A method and system for sensitivity analysis of key assembly error elements

Abstract

The present invention provides a method and system for sensitivity analysis of key assembly error elements, including: S1: construction of an assembly error transmission mechanism model under the action of multiple error sources; S2: construction of an assembly error transmission data model supported by small sample data; S3: construction of a mechanism-data organic fusion model; S4: construction of a Sobol global sensitivity calculation model for assembly accuracy data analysis; S5: global sensitivity calculation driven by mechanism-data fusion; S6: identification of assembly key error elements under sensitivity numerical analysis; S7: verification of the effectiveness of a sensitivity analysis method for key assembly error elements. The present invention comprehensively considers the influence of the fusion of mechanism and data on the sensitivity analysis results, and can realize rapid and effective identification of key error elements of complex products.

Description

[Technical field]

[0001] The present invention relates to the technical field of digital assembly coordination process of mechanical products in the field of mechanical manufacturing, and in particular to a sensitivity analysis method and system for key assembly error elements. [Background technology]

[0002] The identification of assembly error elements is an important part of the assembly process of aerospace products, which directly affects the optimization efficiency of errors in each link of the product and the service performance of the product. In the optimization process of the assembly error link of modern products, tolerance analysis software and statistical analysis methods are often used to identify key error elements. The effectiveness of error identification depends on the precise construction of the assembly error transfer model. The existing sensitivity analysis methods of key assembly error elements are generally carried out in tolerance analysis software such as 3DCS or VSA. However, the sensitivity analysis results in the above tolerance analysis software are mostly calculated under ideal conditions, and the relevant data generated on site and the deformation calculation data are not fully utilized, which makes the sensitivity analysis results not very adaptable to the assembly site.

[0003] With the continuous improvement of the digitalization and intelligence of modern product production, modern manufacturing products have higher requirements for product optimization and production efficiency, making it difficult for the traditional identification of key error links in product assembly based on engineering experience to meet current production needs; at the same time, for the current simulation-based key error link identification method, there are problems such as failure to consider the actual physical property data on site and insufficient data utilization in the assembly simulation modeling stage. At the same time, most of them fail to consider the deformation caused by the force during the product assembly process. The sensitivity analysis results of the assembly error elements obtained based on this have limited improvement effects in the subsequent optimization links, which brings certain difficulties to further improve the product assembly accuracy.

[0004] Therefore, it is necessary to study a key assembly error element sensitivity analysis method and system to address the deficiencies of the prior art in order to solve or alleviate one or more of the above problems. [Summary of the invention]

[0005] In view of this, the present invention provides a key assembly error element sensitivity analysis method and system. Based on the mechanism and data fusion drive, a key assembly error element sensitivity analysis method driven by mechanism and data fusion is constructed. The influence of the fusion of mechanism and data on the sensitivity analysis results is comprehensively considered, and the key error elements of complex products can be quickly and accurately identified.

[0006] In one aspect, the present invention provides a key assembly error element sensitivity analysis method, which is used to identify key error elements in the production process of composite wing box assembly parts, and the analysis method comprises the following steps:

[0007] S1: According to the assembly process and structural characteristics of the product, the assembly error transmission mechanism model under the action of multiple error sources is constructed;

[0008] S2: Based on the assembly error transmission mechanism model and finite sample data, an assembly error transmission data prediction model is constructed;

[0009] S3: Construct a mechanism-data organic fusion model through the assembly error transmission mechanism model and the assembly error transmission data prediction model;

[0010] S4: Construct a global sensitivity calculation model for assembly accuracy data analysis through a mechanism-data organic fusion model;

[0011] S5: Global sensitivity calculation is performed through the mechanism-data organic fusion model and the Sobol global sensitivity calculation model;

[0012] S6: Identify assembly key error elements through global sensitivity calculation results;

[0013] S7: Verify the effectiveness of the assembly key error element identification results.

[0014] According to the above aspects and any possible implementation manner, an implementation manner is further provided, wherein S1 specifically includes:

[0015] S11: According to the assembly process and structural characteristics of the product, determine the various error sources existing in the product assembly process, obtain relevant data of the error sources, and use a small displacement screw model to represent the changes of the corresponding characteristics in the assembly process;

[0016] S12: According to the assembly relationship between the components and the positioning and clamping conditions of the tooling, combined with the error sources analyzed in S11, an assembly error transmission relationship diagram at the key assembly link is established;

[0017] S13: According to the assembly error transfer relationship diagram and the small displacement screw model, the Jacobi screw model is used to construct the assembly error transfer mechanism model.

[0018] According to the above aspects and any possible implementation manner, an implementation manner is further provided, wherein S2 specifically includes:

[0019] S21: according to the structure of the assembly error mechanism model, determine the collection requirements of the assembly error transfer data prediction model, and define the data output, wherein the output data is the deviation between the assembly accuracy calculation value of the mechanism model and the measured assembly value;

[0020] S22: Determine data collection requirements based on data output, where the collection requirements include input data and training data in the model training and construction process;

[0021] S23: Limited data collection is performed on input data and training data, and the data is preprocessed to remove outliers. The assembly error transfer data prediction model is optimized through a support vector regression prediction model, and the numerical values ​​of the mechanism calculation assembly accuracy and the measured assembly accuracy are predicted through the optimized prediction model.

[0022] According to the above aspects and any possible implementation, an implementation is further provided, wherein S3 specifically includes:

[0023] S31: Obtain product data, substitute the product data into the assembly error transmission mechanism model, calculate the mechanism theory calculation value of the assembly accuracy, obtain the error transmission measured data, use the error transmission measured data and the mechanism theory calculation value as training samples, and the training prediction value is the assembly error transmission data prediction model of the assembly mechanism calculation deviation;

[0024] S32: obtaining a predicted value of a mechanism calculation deviation through an assembly error transfer data prediction model;

[0025] S33: adding the calculated assembly accuracy mechanism calculation deviation prediction value to the assembly accuracy mechanism calculation theoretical value to obtain a compensated assembly accuracy calculation value;

[0026] S34: A mechanism-data organic fusion model for assembly accuracy prediction is constructed through the compensated assembly accuracy calculation value.

[0027] According to the aspects described above and any possible implementation methods, an implementation method is further provided, wherein the error transfer measured data includes but is not limited to the assembly accuracy measured data collected on-site, the process data read in the assembly process plan, related simulation data and error data.

