly yield of the preset assembly.

Description

Component size classification method and production equipment

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to the Chinese patent application filed with the China Patent Office on February 2, 2024, with application number 202410150681.3 and application name “Size Grading Method and Related Equipment of Rotating Shaft Assembly”, the entire contents of which are incorporated by reference into this application. Technical Field

[0003] The present application relates to the technical field of production of smart terminals, and in particular to a component size grading method and production equipment. Background Art

[0004] When assembling multiple components into a mechanical assembly, the combination of components of different sizes can affect the assembly yield. Therefore, it is necessary to perform tolerance binning on the component sizes of the mechanical assembly. However, the methods used in related art to determine the component size binning scheme for the mechanical assembly through tolerance binning are complex and time-consuming.

[0005] Summary of the Invention

[0006] In view of the above, it is necessary to provide a component size grading method and production equipment to solve the problem that the method of determining the component size grading scheme of mechanical components through tolerance grading in the above-mentioned related technologies is relatively complicated and time-consuming.

[0007] In a first aspect, an embodiment of the present application provides a component size grading method, which is applied to production equipment, and the method includes: obtaining multiple dimension data of multiple components; determining multiple grading schemes of the multiple components based on the multiple dimension data; determining a matching scheme of the multiple components based on the multiple grading schemes of the multiple components; calculating the fitness score of each matching scheme; and determining a matching scheme of a preset component based on the fitness score of each matching scheme.

[0008] Through the above technical solution, the grading and matching schemes of multiple size data of multiple components are automatically optimized, which improves the efficiency and accuracy of grading and matching the size data of preset components and effectively improves the assembly yield of preset components.

[0009] In one possible implementation, determining multiple grading schemes for the multiple components based on the multiple dimensional data includes: setting grading levels of the multiple dimensional data of the multiple components and a grading threshold for each level; and determining the grading levels of the multiple dimensional data of the multiple components and a grading threshold for each level as multiple grading schemes for the multiple components.

[0010] Through the above technical solution, a genetic algorithm is used to automatically optimize the component size data grading scheme, thereby improving the efficiency and intelligence of component size grading and matching.

[0011] In one possible implementation, determining a matching scheme for the multiple components based on the multiple grading schemes of the multiple components includes: determining the quantile corresponding to each grading threshold of the multiple size data of the multiple components based on an inverse cumulative distribution function; grouping the multiple size data of the multiple components according to the quantile corresponding to each grading threshold; and matching the grouped size data of different components to obtain a matching scheme for the multiple size data of the multiple components.

[0012] Through the above technical solution, a genetic algorithm is used to automatically optimize the component size data matching solution, thereby improving the efficiency and intelligence of component size grading and matching.

[0013] In one possible implementation, determining the quantile corresponding to each bin threshold of the multiple dimensional data of the multiple components based on the inverse cumulative distribution function includes: using the inverse cumulative distribution function of the normal distribution to calculate the dimensional data corresponding to each bin threshold, and using the calculated dimensional data corresponding to each bin threshold as the quantile.

[0014] Through the above technical solution, the quantile corresponding to the bin threshold can be quickly calculated, thereby improving the matching efficiency of the size data of multiple components.

[0015] In a possible implementation, the calculating the fitness score of each collocation scheme includes: using a fitness function to calculate the fitness score of each collocation scheme according to the grouping size data of the different components.

[0016] Through the above technical solution, the fitness score of each grading scheme can be accurately calculated.

[0017] In a possible implementation, determining the collocation scheme of the preset component based on the fitness score of each collocation scheme includes: if the fitness score of any collocation scheme is greater than or equal to a preset fitness score threshold, determining the any collocation scheme as the collocation scheme of the preset component; or determining the collocation scheme with the highest fitness score among multiple collocation schemes as the collocation scheme of the preset component.

[0018] Through the above technical solution, the matching scheme of the preset components can be accurately determined.

[0019] In one possible implementation, the fitness function is used to calculate the fitness score of each matching scheme based on the grouping size data of the different components, including: inputting the grouping size data of the different components corresponding to each matching scheme into a prediction model to obtain multiple prediction results for each matching scheme; and using the fitness function to obtain the fitness score of each matching scheme based on the multiple prediction results of each matching scheme.

[0020] Through the above technical solution, the fitness score of each matching solution can be accurately calculated.

[0021] In a possible implementation, the preset component is a shaft component, and the shaft component has a flattening angle in a flattened state. The fitness function is a calculation formula for the yield of all flattening angle prediction values ​​under the matching scheme.

[0022] Through the above technical solution, the calculation formula of the yield of the flattening angle prediction value is used as the fitness function, which can improve the calculation efficiency of the fitness score.

[0023] In one possible implementation, the fitness function is used to calculate the fitness score of each matching scheme based on the group size data of the different components, including: inputting the group size data of the different components corresponding to each matching scheme into a flattening angle prediction model to obtain multiple predicted flattening angles for each matching scheme; calculating the yield of the multiple predicted flattening angles of each matching scheme to obtain the fitness score of each matching scheme.

[0024] Through the above technical solution, the yield of the predicted flattening angle is used as the fitness score of the matching scheme, which facilitates the evaluation of different matching schemes.

[0025] In one possible implementation, the method further includes: extracting multiple features that affect the flattening angle of the shaft assembly, wherein the multiple features include dimensional data of components of the shaft assembly; and establishing the flattening angle prediction model based on the multiple features and the flattening angle data.

[0026] Through the above technical solution, a flattening angle prediction model is established based on multiple features that affect the flattening angle of the shaft assembly, thereby improving the efficiency of determining corresponding flattening angle data based on the dimensional data of multiple components.

[0027] In one possible implementation, the extracting of multiple features that affect the flattening angle of the shaft assembly includes: obtaining force information of the element of the shaft assembly in a flattened state; obtaining sampling data corresponding to the dimensional data of the element based on the force information of the element of the shaft assembly in a flattened state; and performing regression analysis on the sampling data corresponding to the dimensional data of the element to obtain the multiple features that affect the flattening angle of the shaft assembly.

[0028] Through the above technical solution, multiple features that affect the flattening angle in the shaft assembly can be quickly and accurately determined, effectively improving the efficiency of component size classification and matching.

[0029] In one possible implementation, the sampling data corresponding to the dimensional data of the element are subjected to regression analysis to obtain the multiple features that affect the flattening angle of the rotating shaft assembly, including: standardizing the sampling data; dividing the sampling data into a training set, and using the training set to train a fitted cable regression model; obtaining coefficients of all sampling data based on the trained cable regression model; using the absolute value of the coefficient as the contribution of the dimensional data of the element to the flattening angle, sorting the dimensional data of the element in descending order of the contribution, and selecting multiple dimensional data ranked first according to a preset percentage to obtain the multiple features.

[0030] Through the above technical solution, the Lasso regression model is used to screen out the features that affect the flattening angle from a large amount of dimensional data. Subsequently, a flattening angle prediction model is established based on the features, and the grading and matching schemes are automatically optimized, which effectively improves the efficiency of establishing the flattening angle prediction model and the efficiency of optimizing the grading and matching schemes.

[0031] In a possible implementation, establishing the flattening angle prediction model based on the multiple features and the flattening angle data includes: training a preset deep learning model based on the multiple features and the flattening angle data to obtain the flattening angle prediction model.

[0032] Through the above technical solution, an accurate flattening angle prediction model can be obtained by training the preset deep learning model through the component size data and the flattening angle data of the shaft assembly, thereby improving the accuracy of the flattening angle prediction results.

[0033] In one possible implementation, the preset deep learning model is trained according to the multiple features and the flattening angle data to obtain the flattening angle prediction model, including: establishing a training sample set according to multiple groups of features and the flattening angle data corresponding to each group of features; using a group of features and the corresponding flattening angle data in the training sample set as training data, inputting the preset deep learning model, and calculating the loss function value of the preset deep learning model; if the loss function value of the preset deep learning model is greater than a preset value, adjusting the parameters of the preset deep learning model, and continuing to train the preset deep learning model until the loss function value of the preset deep learning model is less than or equal to the preset value, determining that the training of the preset deep learning model is completed, and determining the trained preset deep learning model as the flattening angle prediction model.