[0028] According to the above aspects and any possible implementation manner, an implementation manner is further provided, wherein S4 specifically includes:

[0029] S41: setting the value range of each assembly parameter and performing random sampling in the finite data, calculating the theoretical calculation accuracy of the mechanism model according to the sampling results, and establishing the first matrix and the second matrix according to the sampling data;

[0030] S42: exchanging the same columns in the first matrix and the second matrix to obtain a third matrix corresponding to the first matrix and a fourth matrix corresponding to the second matrix;

[0031] S43: inputting the error source parameters in the first matrix, the second matrix, the third matrix and the fourth matrix into a mechanism-data organic fusion model, and obtaining four groups of assembly accuracy data sets;

[0032] S44: The first-order sensitivity and global sensitivity of four sets of assembly parameters are calculated by using four sets of assembly accuracy data sets;

[0033] S45: The Sobol global sensitivity calculation model is constructed through four sets of assembly accuracy data sets, first-order sensitivity and global sensitivity.

[0034] According to the above aspects and any possible implementation manner, an implementation manner is further provided, wherein S5 specifically includes:

[0035] S51: reading the training data of assembly parameters and assembly accuracy in the limited product samples, using the training data to obtain a support vector regression model that can predict the assembly accuracy, reading the test data of assembly parameters and assembly accuracy of the limited product samples, and verifying the accuracy of the model;

[0036] S52: using Sobol sequence sampling to obtain assembly parameter samples, and constructing a Sobol sampling matrix at the same time, extracting multiple sets of assembly parameter input data and substituting them into the established support vector regression model to obtain multiple sets of assembly parameter output data;

[0037] S53: Substitute the input data of multiple groups of assembly parameters and their corresponding output data into the Sobol global sensitivity calculation model to calculate and analyze the first-order sensitivity and the global sensitivity.

[0038] According to the above aspects and any possible implementation manner, an implementation manner is further provided, wherein S6 specifically includes:

[0039] S61: collecting the first-order sensitivity and global sensitivity values ​​of each assembly parameter, summing the global sensitivity value and the first-order sensitivity value of each assembly parameter, judging whether the sum is 1, and determining the validity of the sensitivity result calculation;

[0040] S62: sorting the global sensitivity values ​​of the various assembly error parameters from large to small, determining the key error elements that have a greater impact on the assembly accuracy, and at the same time determining the assembly parts that have a greater impact on the assembly accuracy according to the part classification.

[0041] According to the above aspects and any possible implementation manner, an implementation manner is further provided, wherein S7 specifically includes:

[0042] S71: Based on the engineering digital model provided by the design department, the model format in the engineering digital model is uniformly converted into the JT format using computer software, and the model is imported into the VSA tolerance simulation analysis software, the tolerance data in the design scheme is read, the tolerance of each feature of the part is defined, the assembly process scheme of the product is read, the assembly sequence of the parts, the clamping scheme and the matching relationship between the parts are determined, and the matching relationship between the parts is set in the software;

[0043] S72: Set the measurement operation of assembly accuracy in the VSA software, set the measurement of position changes of multiple points, and measure the position changes between points to complete the measurement of assembly accuracy, and rely on the software's own functions to complete the calculation of sensitivity values;

[0044] S73: Read the sensitivity calculation result in the VSA software and compare it with the relevant results obtained by Sobol calculation, analyze and calculate the deviation between the simulation value and the calculated value, and ensure the accuracy of the sensitivity analysis calculation method driven by mechanism-data fusion.

[0045] According to the above aspects and any possible implementation, a key assembly error element sensitivity analysis system is further provided, the analysis system is used for identifying key error elements of composite wing box assembly parts during the production process, the analysis system comprises:

[0046] The assembly error transmission mechanism model building module is used to build the assembly error transmission mechanism model under the action of multiple error sources according to the assembly process and structural characteristics of the product;

[0047] An assembly error transfer data prediction model building module is used to build an assembly error transfer data prediction model based on the assembly error transfer mechanism model and finite sample data;

[0048] A mechanism-data organic fusion model building module is used to build a mechanism-data organic fusion model through an assembly error transmission mechanism model and an assembly error transmission data prediction model;

[0049] The global sensitivity calculation model construction module is used to construct the global sensitivity calculation model for assembly accuracy data analysis based on the mechanism-data organic fusion model;

[0050] Global sensitivity calculation module, used to perform global sensitivity calculation through mechanism-data organic fusion model and Sobol global sensitivity calculation model;

[0051] Key error element identification module, used to identify assembly key error elements through global sensitivity calculation results;

[0052] The validity verification module is used to verify the validity of the assembly key error element identification results.

[0053] Compared with the prior art, the present invention can achieve the following technical effects:

[0054] 1) The present invention establishes a mechanism model and a data model of error transmission respectively, uses the calculation results of the two as proxy functions, and adopts the Sobol sensitivity analysis method to establish a global sensitivity calculation model for assembly accuracy data analysis to complete the identification of key assembly error elements;

[0055] 2) The present invention constructs a prediction model driven by the fusion of a mechanism model and a data model, which can effectively identify key assembly error elements only by relying on small sample data in the production process, and solves the problem that the data-driven model requires a large amount of data support;

[0056] 3) The present invention can avoid the shortcomings of existing commercial tolerance analysis software, such as low calculation accuracy and poor interpretability of assembly accuracy analysis results caused by not considering the actual physical property data of the assembly site during sensitivity simulation, and the calculation results are more adaptable to the assembly site.

[0057] Of course, any product implementing the present invention does not necessarily need to achieve all of the above-mentioned technical effects at the same time.