[0034] Through the above technical solution, multiple features that affect the flattening angle of the shaft assembly are used to train a preset deep learning model to establish a flattening angle prediction model, which can improve the accuracy of the flattening angle prediction value output by the flattening angle prediction model.

[0035] In a second aspect, an embodiment of the present application provides a component size grading method, which is applied to production equipment, and the method includes: obtaining multiple size data of multiple first components and multiple second components; determining multiple grading schemes of the multiple first components and the multiple second components based on the multiple size data; determining matching schemes of the multiple first components and the multiple second components based on the multiple grading schemes of the multiple first components and the multiple second components; calculating the fitness score of each matching scheme; and determining matching schemes of multiple preset components based on the fitness score of each matching scheme, wherein each preset component includes a first component and a second component.

[0036] In one possible implementation, determining multiple grading schemes for the multiple first elements and the multiple second elements based on the multiple dimensional data includes: setting grading levels of the multiple dimensional data of the multiple first elements and the multiple second elements and the grading threshold of each level; and determining the grading levels of the multiple dimensional data of the multiple first elements and the multiple second elements and the grading threshold of each level as multiple grading schemes for the multiple first elements and the multiple second elements.

[0037] In one possible implementation, determining a matching scheme of the multiple first elements and the multiple second elements based on the multiple grading schemes of the multiple first elements and the multiple second elements includes: determining the quantile corresponding to each grading threshold of the multiple size data of the multiple first elements and the multiple second elements according to the inverse cumulative distribution function; grouping the multiple size data of the multiple first elements and the multiple second elements according to the quantile corresponding to each grading threshold; and matching the grouped size data of the multiple first elements and the multiple second elements to obtain a matching scheme of the multiple size data of the multiple first elements and the multiple second elements.

[0038] In a possible implementation, the calculating the fitness score of each collocation scheme includes: using a fitness function to calculate the fitness score of each collocation scheme according to the grouping size data of the plurality of first elements and the plurality of second elements.

[0039] In a possible implementation, the preset component is a shaft component, and the shaft component has a flattening angle in a flattened state. The fitness function is a calculation formula for the yield of all flattening angle prediction values ​​under the matching scheme.

[0040] In one possible implementation, the fitness function is used to calculate the fitness score of each matching scheme based on the group size data of the multiple first elements and the multiple second elements, including: inputting the group size data of the multiple first elements and the multiple second elements corresponding to each matching scheme into a flattening angle prediction model to obtain multiple predicted flattening angles for each matching scheme; calculating the yield of the multiple predicted flattening angles of each matching scheme to obtain the fitness score of each matching scheme.

[0041] In a third aspect, an embodiment of the present application provides a production device, which includes: a feeding part for receiving a plurality of components; a processing part for executing the above-mentioned component size classification method and determining a matching scheme of preset components; a sorting part for matching the plurality of components according to the matching scheme of the preset components; and a discharging part for outputting the plurality of components that have been matched.

[0042] In a possible implementation, the production equipment further includes: a measuring component configured to measure a plurality of dimensional data of the plurality of components, and send the measured plurality of dimensional data of the plurality of components to the processing component.

[0043] In a possible implementation, the production equipment further includes: an assembly component for assembling the plurality of matched components into the preset assembly.

[0044] In a fourth aspect, an embodiment of the present application provides a production device, comprising: a feeding part for receiving a plurality of first components and a plurality of second components; a processing part for executing the above-mentioned component size classification method and determining a matching scheme of a plurality of preset components; a sorting part for matching the plurality of first components and the plurality of second components according to the matching scheme of the plurality of preset components; and a discharging part for outputting the plurality of first components and the plurality of second components that have completed matching.

[0045] In a fifth aspect, the present application provides a computer storage medium storing program instructions. When the program instructions are executed on a production device, the processor of the production device executes the above-mentioned component size binning method.

[0046] In addition, the technical effects brought about by the second to fifth aspects can be found in the descriptions of the methods of each design in the above method section, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] FIG1 is a three-dimensional diagram of a terminal device provided in an embodiment of the present application in a folded state.

[0048] FIG2 is a three-dimensional diagram of a terminal device provided in an embodiment of the present application in a flattened state.

[0049] FIG3 is a perspective view of another foldable terminal device provided by an embodiment of the present application.

[0050] FIG4 is a flow chart of a component size binning method provided in an embodiment of the present application.

[0051] FIG5 is a flow chart of a component size binning method provided in another embodiment of the present application.

[0052] FIG6 is a flow chart of a component size classification method provided in another embodiment of the present application.

[0053] FIG7 is a flowchart of feature extraction provided by an embodiment of the present application.

[0054] FIG8 is a flowchart of model training provided by an embodiment of the present application.

[0055] FIG9 is a perspective view of a main swing arm provided in an embodiment of the present application.

[0056] 10 to 12 are three-dimensional views of a base provided in one embodiment of the present application.

[0057] FIG13 is a perspective view of a secondary swing arm provided in an embodiment of the present application.

[0058] 14 to 16 are three-dimensional views of a wedge block provided in accordance with an embodiment of the present application.

[0059] Figure 17 is a structural diagram of a preset deep learning model provided in an embodiment of the present application.

[0060] FIG18 is a flow chart of optimizing the bin positions and bin thresholds provided in one embodiment of the present application.

[0061] FIG19 is a schematic diagram of the bin positions and bin thresholds provided in an embodiment of the present application.

[0062] FIG20 is a schematic diagram of the architecture of the production equipment provided in one embodiment of the present application.

[0063] FIG21 is a schematic diagram of an application scenario of a component size classification method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0064] The terms "first" and "second" involved in the embodiments of the present application are for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more features. In the description of the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used in the specification of this application are for the purpose of describing specific embodiments only and are not intended to limit this application. It should be understood that, unless otherwise specified in this application, " / " means or. For example, A / B can mean A or B. "And / or" in this application is merely a way to describe the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. "At least one" means one or more. "Multiple" means two or more than two. For example, at least one of a, b or c can mean: a, b, c, a and b, a and c, b and c, a, b and c. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0066] For ease of understanding, some concepts related to the embodiments of the present application are exemplarily described for reference:

[0067] Finite element simulation: It is a numerical simulation method that regards a continuum as a discrete set of finite units connected to each other in a certain way. It is used to solve thermal, mechanical, electromagnetic and other problems of the continuum. Finite element simulation is based on the finite element method, which decomposes complex structures or systems into a finite number of simple elements. Numerical calculations are performed on these elements to derive the behavior of the overall system.

[0068] Monte Carlo sampling, also known as statistical simulation sampling, is a method of performing approximate numerical calculations by randomly sampling from a probability model. Specifically, simulation refers to replacing or simulating certain characteristics or partial states of a real or abstract system with another system (called a simulation model). To solve a problem, it is transformed into a problem that requires solving a probability model. A large number of random numbers that conform to the model are then generated and analyzed to solve the problem. This method is called random simulation sampling, also known as Monte Carlo sampling.

[0069] Tolerance: The permissible variation in the actual parameter value. Parameters can be geometric parameters in machining, or parameters in physics, chemistry, electricity, and other disciplines. In mechanical manufacturing, the purpose of establishing tolerances is to determine the geometric parameters of a product so that their variation is within a certain range to ensure interchangeability or compatibility.

[0070] Position tolerance: used to specify the position tolerance or allowable range of position deviation of a part.

[0071] Flattening angle: A parameter that describes the relationship between displacement and force under load, particularly in mechanical engineering and structural analysis. Flattening angle describes the distortion or deviation of an object under load and can be used to understand the deformation state of a structure.

[0072] When assembling multiple components into a mechanical assembly, the combination of components of different sizes can affect the assembly yield. Therefore, it is necessary to perform tolerance binning on the component sizes of the mechanical assembly. However, the methods used in related art to determine the component size binning scheme for the mechanical assembly through tolerance binning are complex and time-consuming.

[0073] With the development of smart terminal technology, users have an increasingly strong demand for large-screen terminals. However, large-screen terminals are not easy to carry due to their large screen size. Foldable terminal devices use a folding screen method to increase the screen size without significantly increasing the overall size of the device, becoming the current mainstream large-screen terminal solution.

[0074] Foldable terminal devices typically use a hinge assembly to achieve a folding function. However, during the assembly process of the foldable terminal device, the hinge assembly is prone to defects such as under-expansion or over-expansion. As a result, the production yield of the hinge assembly is low, which in turn leads to a low production yield of the foldable terminal device.

[0075] Refer to Figure 1, which is a perspective view of a terminal device provided in an embodiment of the present application in a folded state. The terminal device 200 provided in an embodiment of the present application is a foldable terminal device, comprising a first display screen 10, a second display screen 20, and a hinge assembly 30. The first display screen 10 is the outer screen of the terminal device 200, and the second display screen 20 is the inner screen of the terminal device 200. The terminal device 200 is in a folded state or a flattened state based on the rotation of the hinge assembly 30. In the folded state, the terminal device 100 uses the first display screen 10.

[0076] Referring to FIG. 2 , which is a perspective view of a terminal device according to an embodiment of the present application in a flattened state, a user can drive the hinge assembly 30 to rotate the terminal device 200 from the folded state shown in FIG. 1 to the flattened state shown in FIG. 2 . In the flattened state, the terminal device 100 utilizes the second display screen 20 , which includes a first display area 21 , a second display area 22 , and a third display area 23 .

[0077] 3 , which is a perspective view of another foldable terminal device according to an embodiment of the present application, the terminal device 200 includes a display screen 40 and a hinge assembly 50 . The display screen 40 includes a first display area 41 , a second display area 42 , and a third display area 43 .

[0078] Taking the hinge assembly of a foldable terminal as an example, the hinge assembly is a critical component that enables both the folding and unfolding functions of the foldable terminal. However, during the assembly process of the foldable terminal, due to limitations in assembly technology, the assembled hinge assembly is prone to defects such as under-expansion and over-expansion, resulting in a low mass production yield of the hinge assembly. Under-expansion and over-expansion of the hinge assembly can cause the display screen of the foldable terminal to be under-expansion or over-expansion accordingly. Under-expansion of the display screen refers to a flattening angle of less than 180 degrees when the foldable terminal is flattened, i.e., the angle between the first display area 21 or the third display area 23 and the second display area 22 is less than 180 degrees. Over-expansion of the display screen refers to a flattening angle of greater than 180 degrees when the foldable terminal is flattened, i.e., the angle between the first display area 21 or the third display area 23 and the second display area 22 is greater than 180 degrees. Both under-expansion and over-expansion can cause the display screen of the foldable terminal to be uneven when flattened, affecting the image quality of each display area and reducing the user experience.

[0079] Due to the low precision assembly threshold and the large processing tolerance of materials, the current assembly of foldable terminal devices cannot adopt the blind assembly method, and tolerance grading technology needs to be adopted. However, the structure of the hinge assembly is complex, and the flattening angle is easily affected by various factors such as material deformation and processing tolerance fluctuations. It is technically difficult to establish a high-accuracy automatic grading solution. In the scheme of grading and matching based on traditional tolerance simulation, the grading threshold needs to be calculated in a step-by-step manner using the Design of Experiments (DOE). Continuous debugging is required to find the threshold with the optimal yield. The calculation method is not intelligent enough, and each time requires a long calculation cycle (for example, two days), and the calculation time is long. In addition, the DOE step calculation has a large gradient. For example, if 10% is used as a threshold step, it is easy to miss the optimal threshold solution.

[0080] In order to solve the above problems, an embodiment of the present application provides a component size grading method, which performs intelligent and automatic optimization of the component size grading scheme, reduces the difficulty of implementing component size grading, shortens the time required to determine the component size grading scheme of the mechanical component, and effectively improves the assembly yield of the mechanical component, for example, improving the yield of the flattening angle of the shaft component.

[0081] Refer to FIG4 , which is a flow chart of a component size binning method provided in one embodiment of the present application. The method is applied to a production device and includes:

[0082] S101, obtaining multiple dimension data of multiple components.

[0083] In one embodiment of the present application, multiple dimension data of multiple components can be measured by a measuring device. In another embodiment of the present application, the memory of the production device can also pre-store multiple dimension data of multiple components, so that the multiple dimension data of multiple components can be obtained from the memory of the production device. In another embodiment of the present application, the production device can also receive input of multiple dimension data of multiple components. In one embodiment of the present application, the multiple components can be components used to be assembled into a preset component, for example, the preset component is a hinge assembly of a foldable terminal device, and the multiple components include a main swing arm, a base, a secondary swing arm, and a wedge block of the hinge assembly.

[0084] S102, determining multiple binning schemes for multiple components based on multiple dimension data.

[0085] In one embodiment of the present application, binning levels of multiple dimensional data of multiple components and binning thresholds for each level are set, and the binning levels of multiple dimensional data of multiple components and binning thresholds for each level are determined as multiple binning schemes for multiple components.

[0086] S103: determining a matching scheme for the plurality of components according to the plurality of grading schemes for the plurality of components.

[0087] In one embodiment of the present application, the quantile corresponding to each bin threshold of multiple dimensional data of multiple components is determined according to the inverse cumulative distribution function, the multiple dimensional data of the multiple components are grouped according to the quantile corresponding to each bin threshold, and the grouped dimensional data of different components are matched to obtain a matching scheme for the multiple dimensional data of the multiple components.

[0088] S104, calculating the fitness score of each matching scheme.

[0089] In one embodiment of the present application, a fitness function is used to calculate the fitness score of each combination scheme based on the group size data of different components. The group size data of different components corresponding to each combination scheme is input into a prediction model to obtain multiple prediction results for each combination scheme. The fitness function is then used to calculate the fitness score of each combination scheme based on the multiple prediction results.

[0090] In one embodiment of the present application, the pre-set component is a rotating shaft component, which has a flattening angle when flattened. The fitness function is a formula for calculating the yield of all predicted flattening angles under a given combination scheme, and the prediction model is a flattening angle prediction model. Grouped dimension data of different components corresponding to each combination scheme is input into the flattening angle prediction model to obtain multiple predicted flattening angles for each combination scheme. The yield of each predicted flattening angle is calculated for each combination scheme, and a fitness score is obtained for each combination scheme.

[0091] S105: Determine a matching scheme for the preset components based on the fitness score of each matching scheme.

[0092] In one embodiment of the present application, if the fitness score of any combination scheme is greater than or equal to a preset fitness score threshold, the combination scheme is determined as the combination scheme of the preset component. In another embodiment of the present application, the combination scheme with the highest fitness score among multiple combination schemes can also be determined as the combination scheme of the preset component.

[0093] In one embodiment of the present application, the component size classification method also includes: extracting multiple features that affect the flattening angle of the shaft assembly, wherein the multiple features include multiple dimensional data of multiple components of the shaft assembly; and establishing a flattening angle prediction model based on the multiple features and the flattening angle data.

[0094] In one embodiment of the present application, multiple features that affect the flattening angle of the shaft assembly are extracted, including: obtaining force information of multiple components of the shaft assembly in a flattened state; obtaining sampling data corresponding to the dimensional data of the components based on the force information of the components of the shaft assembly in the flattened state; and performing regression analysis on the sampling data corresponding to the dimensional data of the components to obtain multiple features that affect the flattening angle of the shaft assembly.