Brief Description of the Drawings

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0059] Figure 1 It is a wing box structure and key assembly features provided by an embodiment of the present invention;

[0060] Figure 2 It is a flow chart of a mechanism-data model fusion construction method provided by an embodiment of the present invention;

[0061] Figure 3 is a sensitivity analysis flow chart provided by an embodiment of the present invention;

[0062] Figure 4 This is a straight seam gap calculation process under the mechanism-data hybrid drive provided by an embodiment of the present invention;

[0063] Figure 5 is an error distribution diagram of a VSA simulation straight-to-seam gap provided by an embodiment of the present invention;

[0064] Figure 6 is the sensitivity value of each assembly parameter of the straight-to-seam gap simulated by the VSA provided by an embodiment of the present invention;

[0065] Figure 7is an error distribution diagram of a straight-to-slit step difference simulated by a VSA according to an embodiment of the present invention;

[0066] Figure 8 is the sensitivity value of each assembly parameter of the VSA simulation straight seam step difference provided by an embodiment of the present invention;

[0067] Fig. 9 It is a flow chart of a key assembly error element sensitivity analysis method driven by mechanism and data fusion proposed by the present invention. [Specific implementation method]

[0068] In order to better understand the technical solution of the present invention, the embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0069] It should be clear that the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "said" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0071] The present invention provides a key assembly error element sensitivity analysis method, which is used to identify key error elements in the production process of composite wing box assembly parts. The analysis method comprises the following steps:

[0072] S1: According to the assembly process and structural characteristics of the product, the assembly error transmission mechanism model under the action of multiple error sources is constructed;

[0073] S2: Based on the assembly error transmission mechanism model and finite sample data, an assembly error transmission data prediction model is constructed;

[0074] S3: Construct a mechanism-data organic fusion model through the assembly error transmission mechanism model and the assembly error transmission data prediction model;

[0075] S4: Construct a global sensitivity calculation model for assembly accuracy data analysis through a mechanism-data organic fusion model;

[0076] S5: Global sensitivity calculation is performed through the mechanism-data organic fusion model and the Sobol global sensitivity calculation model;

[0077] S6: Identify assembly key error elements through global sensitivity calculation results;

[0078] S7: Verify the effectiveness of the assembly key error element identification results.

[0079] The S1 specifically includes:

[0080] S11: According to the assembly process and structural characteristics of the product, determine the various error sources existing in the product assembly process, obtain relevant data of the error sources, and use a small displacement screw model to represent the changes of the corresponding characteristics in the assembly process;

[0081] S12: According to the assembly relationship between the components and the positioning and clamping conditions of the tooling, combined with the error sources analyzed in S11, an assembly error transmission relationship diagram at the key assembly link is established;

[0082] S13: According to the assembly error transfer relationship diagram and the small displacement screw model, the Jacobi screw model is used to construct the assembly error transfer mechanism model.

[0083] The S2 specifically includes:

[0084] S21: according to the structure of the assembly error mechanism model, determine the collection requirements of the assembly error transfer data prediction model, and define the data output, wherein the output data is the deviation between the assembly accuracy calculation value of the mechanism model and the measured assembly value;

[0085] S22: Determine data collection requirements based on data output, where the collection requirements include input data and training data in the model training and construction process;

[0086] S23: Limited data collection is performed on input data and training data, and the data is preprocessed to remove outliers. The assembly error transfer data prediction model is optimized through a support vector regression prediction model, and the numerical values ​​of the mechanism calculation assembly accuracy and the measured assembly accuracy are predicted through the optimized prediction model.

[0087] The limited data described in the present invention is small sample data, specifically the data needed to be used in the training model in the present invention, such as actual assembly accuracy data collected from on-site measurements, process data read in the assembly process plan, related simulation data and error data.

[0088] The S3 specifically includes:

[0089] S31: Obtain product data, substitute the product data into the assembly error transmission mechanism model, calculate the mechanism theory calculation value of the assembly accuracy, obtain the error transmission measured data, use the error transmission measured data and the mechanism theory calculation value as training samples, and the training prediction value is the assembly error transmission data prediction model of the assembly mechanism calculation deviation;

[0090] S32: obtaining a predicted value of a mechanism calculation deviation through an assembly error transfer data prediction model;

[0091] S33: adding the calculated assembly accuracy mechanism calculation deviation prediction value to the assembly accuracy mechanism calculation theoretical value to obtain a compensated assembly accuracy calculation value;

[0092] S34: A mechanism-data organic fusion model for assembly accuracy prediction is constructed through the compensated assembly accuracy calculation value.

[0093] The error transfer measured data include but are not limited to the assembly accuracy measured data collected on-site, the process data read in the assembly process plan, the relevant simulation data and the error data.

[0094] The S4 specifically includes:

[0095] S41: setting the value range of each assembly parameter and performing random sampling in the finite data, calculating the theoretical calculation accuracy of the mechanism model according to the sampling results, and establishing the first matrix and the second matrix according to the sampling data;

[0096] S42: exchanging the same columns in the first matrix and the second matrix to obtain a third matrix corresponding to the first matrix and a fourth matrix corresponding to the second matrix;

[0097] S43: inputting the error source parameters in the first matrix, the second matrix, the third matrix and the fourth matrix into a mechanism-data organic fusion model, and obtaining four groups of assembly accuracy data sets;

[0098] S44: The first-order sensitivity and global sensitivity of four sets of assembly parameters are calculated by using four sets of assembly accuracy data sets;

[0099] S45: The Sobol global sensitivity calculation model is constructed through four sets of assembly accuracy data sets, first-order sensitivity and global sensitivity.

[0100] The S5 specifically includes:

[0101] S51: reading the training data of assembly parameters and assembly accuracy in the limited product samples, using the training data to obtain a support vector regression model that can predict the assembly accuracy, reading the test data of assembly parameters and assembly accuracy of the limited product samples, and verifying the accuracy of the model;

[0102] S52: using Sobol sequence sampling to obtain assembly parameter samples, and constructing a Sobol sampling matrix at the same time, extracting multiple sets of assembly parameter input data and substituting them into the established support vector regression model to obtain multiple sets of assembly parameter output data;

[0103] S53: Substitute the input data of multiple groups of assembly parameters and their corresponding output data into the Sobol global sensitivity calculation model to calculate and analyze the first-order sensitivity and the global sensitivity.

[0104] The S6 specifically includes:

[0105] S61: collecting the first-order sensitivity and global sensitivity values ​​of each assembly parameter, summing the global sensitivity value and the first-order sensitivity value of each assembly parameter, judging whether the sum is 1, and determining the validity of the sensitivity result calculation;

[0106] S62: sorting the global sensitivity values ​​of the various assembly error parameters from large to small, determining the key error elements that have a greater impact on the assembly accuracy, and at the same time determining the assembly parts that have a greater impact on the assembly accuracy according to the part classification.

[0107] The S7 specifically includes:

[0108] S71: Based on the engineering digital model provided by the design department, the model format in the engineering digital model is uniformly converted into the JT format using computer software, and the model is imported into the VSA tolerance simulation analysis software, the tolerance data in the design scheme is read, the tolerance of each feature of the part is defined, the assembly process scheme of the product is read, the assembly sequence of the parts, the clamping scheme and the matching relationship between the parts are determined, and the matching relationship between the parts is set in the software;

[0109] In a specific embodiment, the computer software may preferably be converted into a format using CATIA software, or converted into a model format provided in VSA software.

[0110] S72: Set the measurement operation of assembly accuracy in the VSA software, set the measurement of position changes of multiple points, and measure the position changes between points to complete the measurement of assembly accuracy, and rely on the software's own functions to complete the calculation of sensitivity values;

[0111] S73: Read the sensitivity calculation result in the VSA software and compare it with the relevant results obtained by Sobol calculation, analyze and calculate the deviation between the simulation value and the calculated value, and ensure the accuracy of the sensitivity analysis calculation method driven by mechanism-data fusion.

[0112] The present invention also provides a key assembly error element sensitivity analysis system, which is used to identify key error elements in the production process of composite wing box assembly parts. The analysis system includes:

[0113] The assembly error transmission mechanism model building module is used to build the assembly error transmission mechanism model under the action of multiple error sources according to the assembly process and structural characteristics of the product;

[0114] An assembly error transfer data prediction model building module is used to build an assembly error transfer data prediction model based on the assembly error transfer mechanism model and finite sample data;

[0115] A mechanism-data organic fusion model building module is used to build a mechanism-data organic fusion model through an assembly error transmission mechanism model and an assembly error transmission data prediction model;

[0116] The global sensitivity calculation model construction module is used to construct the global sensitivity calculation model for assembly accuracy data analysis based on the mechanism-data organic fusion model;

[0117] Global sensitivity calculation module, used to perform global sensitivity calculation through mechanism-data organic fusion model and Sobol global sensitivity calculation model;

[0118] Key error element identification module, used to identify assembly key error elements through global sensitivity calculation results;

[0119] The validity verification module is used to verify the validity of the assembly key error element identification results.

[0120] The principle of the present invention is as follows:

[0121] like Fig. 9 As shown, the present invention provides a key assembly error element sensitivity analysis method driven by mechanism and data fusion, the assembly accuracy method is used to identify key error elements of composite wing box assembly parts during the production process, and the key error element identification includes the following steps:

[0122] S1: Construction of assembly error transmission mechanism model under the action of multiple error sources;

[0123] S2: Construction of assembly error transfer data model supported by small sample data;

[0124] S3: Construction of mechanism-data organic fusion model;

[0125] S4: Construction of Sobol global sensitivity calculation model for assembly accuracy data analysis;

[0126] S5: Mechanism-data fusion driven global sensitivity calculation;

[0127] S6: Identification of critical error elements of assembly using sensitivity numerical analysis;

[0128] S7: Verification of the effectiveness of the sensitivity analysis method for key assembly error elements.

[0129] The S1 specifically includes:

[0130] S11: According to the assembly process and structural characteristics of the product, find out the various error sources in the product assembly process, and use the small displacement screw model to represent the changes of the corresponding characteristics in the assembly process.

[0131] S12: According to the assembly relationship between the components and the tooling positioning conditions, combined with the key error sources analyzed in S11, an assembly error transfer relationship diagram at the key assembly link is established.

[0132] S13: According to the assembly error transmission relationship diagram in the product assembly process and the small displacement screw model of each key feature, the Jacobi screw model is used to construct the assembly error transmission mechanism model.

[0133] The S2 specifically includes:

[0134] S21: First, according to the structure of the assembly error mechanism model, the data model collection requirements are clarified and the data output is defined. The output data is the deviation between the assembly accuracy calculation value of the mechanism model and the measured assembly value.

[0135] S22: According to the data output of the data model, clarify the data collection requirements, which mainly include input data and training data in the model training and construction process.

[0136] S23: Collect input data and training data required for data model construction, preprocess the data to remove outliers, and then predict the numerical values ​​of mechanism calculation assembly accuracy and measured assembly accuracy through support vector regression prediction model.

[0137] The S3 specifically includes:

[0138] S31: Set the value range of each assembly parameter and perform random sampling. According to the sampling results, the theoretical calculation accuracy of the mechanism model is calculated at the same time. Matrix A (first matrix) and matrix B (second matrix) are established based on the sampling data (2m sampling samples, k assembly parameters, m and k are both natural numbers).

[0139] S32: Exchange the i-th column (equivalent to the same column mentioned above) of matrix A (first matrix) and matrix B (second matrix) to obtain matrix S i (The third matrix) and the matrix S -i (Fourth Matrix).

[0140] S33: transform the matrices A, B, S i ,S -i The assembly parameter data such as tooling positioning accuracy, component feature manufacturing error, etc. in the first matrix, the second matrix, the third matrix, and the fourth matrix are brought into the established mechanism-data fusion assembly accuracy prediction model to obtain the corresponding assembly accuracy data set y A,y B , (i.e., four sets of assembly accuracy data sets), and then further calculate the assembly parameters x i The first-order sensitivity With global sensitivity

[0141] The S4 specifically includes:

[0142] S41: Measure and collect the actual data such as assembly accuracy generated on site, and read the process data, related simulation data and error data in the assembly process plan, and use these data as training samples. The training prediction value is a support vector regression model of the assembly mechanism calculation deviation.

[0143] S42: Substitute the product-related data into the assembly error mechanism calculation model to calculate the theoretical calculation value of the assembly accuracy, and use the calculation result as the input of the obtained training model to obtain the predicted value of the mechanism calculation deviation through the data model.

[0144] S43: Adding the calculated assembly accuracy mechanism calculation deviation prediction value to the assembly accuracy mechanism calculation theoretical value to obtain a compensated assembly accuracy calculation value.

[0145] The S5 specifically includes:

[0146] S51: Read the training data of assembly parameters and assembly accuracy of product samples, and use these data to train a support vector regression model that can predict assembly accuracy. Read the test data of assembly parameters and assembly accuracy of product samples to verify the accuracy of the model.

[0147] S52: Use Sobol sequence sampling to obtain assembly parameter samples and construct the Sobol sampling matrix AB i , multiple sets of assembly parameter input data are extracted and substituted into the established support vector regression model to further obtain multiple sets of assembly parameter output data.

[0148] S53: Substitute the input data of multiple groups of assembly parameters and their corresponding output data into the calculation formulas of the first-order sensitivity and the global sensitivity respectively to complete the sensitivity analysis of the mechanism-data fusion.