[0095] In one embodiment of the present application, regression analysis is performed on the sampled data corresponding to the dimensional data of the component to obtain multiple features that affect the flattening angle of the shaft assembly, including: standardizing the sampled data; dividing the sampled data into a training set, and using the training set to train a fitted cable regression model; obtaining coefficients of all sampled data based on the trained cable regression model; using the absolute value of the coefficient as the contribution of the dimensional data of the component to the flattening angle, sorting the dimensional data of the component in descending order of contribution, and selecting multiple dimensional data ranked first according to a preset percentage to obtain multiple features.

[0096] In one embodiment of the present application, a flattening angle prediction model is established based on multiple features and flattening angle data, including: training a preset deep learning model based on the multiple features and flattening angle data to obtain the flattening angle prediction model.

[0097] In one embodiment of the present application, a preset deep learning model is trained based on multiple features and flattening angle data to obtain a flattening angle prediction model, including: establishing a training sample set based on multiple groups of features and the flattening angle data corresponding to each group of features; using a group of features and the corresponding flattening angle data in the training sample set as training data, inputting the preset deep learning model, and calculating the loss function value of the preset deep learning model; if the loss function value of the preset deep learning model is greater than the preset value, adjusting the parameters of the preset deep learning model, and continuing to train the preset deep learning model until the loss function value of the preset deep learning model is less than or equal to the preset value, determining that the training of the preset deep learning model is completed, and determining the trained preset deep learning model as the flattening angle prediction model.

[0098] Refer to FIG5 , which is a flow chart of a component size binning method provided in another embodiment of the present application. The method is applied to a production device and includes:

[0099] S201 , obtaining a plurality of dimension data of a plurality of first components and a plurality of second components.

[0100] S202 , determining a plurality of grading schemes for a plurality of first components and a plurality of second components according to a plurality of size data.

[0101] S203 : Determine a matching scheme of the plurality of first components and the plurality of second components according to the plurality of grading schemes of the plurality of first components and the plurality of second components.

[0102] S204: Calculate the fitness score of each matching scheme.

[0103] S205 , determining a combination scheme of multiple preset components according to the fitness score of each combination scheme, wherein each preset component includes a first component and a second component.

[0104] Refer to FIG6 , which is a flow chart of a component size binning method provided by another embodiment of the present application. The method is applied to a production device and includes:

[0105] S301 , extracting multiple features that affect the flattening angle of the shaft assembly, where the multiple features include dimensional data of multiple components.

[0106] In one embodiment of the present application, multiple features that affect the flattening angle of the shaft assembly are extracted based on a design drawing of the shaft assembly. The design drawing of the shaft assembly is a three-dimensional model image, which is generated by pre-proportioning a shaft assembly of standard dimensions. The design drawing of the shaft assembly is input into a finite element simulation software. The finite element simulation software is used to perform a force analysis on multiple components of the shaft assembly in a flattened state to determine the force information of the multiple components. The force information of the multiple components includes force contact points and force deformation amounts. The force contact points and force deformation amounts of the multiple components are input into a tolerance simulation software. The tolerance simulation software is used to perform Monte Carlo sampling on the dimensional data of the multiple components to determine the sampling data corresponding to the dimensional data of the multiple components. The sampling data is subjected to regression analysis to determine multiple features in the sampling data, wherein the features include the dimensional data of the multiple components, for example, the dimensional data includes the position tolerance of the components.

[0107] S302: Establish a flattening angle prediction model based on multiple features and flattening angle data.

[0108] In one embodiment of the present application, a preset deep learning model is trained based on multiple features and flattening angle data to obtain a flattening angle prediction model. The preset deep learning model is a fully connected neural network model, and the multiple features and flattening angle data serve as training data for the fully connected neural network model. The multiple features are used as input data for the fully connected neural network model, and the flattening angle data is used as output data for the fully connected neural network model. The fully connected neural network model is trained, and the trained fully connected neural network model is used as the flattening angle prediction model. In other embodiments of the present application, the preset deep learning model may also be other types of machine learning models, such as a convolutional neural network model.

[0109] S303: Determine multiple binning schemes and component matching schemes for the multiple components based on the flattening angle prediction model and the genetic algorithm. The binning scheme includes binning levels and bin thresholds for the dimensional data of each component. The matching scheme includes matching the dimensional data levels of different components.

[0110] In one embodiment of the present application, the binning level of the dimensional data of each component is customized, for example, it can be set to 2 to n levels, where n can be an integer greater than 2. The binning thresholds of the dimensional data of each component are traversed, and the quantile of the binning threshold of each customized bin is determined based on the inverse cumulative distribution function. Monte Carlo sampling calculations of different binning thresholds are performed using a genetic algorithm. The dimensional data of different levels of multiple components are matched, and the dimensional data of different levels of the matched multiple components are input into a flattening angle prediction model to obtain corresponding predicted flattening angles. The comprehensive defective rate of each level after matching is determined based on the predicted flattening angle and a preset flattening angle range. The optimal binning and matching scheme is determined based on the minimum value of the comprehensive defective rate.

[0111] The above-mentioned embodiments of the present application determine multiple features that affect the flattening angle in the shaft assembly through finite element force analysis, Monte Carlo sampling and regression analysis, and train a preset deep learning model based on the features to establish a flattening angle prediction model. The grading scheme and matching scheme of the component's dimensional data are automatically optimized based on the flattening angle prediction model and genetic algorithm, thereby improving the efficiency and accuracy of grading the dimensional data of the shaft assembly and effectively improving the assembly yield of the shaft assembly.

[0112] Please refer to FIG7 , which is a flowchart of feature extraction provided in one embodiment of the present application.

[0113] S3011, inputting the design drawing of the rotating shaft assembly into the finite element simulation software to obtain the force information of multiple components of the rotating shaft assembly in a flattened state.

[0114] In one embodiment of the present application, the design drawing of the shaft assembly is a finite element model generated by pre-proportioning a shaft assembly of standard dimensions. For example, the finite element simulation software may be ANSYS software or Abaqus software, which can perform force analysis on multiple components of the shaft assembly in a flattened state and determine the force information of multiple components in the shaft assembly. The flattened state means that the flattening angle of the shaft assembly is 180 degrees. The force information of multiple components includes, but is not limited to: the force contact point m of the component; i and the load deformation δ i , where the force contact point m i The three-dimensional coordinate values ​​in a three-dimensional rectangular coordinate system are used for representation, and the three-dimensional rectangular coordinate system can be established based on the position of the shaft assembly. In one embodiment of the present application, the multiple components of the shaft assembly include, but are not limited to: a main swing arm, a secondary swing arm, a base, a wedge block, a hinge, a main shaft, a support portion, and a slider.

[0115] In one embodiment of the present application, a geometric model of a shaft assembly (e.g., a design drawing of the shaft assembly) is created or imported using finite element simulation software. The geometric model of the shaft assembly is discretized into a finite element mesh. Material properties, such as elastic modulus, Poisson's ratio, and density, are defined for each component of the shaft assembly. These properties determine the material's response to external loads. Boundary conditions and loading are defined on the finite element model of the shaft assembly. Boundary conditions and loading include constraints (e.g., fixed supports or axial constraints) and loading conditions (e.g., force, pressure, temperature, etc.). The boundary conditions and loading in this embodiment of the present application assume that the shaft assembly is in a flattened state. The finite element model of the shaft assembly is solved using a numerical method (e.g., the finite element method), solving a system of linear or nonlinear equations on the mesh to obtain the node displacements and stress distributions in the finite element model. The results are visualized and analyzed using the finite element simulation software. By examining the displacement, stress, deformation, and other results, the stress and deformation of the shaft assembly under a given load (i.e., in the flattened state) are evaluated, and the contact conditions between the multiple components of the shaft assembly and the stress deformation of the multiple components are determined. The contact between multiple components is represented by force-bearing contact points.