[0149] The S6 specifically includes:

[0150] S61: Collect the first-order sensitivity and global sensitivity values ​​of each assembly parameter, sum the global sensitivity value and the first-order sensitivity value of each assembly parameter, determine whether the sum is 1, and determine the validity of the sensitivity result calculation.

[0151] S62: sorting the global sensitivity values ​​of the various assembly error parameters from large to small, determining several key error elements that have a greater impact on the assembly accuracy, and at the same time determining several assembly parts that have a greater impact on the assembly accuracy according to the part classification.

[0152] The S7 specifically includes:

[0153] S71: Based on the engineering digital model provided by the design department, CATIA software is used to convert the model format into JT format, and the model is imported into VSA tolerance simulation analysis software, the tolerance data in the design scheme is read, and the tolerance of each feature of the part is defined. The assembly process scheme of the product is read, the assembly sequence of the parts, the clamping scheme and the matching relationship between the parts are determined, and the matching relationship between the parts is set in the software.

[0154] S72: Set up the measurement operation of assembly accuracy in the VSA software, set up the measurement of position changes of multiple points, and measure the position changes between points to complete the measurement of assembly accuracy. At the same time, rely on the software's own functions to complete the calculation of sensitivity values.

[0155] S73: Read the sensitivity calculation result in the VSA software and compare it with the relevant results obtained by Sobol calculation, analyze and calculate the deviation between the simulation value and the calculated value, and ensure the accuracy of the sensitivity analysis calculation method driven by mechanism-data fusion.

[0156] Embodiment 1:

[0157] In this embodiment, the key assembly error element identification of the straight seam gap and step difference of a certain type of wing box assembly is taken as an example. Figure 1 As shown in the figure, the wing box is composed of machined ribs, front and rear beams, frame plates, partitions and composite skins. The assembly sequence is from outside to inside, and the wing box frame and composite wall panels are assembled in sequence. At the same time, during the assembly process, it is necessary to control the assembly accuracy of the left straight seam gap and step difference to ensure the assembly quality. Taking the key error link identification of the left straight seam gap and step difference of this product as an example, the implementation steps of the key assembly error element sensitivity analysis method driven by mechanism and data fusion are explained.

[0158] The overall idea adopted by the present invention to solve the technical problem is as follows: first, according to the assembly process scheme of the product, multiple existing error sources are read, including the tolerance of the assembly contact feature surface of each part, the positioning accuracy of the assembly tooling, the part position tolerance information, the measured assembly accuracy value, and the mechanism model assembly accuracy calculation value, and the assembly error transmission mechanism model and data model are respectively established, so as to establish a Sobol global sensitivity calculation model for assembly accuracy, so as to perform sensitivity analysis on each assembly error link of the product; secondly, a fusion modeling method of the error transmission mechanism model and the measured / simulated data model is proposed, and the deviation between the mechanism model assembly accuracy calculation value and the measured assembly value is used as the predicted output data, and a support vector regression method is used to construct a prediction model of the assembly accuracy mechanism calculation deviation value driven by mechanism-data fusion; finally, relying on the established mechanism-data fusion driven assembly error prediction model, the Sobol sensitivity analysis method is used to calculate the sensitivity value of each assembly error link, and the calculation results are sorted by size, and the key assembly error elements are analyzed and identified, and the assembly process is simulated by VSA software, and the validity of the sensitivity value is verified by comparing the simulation and calculation results. The above method can effectively guide the subsequent tolerance optimization process and provide an effective solution to the problems of difficulty in identifying key assembly errors and low efficiency.

[0159] The present invention provides a key assembly error element sensitivity analysis method driven by mechanism and data fusion, specifically a key assembly error element sensitivity analysis method considering the actual physical characteristics of the assembly site and assembly accuracy data, which can be used for identifying and improving the key assembly error links of mechanical products. The key assembly error element sensitivity analysis method driven by mechanism and data fusion comprises the following steps:

[0160] S1: Construction of assembly error transmission mechanism model under the action of multiple error sources;

[0161] Specifically, firstly according to Figure 1 The overall structure and key assembly features of the wing box show that the wing box is composed of machined ribs, front and rear beams, frame plates, partitions and composite skins, and the assembly sequence is from the outside to the inside. It can be analyzed that the error sources in the assembly process of the product mainly include component manufacturing errors, deformation errors caused by contact between parts, and assembly tooling positioning and clamping errors.

[0162] Further, a global coordinate system of the assembly system is created, and then a local coordinate system is created on the mating surface of each part. According to the changes caused by the above-mentioned various errors, small displacement spinors are used to represent them, and the spinor matrix form of the changes of various error sources can be obtained. The geometric function elements of the part errors of the present invention are shown in Table 1.

[0163] Table 1

[0164]

[0165]

[0166] According to the assembly sequence of the wing box, the matching relationship of the characteristic surfaces of each component and the tooling positioning method, an assembly error transfer relationship diagram is constructed.

[0167] Furthermore, the position error conversion matrix can be obtained through the assembly error transfer relationship diagram obtained above and the spinor matrix form of the changes of various error sources, which is used to describe the transfer relationship between the positioning tooling, the inside of the part and the mating surface in the product assembly process;

[0168] The effects of the above-mentioned parts' own manufacturing errors, assembly deformation errors, assembly tooling positioning and clamping errors on product assembly are converted into changes in key assembly geometric features. The values ​​in the small displacement screw model are updated, and the Jacobian screw matrix representing the error transfer relationship is corrected to establish a mechanism model that characterizes the product assembly error transfer process.

[0169] S2: Construction of assembly error transfer data model supported by small sample data;

[0170] Specifically, according to the assembly function requirements of the wing box overall structure and key assembly features, the output data of the wing assembly data model is determined to be the deviation between the assembly accuracy calculation value of the mechanism model and the measured assembly value.

[0171] Furthermore, according to the source of data, it is mainly divided into input data and training data in the process of model training and construction, where the input data includes the error source information of the contact and matching surfaces of each assembly, including the tolerance of the contact feature surface of each part assembly, the positioning accuracy of the assembly tooling, the part position tolerance information, the measured assembly accuracy value, the calculated value of the mechanism model assembly accuracy, etc. The training data is a small sample set of tolerance data of each component link.

[0172] Furthermore, the numerical values ​​of the mechanism calculation assembly accuracy and the measured assembly accuracy are predicted by the support vector regression prediction model, namely:

[0173]

[0174] In the formula, f(x1,…x n ) is the compensation value of the calculation result of the mechanism model, y m is the measured product assembly accuracy, y t (x1,…x n ) is the theoretical assembly deviation calculated by the mechanism model, is the feature variation error of part feature 1, is the feature variation error of part feature 1. Part feature 1 and part feature 2 together constitute the assembly accuracy error.

[0175] S3: Construction of mechanism-data organic fusion model;

[0176] Specifically, the value range of each assembly parameter is set first, and the Sobol Sequence sampling method is used to randomly sample the assembly input parameters. The results are then substituted into the mechanism model to obtain a sample set of theoretical calculation results of assembly accuracy.

[0177] Furthermore, according to the obtained sample sets of assembly parameters and theoretically calculated values ​​of assembly accuracy, matrices A and B (2m sampling samples, k assembly parameters) are established:

[0178]

[0179]

[0180] Furthermore, the matrix A is swapped with the i-th column of the matrix B to obtain the matrix S i With S -i :

[0181]

[0182]

[0183] Furthermore, the matrices A, B, S i ,S -i The assembly parameter data such as tooling positioning accuracy and component feature manufacturing error in the above mechanism-data fusion assembly accuracy prediction model are brought into the corresponding assembly accuracy data set y A ,y B , have:

[0184]

[0185] Further, the assembly parameter x is calculated i The first-order sensitivity With global sensitivity Complete the construction of Sobol global sensitivity calculation model:

[0186]

[0187]

[0188] S4: Construction of Sobol global sensitivity calculation model for assembly accuracy data analysis;

[0189] Specifically, the actual manufacturing accuracy and deformation deviation values ​​of the product ribs, beams, walls, partitions, upper and lower wall panels are measured and collected first. After the wing box product is assembled, the assembly gap and assembly step difference accuracy data of the left straight seam are measured, and the above-collected data sets are used as training samples to train the constructed assembly mechanism calculation deviation prediction model based on support vector regression. The main process is as follows: Figure 2 shown.

[0190] Furthermore, the design tolerance information of the component parts (ribs, beams, walls, partitions, upper and lower wall panels, etc.) of the wing box product is input into the established assembly error mechanism calculation model, and the straight seam gap error and step difference error of the wing box product are calculated respectively.

[0191] Furthermore, the measured accuracy data of the straight seam gap and step difference generated during the wing box assembly process, the assembly process data, the simulation data and the error theory calculation data are read, and substituted into the established support vector regression prediction mathematical model to calculate the mechanism calculation deviation prediction values ​​of the product straight seam gap and step difference respectively.

[0192] Furthermore, the calculated deviation prediction value of the mechanism calculation of the straight seam gap and step difference of the product is added to the corresponding gap and step difference mechanism calculation value to complete the correction of the theoretical calculation value of the mechanism model, and the compensated straight seam gap of the wing box and the compensated assembly accuracy calculation value are obtained, completing the organic integration of the error transmission mechanism model and the simulation / measured data model.

[0193] S5: Numerical calculation of global sensitivity driven by mechanism-data model fusion;

[0194] The sensitivity analysis calculation process under data-mechanism hybrid drive is as follows Figure 3 As shown, specifically, first read the small sample data of the product production process (including the shape and position errors of the product ribs, beams, walls, partitions, upper and lower wall panels, the positioning errors of each tooling, the straight seam gap and step difference accuracy of the product, etc.), and use 70% of the data samples to build the support vector regression model, and 30% to verify the accuracy.

[0195] Furthermore, the model input data is defined as multiple error sources in the assembly process, and the output data is the compensated assembly accuracy data. The model is trained. After the training is completed, the remaining 30% of the sample data is used to verify the model accuracy and calculate the mean absolute error MAE, root mean square error RMSE, and absolute coefficient R 2 They are 0.052176, 0.007424, and 0.079945 respectively, and the accuracy meets the requirements.

[0196] Furthermore, the value range of each assembly parameter is set, and the Sobol sequence method is used to randomly sample the geometric tolerances of the left and right walls, frames and other components, construct the sampling matrices A and B of each error link of the product, and exchange the i-th column of matrix A with matrix B to obtain matrix S i With S -i .

[0197] Furthermore, the calculation process of the straight seam gap is as follows: Figure 4 As shown, the matrices A, B, S i ,S -i The assembly parameter data such as tooling positioning accuracy and component feature manufacturing error in the above mechanism-data fusion assembly accuracy prediction model are brought into the corresponding assembly accuracy data set y A ,y B , The calculated results are substituted into the global sensitivity calculation formula to obtain the global sensitivity analysis values ​​of the gap and step difference between the straight seams between the skins.

[0198] S6: Identification of critical error elements of assembly using sensitivity numerical analysis;

[0199] Specifically, the sensitivity analysis results of the straight seam gap and the straight seam step difference between the skins are read respectively, as shown in Tables 2 and 3.

[0200] Table 2

[0201]

[0202]

[0203] Table 3

[0204]

[0205] Furthermore, the sensitivity results in the above table are sorted, and the global sensitivity of the gap between the straight seams of the No. 1 and No. 2 wall panels is as follows from large to small:

[0206] x6>x5>x4>x3>x2>x1>x 23 >x 18 >x 24 >x 17 >x 16 >x 13 >x 15 >x 14 >x 20 >x 10 >x8

[0207] >x 19 >x 22 >x 21 >x7>x9>x11 >x 12

[0208] The global sensitivity of the step difference between the straight seams of the 1st and 2nd wall panels is as follows from large to small:

[0209] x4>x3>x 15 >x 16 >x 14 >x 24 >x 23 >x 12 >x 11 >x 13 >x1>x2>x 22 >x 21 >x 17 >x8>x6

[0210] >x 10 >x9>x 19 >x7>x5>x 18 >x 20

[0211] Further analysis of the sensitivity ranking results showed that the main factors affecting the assembly accuracy of the straight seam gap are the manufacturing error of the upper wall panel edge, the manufacturing error of the upper wall panel positioning hole, and the positioning error of the upper wall panel assembly tooling; the main factors affecting the step difference accuracy of the straight seam are the manufacturing error in the wall panel thickness direction, the adsorption positioning error of the upper wall panel assembly tooling, the height tolerance and contour of each frame and the left and right walls. The sensitivity values ​​of the remaining assembly error elements to the outer shape gap and step difference are relatively small.