[0116] S3012: Input the force information of the multiple components of the rotating shaft assembly in the flattened state into the tolerance simulation software to obtain sampling data corresponding to the dimensional data of the multiple components.

[0117] In one embodiment of the present application, the tolerance simulation software is Monte Carlo simulation software. Based on the force information of the multiple components of the shaft assembly in a flattened state, the tolerance simulation software performs a predetermined number of Monte Carlo sampling on the dimensional data of the multiple components to obtain sampled data corresponding to the dimensional data of the multiple components. For example, the predetermined number of samplings may be one million, one million and two hundred thousand, or another number, which is not limited to this embodiment of the present application. The dimensional data of the multiple components includes, but is not limited to, positional tolerances of the surfaces of the components.

[0118] In one embodiment of the present application, the force information of multiple components is input as virtual features into the three-dimensional tolerance simulation software of the mechanism, and the dimensional data of each component matching the force information of the corresponding component is output through the tolerance simulation software. The dimensional data of each component matching the force information of the corresponding component is subjected to a preset number of Monte Carlo sampling to obtain the sampling data corresponding to the dimensional data of each component.

[0119] S3013: Perform regression analysis on the sampled data corresponding to the dimensional data of the multiple components to obtain multiple features that affect the flattening angle of the shaft assembly.

[0120] In one embodiment of the present application, a regression analysis is performed on sampled data corresponding to the dimensional data of multiple components using a preset regression analysis model to obtain the contribution of each type of dimensional data for each component, identify key influencing factors, and sort the multiple dimensional data of the multiple components in descending order according to the contribution. In other words, the multiple dimensional data of the multiple components are sorted in descending order of contribution, and the component dimensional data that ranks at the top by a preset percentage is used as the feature that affects the flattening angle of the shaft assembly. For example, the preset regression analysis model is a LASSO regression analysis model, and the preset percentage is 90%, 95%, or another percentage.

[0121] In one embodiment of the present application, the sampled data is standardized, for example, the sampled data is standardized to a mean of 0 and a standard deviation of 1, eliminating the differences in measurement units and proportions between different feature variables. The sampled data is divided into a training set and a test set. The training set includes multiple groups of training data, the input data of each group of training data is sampled data, that is, the dimensional data of the component, and the output data is the flattening angle of the shaft assembly. The test set includes multiple groups of test data, the input data of each group of test data is sampled data, that is, the dimensional data of the component, and the output data is the flattening angle of the shaft assembly. The training set is used to fit the LASSO regression model. During the fitting process, the sparsity of the features is controlled by adjusting the L1 regularization parameter. For example, the optimal L1 regularization parameter can be found using methods such as cross-validation or grid search. Based on the trained Lasso regression model, the coefficients of all features are obtained, the absolute values ​​of the coefficients are sorted in descending order, and the features are screened out. The performance of the model is evaluated using the test set, for example, using indicators such as root mean square error (RMSE) and R square to evaluate the model performance.

[0122] In one embodiment of the present application, the absolute value of the coefficient is used as the contribution of each feature (i.e., the dimensional data of each element) to the flattening angle of the shaft assembly, and multiple features are sorted in descending order according to the contribution, and the features that rank at the top by a preset percentage are selected to obtain the final features.

[0123] The above-mentioned embodiment of the present application can accurately screen out multiple features that affect the flattening angle in the shaft assembly through finite element force analysis, Monte Carlo sampling and regression analysis.

[0124] Please refer to FIG8 , which is a flowchart of model training provided in one embodiment of the present application.

[0125] S3021: Establish a training sample set based on multiple groups of features and the flattening angle data corresponding to each group of features.

[0126] In one embodiment of the present application, the features corresponding to each flattening angle data include dimensional data of multiple components, and the dimensional data of the components are dimensional data of the key structural feature surfaces of the components. For example, the multiple components include a main swing arm A, a base B, an auxiliary swing arm C, and a wedge block D. Refer to Figure 9, which is a stereoscopic view of the main swing arm provided in one embodiment of the present application. The dimensional data of the key structural feature surfaces of the main swing arm 11 include the position tolerance A1 of the matching hole 111 between the main swing arm and the wedge block, the position tolerance A2 of the stop surface 112 between the main swing arm and the base, and the position tolerance A3 of the matching arc surface 113 between the main swing arm and the base. Refer to Figures 10-12, which are stereoscopic views of the base provided in one embodiment of the present application. The dimensional data of the key structural feature surfaces of the base 12 include the position tolerance B1 of the matching hole pin 121 between the base and the auxiliary swing arm, the position tolerance B2 of the stop surface 122 between the base and the auxiliary swing arm, the position tolerance B3 of the matching arc surface 123 between the base and the main swing arm, and the position tolerance B4 of the stop surface 124 between the base and the main swing arm. Refer to Figure 13, which is a three-dimensional view of the auxiliary swing arm provided in an embodiment of the present application. The dimensional data of the key structural feature surfaces of the auxiliary swing arm 13 include the position tolerance C1 of the matching surface 131 between the auxiliary swing arm and the wedge block, the position tolerance C2 of the stop surface 132 between the auxiliary swing arm and the base, and the position tolerance C3 of the matching hole 133 between the auxiliary swing arm and the base. Refer to Figures 14-16, which are three-dimensional views of the wedge block provided in an embodiment of the present application. The dimensional data of the key structural feature surfaces of the wedge block 14 include the position tolerance D1 of the matching hole 141 between the wedge block and the main swing arm, and the position tolerance D2 of the matching surface 142 between the wedge block and the auxiliary swing arm.

[0127] S3022: A set of features and corresponding flattening angle data are used as training data, input into a preset deep learning model, and a loss function value of the preset deep learning model is calculated.

[0128] Refer to Figure 17, which is a schematic diagram of the structure of a preset deep learning model provided in one embodiment of the present application. In one embodiment of the present application, the preset deep learning model is a fully connected neural network model (also known as a multi-layer perceptron). The fully connected neural network model is initialized and imported into the deep learning framework, and initial parameters are set. The set of features is then used as input data for the fully connected neural network model, and the flattening angle data is used as output data for the fully connected neural network model. The fully connected neural network model is trained, and the loss function of the preset deep learning model is calculated.

[0129] In one embodiment of the present application, a fully connected neural network model includes an input layer, two hidden layers, and an output layer. The input layer includes 34 neurons, corresponding to the first parameter of the linear transformation function nn.Linear(34, 128). The first hidden layer is a fully connected layer, defined by the linear transformation function self.fc1 = nn.Linear(34, 128), and includes 128 neurons. The first hidden layer also includes a Dropout layer (self.dropout1 = nn.Dropout(p = 0.5)) to reduce overfitting. The second hidden layer is also a fully connected layer, defined by the linear transformation function self.fc2 = nn.Linear(128, 64), and includes 64 neurons. The second hidden layer also includes a Dropout layer (self.dropout2 = nn.Dropout(p = 0.5)) to reduce overfitting. The output layer is also a fully connected layer, defined by the linear transformation function self.fc3 = nn.Linear(64, 1), and includes 1 neuron, which is used to generate the final output of the fully connected neural network model.

[0130] In one embodiment of the present application, the activation function of the fully connected neural network model is self.relu=nn.ReLU(), and the operation code of the fully connected neural network model is:

[0131] def__init__(self):

[0132] super(Net,self).__init__();

[0133] self.fc1=nn.Linear(34,128);

[0134] self.dropout1=nn.Dropout(p=0.5);

[0135] self.fc2=nn.Linear(128,64);

[0136] self.dropout2=nn.Dropout(p=0.5);

[0137] self.fc3 = nn.Linear(64,1);

[0138] self.relu = nn.ReLU().

[0139] In one embodiment of the present application, the loss function of the preset deep learning model is data loss, and the data loss is the mean square error (MSE) between the predicted flattening angle and the actual flattening angle data.