[0212] S7: Verification of the effectiveness of the sensitivity analysis method for key assembly error elements.

[0213] Specifically, first convert the digital models of the skin, frame, bulkhead and other parts in the product into JT format, create an assembly environment in the VSA software and import the assembly parts. At the same time, read the shape and position tolerance data of the ribs, beams, walls, bulkheads, composite wall panels and other parts, and complete the definition of each tolerance information of the product by adding features in the simulation environment. The tolerance information is shown in Table 4 below. For other parts that do not explicitly require tolerances, the initial tolerance input is performed according to the HB5800 standard.

[0214] Table 4

[0215]

[0216]

[0217] Furthermore, the parts are assembled according to the assembly process plan, and the assembly operations are established in the order of left and right walls, 1st frame + 4th frame, 2nd frame + 3rd frame, front section of middle wall, rear section of middle wall, middle section of middle wall, upper wall panel, and lower wall panel.

[0218] Furthermore, two measurement points are set on both sides of the straight seam gap formed by the upper wall panel of the product as measurement features. Then, the deviation measurement operation of the two measurement points in the X direction and the Y direction is established to achieve the measurement of the straight seam gap and the straight seam step difference accuracy. Then, the tolerance analysis simulation is performed to obtain the simulation distribution results and sensitivity values ​​of the straight seam gap and the straight seam step difference, respectively. The calculation results are shown in Table 5 and Table 6, respectively. The error distribution and sensitivity calculation results of the straight seam gap are shown in Table 5 and Table 6. Figure 5 and Figure 6 As shown in the figure, the error distribution and sensitivity calculation results of the straight seam step difference are as follows Figure 7 and Figure 8 shown.

[0219] Table 5 Table 6

[0220]

[0221] Furthermore, the sensitivity analysis results in the VSA software are recorded and compared with the sensitivity analysis results obtained by the mechanism-data fusion driven calculation. The sensitivity calculation values ​​of several key error elements that contribute most to the assembly accuracy sensitivity are listed, and the deviation of the two calculation methods is calculated to ensure the effectiveness of the sensitivity calculation method.

[0222] (1) The factors that have a greater impact on the assembly accuracy of the straight seam gap between the skins are mainly the tooling positioning error of the wall panel parts, the tooling positioning error of the frame parts, and the manufacturing error of the wall panel parts, which together account for 71.52%, and the remaining error elements account for 28.48% (manufacturing error of the frame parts composed of the frame and wall); the deviation between the VSA software simulation results and the calculation results of the present invention is 5.53%.

[0223] (2) The factors that have a greater impact on the assembly accuracy of the straight seam step difference between the skins are mainly the tooling positioning errors of each part and the manufacturing errors of the skeleton parts, which together account for 91.55%, and the remaining error elements (thickness errors of individual wall panel parts) account for 8.45%; the deviation between the VSA software simulation results and the calculation results of the present invention is 7.05%.

[0224] The present invention provides a key assembly error element sensitivity analysis method driven by mechanism and data fusion, which provides an effective solution to the problems of failure to consider the actual physical property data on site and insufficient data utilization in the assembly simulation modeling stage, which makes it difficult to identify key error elements.

[0225] The present invention mainly includes two parts, namely: the construction of a mechanism calculation deviation compensation model based on support vector regression and the sensitivity analysis solution under data-mechanism hybrid drive. After example optimization verification, the present invention can achieve:

[0226] 1) The mechanism model and data model of error transmission are established respectively, and the calculation results of the two are used as proxy functions. The Sobol sensitivity analysis method is adopted to establish a global sensitivity calculation model for assembly accuracy data analysis to complete the identification of key assembly error elements.

[0227] 2) Construct a prediction model driven by the fusion of mechanism model and data model, which can effectively identify key assembly error elements by relying only on small sample data, mechanism model and measured data in the production process, and solve the problem that data-driven models require a large amount of data support.

[0228] 3) Compared with the VSA tolerance simulation analysis software that calculates the sensitivity values ​​of each error link, the method of the present invention can fully consider the influence of factors such as deformation deviation of thin-walled parts caused by the positioning and clamping of assembly tooling and contact deformation of parts on the assembly error transmission process, and can avoid the shortcomings of VSA software analysis such as low sensitivity calculation accuracy and poor interpretability of assembly accuracy analysis results caused by not considering the actual physical property data of the assembly site, and has better adaptability to the assembly site.

[0229] The above is a detailed introduction to a key assembly error element sensitivity analysis method and system provided by the embodiment of the present application. The description of the above embodiment is only used to help understand the method and its core idea of ​​the present application; at the same time, for those skilled in the art, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

[0230] For example, certain words are used in the specification and claims to refer to specific components. Those skilled in the art should understand that hardware manufacturers may use different nouns to refer to the same component. This specification and claims do not use differences in names as a way to distinguish components, but use differences in the functions of components as the criteria for distinction. As mentioned throughout the specification and claims, "including" and "comprising" are open-ended terms, so they should be interpreted as "including / including but not limited to". "Approximately" means that within an acceptable error range, those skilled in the art can solve the technical problem within a certain error range and basically achieve the technical effect. The subsequent description of the specification is a preferred embodiment of the present application, but the description is for the purpose of illustrating the general principles of the present application, and is not used to limit the scope of the present application. The scope of protection of the present application shall be determined by the definition of the attached claims.

[0231] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a product or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such a product or system. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the product or system including the elements.

[0232] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0233] The above description shows and describes several preferred embodiments of the present application, but as mentioned above, it should be understood that the present application is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the application concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art do not deviate from the spirit and scope of the present application, and should be within the scope of protection of the claims attached to the present application.

Claims

1. A method for sensitivity analysis of key assembly error elements, which is used to identify key error elements in the production process of composite wing box assembly parts, characterized in that: The analytical method comprises the following steps: S1: According to the assembly process and structural characteristics of the product, the assembly error transmission mechanism model under the action of multiple error sources is constructed; S2: Based on the assembly error transmission mechanism model and finite sample data, an assembly error transmission data prediction model is constructed; S3: Construct a mechanism-data organic fusion model through the assembly error transmission mechanism model and the assembly error transmission data prediction model; S4: Construct a global sensitivity calculation model for assembly accuracy data analysis through a mechanism-data organic fusion model; S5: Global sensitivity calculation is performed through the mechanism-data organic fusion model and the Sobol global sensitivity calculation model; S6: Identify assembly key error elements through global sensitivity calculation results; S7: Verify the effectiveness of the assembly key error element identification results; The S1 specifically includes: S11: According to the assembly process and structural characteristics of the product, determine the various error sources existing in the product assembly process, obtain relevant data of the error sources, and use a small displacement screw model to represent the changes of the corresponding characteristics in the assembly process; S12: According to the assembly relationship between the components and the positioning and clamping conditions of the tooling, combined with the error sources analyzed in S11, an assembly error transmission relationship diagram at the key assembly link is established; S13: According to the assembly error transmission relationship diagram and the small displacement screw model, the Jacobi screw model is used to construct the assembly error transmission mechanism model; The S2 specifically includes: S21: according to the structure of the assembly error mechanism model, determine the collection requirements of the assembly error transfer data prediction model, and define the data output, wherein the output data is the deviation between the assembly accuracy calculation value of the mechanism model and the measured assembly value; S22: Determine data collection requirements based on data output, where the collection requirements include input data and training data in the model training and construction process; S23: performing limited data collection on input data and training data, and preprocessing the data to remove outliers, optimizing the assembly error transfer data prediction model through a support vector regression prediction model, and predicting the numerical values ​​of the mechanism calculation assembly accuracy and the measured assembly accuracy through the optimized prediction model; The S3 specifically includes: S31: Obtain product data, substitute the product data into the assembly error transmission mechanism model, calculate the mechanism theory calculation value of the assembly accuracy, obtain the error transmission measured data, use the error transmission measured data and the mechanism theory calculation value as training samples, and the training prediction value is the assembly error transmission data prediction model of the assembly mechanism calculation deviation; S32: obtaining a predicted value of a mechanism calculation deviation through an assembly error transfer data prediction model; S33: adding the calculated assembly accuracy mechanism calculation deviation prediction value to the assembly accuracy mechanism calculation theoretical value to obtain a compensated assembly accuracy calculation value; S34: constructing a mechanism-data organic fusion model for assembly accuracy prediction through the compensated assembly accuracy calculation value; The S4 specifically includes: S41: setting the value range of each assembly parameter and performing random sampling in the finite data, calculating the theoretical calculation accuracy of the mechanism model according to the sampling results, and establishing the first matrix and the second matrix according to the sampling data; S42: exchanging the same columns in the first matrix and the second matrix to obtain a third matrix corresponding to the first matrix and a fourth matrix corresponding to the second matrix; S43: inputting the error source parameters in the first matrix, the second matrix, the third matrix and the fourth matrix into a mechanism-data organic fusion model, and obtaining four groups of assembly accuracy data sets; S44: The first-order sensitivity and global sensitivity of four sets of assembly parameters are calculated by using four sets of assembly accuracy data sets; S45: The Sobol global sensitivity calculation model is constructed through four sets of assembly accuracy data sets, first-order sensitivity and global sensitivity.

2. The analysis method according to claim 1, characterized in that The error transfer measured data includes assembly accuracy measured data collected on-site, process data read from the assembly process plan, related simulation data and error data.

3. The analysis method according to claim 2, characterized in that The S5 specifically includes: S51: reading the training data of assembly parameters and assembly accuracy in the limited product samples, using the training data to obtain a support vector regression model that can predict the assembly accuracy, reading the test data of assembly parameters and assembly accuracy of the limited product samples, and verifying the accuracy of the model; S52: using Sobol sequence sampling to obtain assembly parameter samples, and constructing a Sobol sampling matrix at the same time, extracting multiple sets of assembly parameter input data and substituting them into the established support vector regression model to obtain multiple sets of assembly parameter output data; S53: Substitute the input data of multiple groups of assembly parameters and their corresponding output data into the Sobol global sensitivity calculation model to calculate and analyze the first-order sensitivity and the global sensitivity.

4. The analysis method according to claim 3, characterized in that The S6 specifically includes: S61: collecting the first-order sensitivity and global sensitivity values ​​of each assembly parameter, summing the global sensitivity value and the first-order sensitivity value of each assembly parameter, judging whether the sum is 1, and determining the validity of the sensitivity result calculation; S62: sorting the global sensitivity values ​​of the various assembly error parameters from large to small, determining the key error elements that have a greater impact on the assembly accuracy, and at the same time determining the assembly parts that have a greater impact on the assembly accuracy according to the part classification.

5. The analysis method according to claim 1, characterized in that The S7 specifically includes: S71: Based on the engineering digital model provided by the design department, the model format in the engineering digital model is uniformly converted into the JT format using computer software, and the model is imported into the VSA tolerance simulation analysis software, the tolerance data in the design scheme is read, the tolerance of each feature of the part is defined, the assembly process scheme of the product is read, the assembly sequence of the parts, the clamping scheme and the matching relationship between the parts are determined, and the matching relationship between the parts is set in the software; S72: Set the measurement operation of assembly accuracy in the VSA software, set the measurement of position changes of multiple points, and measure the position changes between points to complete the measurement of assembly accuracy, and rely on the software's own functions to complete the calculation of sensitivity values; S73: Read the sensitivity calculation result in the VSA software and compare it with the relevant results obtained by Sobol calculation, analyze and calculate the deviation between the simulation value and the calculated value, and ensure the accuracy of the sensitivity analysis calculation method driven by mechanism-data fusion.

6. A critical assembly error element sensitivity analysis system, which is used to identify critical error elements in the production process of composite wing box assembly parts. The analysis system is implemented based on the analysis method described in any one of claims 1 to 5, and is characterized in that: The analysis system comprises: The assembly error transmission mechanism model building module is used to build the assembly error transmission mechanism model under the action of multiple error sources according to the assembly process and structural characteristics of the product; An assembly error transfer data prediction model building module is used to build an assembly error transfer data prediction model based on the assembly error transfer mechanism model and finite sample data; A mechanism-data organic fusion model building module is used to build a mechanism-data organic fusion model through an assembly error transmission mechanism model and an assembly error transmission data prediction model; The global sensitivity calculation model construction module is used to construct the global sensitivity calculation model for assembly accuracy data analysis based on the mechanism-data organic fusion model; Global sensitivity calculation module, used to perform global sensitivity calculation through mechanism-data organic fusion model and Sobol global sensitivity calculation model; Key error element identification module, used to identify assembly key error elements through global sensitivity calculation results; The validity verification module is used to verify the validity of the assembly key error element identification results.

Images