[0140] S3023: Determine whether the loss function value of the preset deep learning model is less than or equal to a preset value. If the loss function value of the preset deep learning model is greater than the preset value, execute S3024; if the loss function value of the preset deep learning model is less than or equal to the preset value, execute S3025.

[0141] S3024: Adjust the parameters of the preset deep learning model, and then return to S3022. The parameters of the fully connected neural network model include the connection weights between different layers, the bias value of each neuron, etc.

[0142] S3025: Determine whether the training of the preset deep learning model is completed, and determine the trained preset deep learning model as the flattening angle prediction model.

[0143] In one embodiment of the present application, if the loss function of the preset deep learning model is less than or equal to a preset value, it is determined that the preset deep learning model has converged, and further it is determined that the training of the preset deep learning model is completed.

[0144] The above-mentioned embodiment of the present application uses the characteristics of the rotating shaft assembly and the flattening angle data to train the neural network model to obtain the flattening angle prediction model. By inputting the component size data of the rotating shaft assembly into the flattening angle prediction model, the corresponding predicted flattening angle data can be quickly obtained, thereby improving the calculation efficiency of the flattening angle data.

[0145] Please refer to FIG. 18 , which is a flowchart of optimizing the bin positions and bin thresholds provided in one embodiment of the present application.

[0146] S3031, determining a fitness function of a matching scheme of size data of multiple components.

[0147] In one embodiment of the present application, the fitness function for determining a matching scheme for the dimensional data of multiple components is a calculation formula for the yield of all flattening angle prediction values ​​under the matching scheme, and the yield of all flattening angle prediction values ​​corresponding to each matching scheme is used as a fitness score. The dimensional data of multiple components corresponding to each matching scheme are input into a flattening angle prediction model to obtain multiple flattening angle prediction values ​​corresponding to each matching scheme. A determination is made as to whether each flattening angle prediction value is within a preset flattening angle range. If the flattening angle prediction value is within the preset flattening angle range, the flattening angle prediction value is determined to be qualified. If the flattening angle prediction value is not within the preset flattening angle range, the flattening angle prediction value is determined to be unqualified. The ratio between all qualified flattening angle prediction values ​​and all flattening angle prediction values ​​is calculated to obtain the yield of all flattening angle prediction values. The matching scheme includes the matching of the dimensional data of different components. For example, the preset flattening angle range is 180.2 degrees to 181.8 degrees, that is, greater than or equal to 180.2 degrees and less than or equal to 181.8 degrees.

[0148] In another embodiment of the present application, the fitness function of the matching scheme of the dimensional data of multiple components can also be the reciprocal of the sum of the mean square errors of the dimensional data in each bin of each component in the matching scheme, the reciprocal of the sum of the standard deviations of the dimensional data in each bin of each component in the matching scheme, and the sum of the yields of all flattening angle prediction values.

[0149] In one embodiment of the present application, the binning level of the component's dimensional data is the number of levels after the size is binned, and the binning threshold is the percentage value for binning the dimensional data. Refer to Figure 19, which is a schematic diagram of the binning levels and binning thresholds provided in one embodiment of the present application. For example, for multiple dimensional data of a component that conforms to a normal distribution, the number of binning levels is 5, the number of binning thresholds is 4, binning threshold 1 is 15%, binning threshold 2 is 30%, binning threshold 3 is 70%, and binning threshold 4 is 85%. The binning levels include: bin 1 is 0% to 15%, bin 2 is 15% to 30%, bin 3 is 30% to 70%, bin 4 is 70% to 85%, and bin 5 is 85% to 100%.

[0150] S3032: Set binning levels for the size data of multiple components and a binning threshold for each level, and determine the quantile corresponding to each binning threshold based on an inverse cumulative distribution function.

[0151] In one embodiment of the present application, a genetic algorithm is used to optimize the binning levels and binning thresholds of the dimensional data of multiple components. First, an initial population representing different binning schemes is randomly generated. Multiple components and multiple dimensional data of each component are initialized. The binning levels of the multiple dimensional data of each component are customized to 2 to n levels, where n is a positive integer greater than 2, for example, n is set to 4. For example, the dimensional data of the component is the positional tolerance of the key structural feature surface of the component, and the key structural feature surface refers to a specific structural feature that plays an important role in the performance, reliability, or function of the component.

[0152] In one embodiment of the present application, the target probability is determined, and the dimensional data corresponding to the target probability is calculated using the inverse cumulative distribution function of the preset probability distribution function. The target probability is the bin threshold of multiple dimensional data of an element, and the dimensional data corresponding to each target probability is used as the quantile corresponding to the bin threshold. The multiple dimensional data of the initialized element conform to the normal distribution, and the inverse cumulative distribution function is the inverse cumulative distribution function corresponding to the normal distribution. The inverse cumulative distribution function maps the target probability to the corresponding cumulative distribution function value, and the cumulative distribution function value is used as the quantile corresponding to each target probability. Accordingly, the bin threshold is input into the inverse cumulative distribution function, and the quantile corresponding to the bin threshold is output through the inverse cumulative distribution function, thereby obtaining the dimensional data corresponding to each bin threshold in the multiple dimensional data.

[0153] In one embodiment of the present application, after customizing the binning levels of the dimensional data, the number of binning thresholds for the dimensional data is the level minus one. Each binning threshold of the dimensional data of each component is traversed, and each binning threshold is used as a target probability. The inverse cumulative distribution function is input to obtain the quantile corresponding to the binning threshold. The traversal method can be to traverse each binning threshold at a preset percentage interval, such as a preset percentage of 5%, or to traverse each binning threshold within a preset percentage range, such as a preset percentage range of 15% to 25%.

[0154] For example, if the bin level of the size data is customized to 4, then the number of bin thresholds of the size data is 3. The three bin thresholds are set to 20%, 50%, and 70% respectively. 20%, 50%, and 70% are input as target probabilities into the inverse cumulative distribution function corresponding to the normal distribution, and the quantile corresponding to the bin threshold of 20%, the quantile corresponding to the bin threshold of 50%, and the quantile corresponding to the bin threshold of 70% are obtained.

[0155] S3033: Grouping the size data of the multiple components according to the quantile corresponding to the bin threshold of each bin level.

[0156] In one embodiment of the present application, the dimension data of multiple components are grouped according to the user-defined binning levels and the quantiles corresponding to the binning thresholds of each binning level. For example, if a component has 4 binning levels and the binning thresholds are 20%, 50%, and 70%, the dimension data of the component is grouped according to the quantile corresponding to the binning threshold of 20%, the quantile corresponding to the binning threshold of 50%, and the quantile corresponding to the binning threshold of 70%, resulting in four groups of dimension data, namely, dimension data between 0% and 20%, dimension data between 20% and 50%, dimension data between 50% and 70%, and dimension data between 70% and 100%.

[0157] S3034: Calculate the fitness score of each matching solution based on the grouping size data of the multiple components using a fitness function.

[0158] In one embodiment of the present application, multiple gear positions of different components are arranged and combined to determine a matching scheme for the different components. For example, gear positions 1 to 4 of the main swing arm, gear positions 1 to 4 of the base, gear positions 1 to 4 of the auxiliary swing arm, and gear positions 1 to 4 of the wedge block are matched to generate a matching scheme. The grouped size data of the different components corresponding to the matching scheme is input into a flattening angle prediction model to obtain multiple predicted flattening angles corresponding to the matching scheme. The yield of the multiple predicted flattening angles of the matching scheme is calculated to obtain a fitness score for the matching scheme. The gear positions of different components can be freely matched during the binning process to increase the number of matching schemes, and are not limited to the above examples.

[0159] For example, the dimensional data of the main swing arm A are grouped into first gear AX1, second gear AX2, third gear AX3, and fourth gear AX4; the dimensional data of the base B are grouped into first gear BX1, second gear BX2, third gear BX3, and fourth gear BX4; the dimensional data of the auxiliary swing arm C are grouped into first gear CX1, second gear CX2, third gear CX3, and fourth gear CX4; and the dimensional data of the wedge block D are grouped into first gear DX1, second gear DX2, third gear DX3, and fourth gear DX4. AX1, BX1, CX1, DX1, AX2, BX1, CX1, DX1, ..., AX4, BX4, CX4, DX4 are matched as a matching scheme, which includes all combinations of gear matching between the main swing arm A, the base B, the auxiliary swing arm C, and the wedge block D. The Monte Carlo sampling method is used to select sampled dimensional data from the dimensional data of multiple components corresponding to the matching scheme, and the sampled dimensional data are input into the flattening angle prediction model to obtain the corresponding predicted flattening angle. After multiple samplings, all predicted flattening angles under the matching scheme are obtained, and the yield of all predicted flattening angles is calculated as the fitness score of the matching scheme.

[0160] S3035, determine whether the fitness score is greater than or equal to the preset fitness score threshold. If the fitness score is greater than or equal to the preset fitness score threshold, execute S3036; if the fitness score is less than the preset fitness score threshold, return to execute S3032, reset the binning levels and binning thresholds of the size data of multiple components, and re-determine the quantile corresponding to the binning threshold of each binning level based on the inverse cumulative distribution function. Among them, the fitness score is the yield of multiple predicted flattening angles corresponding to the binning scheme, and the preset fitness score threshold is the preset yield threshold. For example, the preset fitness score threshold is 94%, 95%, 96% or other values.

[0161] S3036: Determine the current matching solution as the optimal matching solution.

[0162] In one embodiment of the present application, the optimal matching solution includes the gear matching of the size data of different components corresponding to the fitness score less than or equal to the fitness score threshold. The binning solution corresponding to the optimal matching solution is the optimal binning solution.

[0163] Based on the selection process of the genetic algorithm, individuals with better performance are selected for reproduction based on the fitness score. Accordingly, the matching scheme with a higher flattening angle yield is selected as the better matching scheme. Based on the crossover process of the genetic algorithm, offspring are generated by exchanging the genes of the parent individuals. Accordingly, by adjusting the gear matching of the size data of different components, different matching schemes are obtained, and the flattening angle yield corresponding to different matching schemes is calculated. Based on the mutation process of the genetic algorithm, certain genes in the individuals are randomly changed to increase genetic diversity. Accordingly, by adjusting the grading level and corresponding grading threshold of the size data of each component, different grading schemes and matching schemes are obtained, and the flattening angle yield corresponding to different matching schemes is calculated. Based on the iterative process of the genetic algorithm, the selection, crossover and mutation processes are repeated, and each generation of the population is evaluated according to the fitness function until the preset number of iterations is reached or the fitness score is stable. Accordingly, the binning levels, binning thresholds, and level combinations of the dimensional data of multiple components are continuously adjusted, thereby adjusting the binning and combination schemes of the dimensional data of the multiple components. The fitness score, i.e., the flattening angle yield, of each combination scheme is calculated until the flattening angle yield is greater than or equal to a preset yield. The combination scheme with a flattening angle yield greater than or equal to the preset yield is then determined as the optimal combination scheme. Alternatively, when the number of adjustments to the combination scheme reaches a preset number, the combination scheme with the highest flattening angle yield among the multiple combination schemes is determined as the optimal combination scheme.

[0164] The above-mentioned embodiment of the present application adopts a genetic algorithm to automatically optimize the dimensional data grading scheme and the matching scheme, thereby improving the efficiency and intelligence of the tolerance grading of the component dimensional data of the shaft assembly.

[0165] In one embodiment of the present application, the size classification method further includes: automatically assembling the shaft assembly according to a determined matching scheme (such as an optimal matching scheme) of the size data of the plurality of components.

[0166] An embodiment of the present application also provides a computer storage medium, in which computer instructions are stored. When the computer instructions are executed on a production device, the production device executes the above-mentioned related method steps to implement the component size classification method in the above-mentioned embodiment.

[0167] Referring to FIG. 20 , an embodiment of the present application further provides a production device 300 comprising: an infeed component 301, a measuring component 302, a processing component 303, a sorting component 304, an outfeed component 305, and an assembly component 306. The processing component 303 is used to execute the above-mentioned component size classification method and determine the optimal matching scheme for the size data of multiple components, that is, the matching scheme for the preset assembly. The infeed component 301 includes, but is not limited to, a carrying device and a conveyor belt, and is used to receive multiple components to be assembled into a preset assembly and to convey the multiple components to the measuring component 302. The measuring component 302 is communicatively connected to the processing component 303, and is used to measure the size data of the multiple components and send the measured size data of the multiple components to the processing component 303. The sorting component 304 can be a robot, and is used to match the multiple components according to the matching scheme for the preset assembly and convey the multiple matched components to the outfeed component 305. The outfeed component 305 is used to output the multiple matched components. The outfeed component 305 can be a conveyor belt. The assembly component 306 is used to assemble the matched multiple components into a preset assembly. In other embodiments of the present application, the production equipment 300 may also not include the measurement component 302, and the size data of the multiple components can be measured in advance.

[0168] Refer to Figure 21, which is a schematic diagram of an application scenario of the component size binning method provided in an embodiment of the present application. Multiple components of different sizes are input into the production equipment 300. The production equipment 300 can apply the component size binning method provided in the embodiment of the present application to determine a matching scheme for multiple size data of multiple components, match the multiple components according to the determined matching scheme, and output multiple components that have been matched. For example, components A1 to An are first components of different sizes, components B1 to Bn are second components of different sizes, components C1 to Cn are third components of different sizes, and components D1 to Dn are fourth components of different sizes. Multiple components of different sizes are input into the production equipment 300, and the production equipment 300 outputs multiple components that have been matched. The multiple components that have been matched include a first component A1 of a specified size, a second component B2 of a specified size, a third component C3 of a specified size, and a third component D4 of a specified size.

[0169] An embodiment of the present application further provides a computer program product. When the computer program product is run on a computer, the computer is caused to execute the above-mentioned related steps to implement the component size classification method in the above-mentioned embodiment.

[0170] In addition, an embodiment of the present application also provides a device, which can specifically be a chip, component or module, and the device may include a connected processor and memory; wherein the memory is used to store computer-executable instructions, and when the device is running, the processor can execute the computer-executable instructions stored in the memory to enable the chip to execute the component size classification method in the above-mentioned method embodiments.

[0171] Among them, the production equipment, computer storage medium, computer program product or chip provided in this embodiment are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0172] Through the description of the above implementation methods, technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0173] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0174] Units described as separate components may or may not be physically separate, and components shown as units may be one physical unit or multiple physical units, that is, they may be located in one place or distributed in multiple places. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0175] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0176] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a device (which can be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not intended to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A component size classification method, applied to production equipment, characterized in that: The method comprises: Get multiple dimension data of multiple components; determining a plurality of binning schemes for the plurality of components according to the plurality of size data; Determining a matching scheme for the plurality of components according to a plurality of grading schemes for the plurality of components; Calculate the fitness score of each collocation scheme; Determine a matching scheme for the preset components based on the fitness score of each matching scheme.

2. The component size binning method according to claim 1, wherein: Determining a plurality of binning schemes for the plurality of components according to the plurality of size data includes: Setting the binning levels of the plurality of dimension data of the plurality of components and the binning threshold of each level; The binning levels of the plurality of size data of the plurality of components and the binning threshold of each level are determined as a plurality of binning schemes for the plurality of components.

3. The component size classification method according to claim 2, wherein: The determining of the matching scheme of the plurality of components according to the plurality of grading schemes of the plurality of components includes: Determine, according to an inverse cumulative distribution function, a quantile corresponding to each bin threshold of the plurality of size data of the plurality of components; Grouping the plurality of size data of the plurality of components according to the quantile corresponding to each bin threshold; The grouped size data of different components are matched to obtain a matching scheme of multiple size data of the multiple components.

4. The component size classification method according to claim 3, wherein: The determining, according to the inverse cumulative distribution function, the quantile corresponding to each bin threshold of the plurality of size data of the plurality of components comprises: The inverse cumulative distribution function of the normal distribution is used to calculate the size data corresponding to each bin threshold, and the calculated size data corresponding to each bin threshold is used as the quantile.

5. The component size classification method according to claim 3, wherein: The calculation of the fitness score of each collocation scheme includes: A fitness function is used to calculate the fitness score of each matching scheme according to the grouping size data of the different components.

6. The component size classification method according to claim 5, wherein: Determining a collocation scheme of preset components according to the fitness score of each collocation scheme includes: If the fitness score of any combination scheme is greater than or equal to a preset fitness score threshold, the any combination scheme is determined as the combination scheme of the preset component; or The collocation scheme with the highest fitness score among the multiple collocation schemes is determined as the collocation scheme of the preset component.

7. The component size binning method according to claim 5, wherein: The fitness function is used to calculate the fitness score of each matching scheme according to the grouping size data of the different components, including: Input the group size data of different components corresponding to each matching scheme into the prediction model to obtain multiple prediction results for each matching scheme; The fitness function is used to obtain a fitness score for each collocation scheme based on the multiple prediction results of each collocation scheme.

8. The component size binning method according to any one of claims 1 to 7, characterized in that: The preset component is a rotating shaft component, and the rotating shaft component has a flattening angle in a flattened state. The fitness function is a calculation formula for the yield of all flattening angle prediction values under the matching scheme.

9. The component size classification method according to claim 8, wherein: The fitness function is used to calculate the fitness score of each matching scheme according to the grouping size data of the different components, including: Inputting the grouped size data of different components corresponding to each matching scheme into the flattening angle prediction model to obtain multiple predicted flattening angles for each matching scheme; The yield of the multiple predicted flattening angles of each matching scheme is calculated to obtain a fitness score of each matching scheme.

10. The component size binning method according to claim 9, wherein: The method further comprises: extracting a plurality of features that affect the flattening angle of the rotating shaft assembly, wherein the plurality of features include a plurality of dimensional data of a plurality of components of the rotating shaft assembly; The flattening angle prediction model is established according to the multiple features and the flattening angle data.

11. The component size binning method according to claim 10, wherein: The extracting of multiple features that affect the flattening angle of the shaft assembly includes: Obtaining force information of the plurality of components of the rotating shaft assembly in a flattened state; acquiring sampling data corresponding to the dimensional data of the plurality of components according to the force information of the plurality of components in the flattened state of the rotating shaft assembly; Regression analysis is performed on sampling data corresponding to the dimensional data of the multiple components to obtain the multiple features that affect the flattening angle of the shaft assembly.

12. The component size binning method according to claim 11, wherein: The regression analysis of the sampled data corresponding to the dimension data of the component to obtain the multiple features affecting the flattening angle of the shaft assembly includes: performing standardization processing on the sampling data; Dividing the sampled data into a training set, and using the training set to train and fit the Lasso regression model; Obtaining coefficients of all sampled data according to the trained Lasso regression model; The absolute value of the coefficient is used as the contribution of the dimension data of the element to the flattening angle, the dimension data of the element are sorted in descending order of the contribution, and the plurality of dimension data sorted first are selected according to a preset percentage to obtain the plurality of features.

13. The component size binning method according to claim 10, wherein: The establishing the flattening angle prediction model according to the multiple features and the flattening angle data includes: A preset deep learning model is trained according to the multiple features and the flattening angle data to obtain the flattening angle prediction model.

14. The component size binning method according to claim 13, wherein: The step of training a preset deep learning model according to the plurality of features and the flattening angle data to obtain the flattening angle prediction model includes: Establish a training sample set based on multiple sets of features and the flattening angle data corresponding to each set of features; Using a set of features and corresponding flattening angle data in the training sample set as training data, inputting them into the preset deep learning model, and calculating the loss function value of the preset deep learning model; If the loss function value of the preset deep learning model is greater than a preset value, adjust the parameters of the preset deep learning model and continue to train the preset deep learning model until the loss function value of the preset deep learning model is less than or equal to the preset value. It is determined that the training of the preset deep learning model is completed, and the trained preset deep learning model is determined as the flattening angle prediction model.

15. A component size classification method, applied to production equipment, characterized in that: The method comprises: Acquiring a plurality of dimension data of a plurality of first components and a plurality of second components; determining a plurality of binning schemes for the plurality of first components and the plurality of second components according to the plurality of size data; Determining a matching scheme of the plurality of first components and the plurality of second components according to a plurality of binning schemes of the plurality of first components and the plurality of second components; Calculate the fitness score of each collocation scheme; According to the fitness score of each matching scheme, a matching scheme of multiple preset components is determined, wherein each preset component includes a first element and a second element.

16. The component size binning method according to claim 15, wherein: Determining a plurality of binning schemes for the plurality of first components and the plurality of second components according to the plurality of size data includes: Setting the binning levels of the plurality of size data of the plurality of first elements and the plurality of second elements and the binning threshold of each level; Binning levels of the plurality of size data of the plurality of first components and the plurality of second components and a binning threshold of each level are determined as a plurality of binning schemes of the plurality of first components and the plurality of second components.

17. The component size binning method according to claim 16, wherein: The determining of a matching scheme of the plurality of first components and the plurality of second components according to a plurality of grading schemes of the plurality of first components and the plurality of second components comprises: Determine, according to an inverse cumulative distribution function, a quantile corresponding to each bin threshold of the plurality of size data of the plurality of first elements and the plurality of second elements; grouping the plurality of size data of the plurality of first components and the plurality of second components according to the quantile corresponding to each bin threshold; The grouping size data of the plurality of first components and the plurality of second components are matched to obtain a matching scheme of a plurality of size data of the plurality of first components and the plurality of second components.

18. The component size binning method according to claim 17, wherein: The calculation of the fitness score of each collocation scheme includes: A fitness function is used to calculate a fitness score of each matching solution according to the grouping size data of the plurality of first components and the plurality of second components.

19. The component size binning method according to any one of claims 15 to 18, characterized in that: The preset component is a rotating shaft component, and the rotating shaft component has a flattening angle in a flattened state. The fitness function is a calculation formula for the yield of all flattening angle prediction values under the matching scheme.

20. The component size binning method according to claim 19, wherein: The step of using a fitness function to calculate the fitness score of each matching scheme according to the grouping size data of the plurality of first components and the plurality of second components includes: Inputting the group size data of the plurality of first components and the plurality of second components corresponding to each matching scheme into a flattening angle prediction model to obtain a plurality of predicted flattening angles for each matching scheme; The yield of the multiple predicted flattening angles of each matching scheme is calculated to obtain a fitness score of each matching scheme.

21. A production equipment, characterized in that, The production equipment includes: an infeed component for receiving a plurality of components; A processing component, configured to execute the component size binning method according to any one of claims 1 to 14, and determine a matching scheme for preset components; A sorting component, used for matching the plurality of components according to the matching scheme of the preset components; The discharging component is used to output the multiple components that have been matched.

22. The production equipment according to claim 21, characterized in that The production equipment also includes: The measuring component is used to measure a plurality of dimension data of the plurality of components and send the measured plurality of dimension data of the plurality of components to the processing component.

23. The production equipment according to claim 21, characterized in that The production equipment also includes: The assembling component is used to assemble the plurality of matched elements into the preset assembly.

24. A production equipment, characterized in that, The production equipment includes: an input component for receiving a plurality of first components and a plurality of second components; A processing component, configured to execute the component size binning method according to any one of claims 15 to 20, and determine a matching scheme of a plurality of preset components; A sorting component, configured to match the plurality of first components and the plurality of second components according to a matching scheme of the plurality of preset components; The discharging component is used to output the plurality of first components and the plurality of second components that have been matched.

Images