Aluminum profile machining quality detection method and system

Abstract

The application discloses an aluminum profile machining quality detection method and system, comprising: acquiring multi-modal defect detection data and user quality acceptance requirements; according to the quality acceptance requirements, performing inversion analysis on the multi-modal defect detection data; according to the analysis result, judging whether the quality acceptance requirements are met, if yes, judging that the quality detection result is qualified; otherwise, judging that the quality detection result is unqualified. Not only suitable for intelligent screening in batch detection, but also providing basis for quality prediction and process feedback in complex use scenarios, having the advantages of high detection precision, strong adaptability, high intelligent degree and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of quality detection, in particular to an aluminum profile processing quality detection method and system. BACKGROUND

[0002] Aluminum profiles are widely used in the fields of construction, transportation, electronics, aerospace, etc. due to their light weight, high strength, good corrosion resistance and other characteristics. With the increasing demand for application performance of aluminum profiles, quality control during the forming process is particularly important. Aluminum profiles are susceptible to process parameters, material defects, uneven cooling speed and other factors during processing, resulting in defects such as cracks, pores, warping and uneven wall thickness, which can seriously affect the structural performance and service life. Traditional quality detection methods mostly rely on manual sampling or surface visual detection, which cannot fully evaluate internal defects and their impact on mechanical properties, and have problems such as detection blind spots, low efficiency, and inconsistent judgment standards.

[0003] With the development of non-destructive testing technology, multi-modal detection methods such as ultrasonic waves, eddy current detection, optical scanning, and X-ray imaging have been gradually applied in aluminum profile defect identification. Although these technologies can achieve higher precision data acquisition, they still face two challenges in practical engineering applications: first, the detection data is diverse and high-dimensional, and there is a lack of unified data fusion and interpretation mechanism; second, there is a clear gap between the detection results and the actual load-bearing capacity of the material, which cannot effectively support quality judgment based on "use performance". In addition, in the face of multiple use requirements under different working conditions, how to combine detection data with structural mechanical properties to achieve a feed-forward evaluation of quality risks is a problem that needs to be solved. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the technical problem solved by the present application is that the existing aluminum profile processing quality detection method cannot meet the demand for performance detection during use.

[0006] To solve the above technical problems, the present application provides the following technical scheme: an aluminum profile processing quality detection method, comprising:

[0007] Obtaining multi-modal defect detection data and user quality acceptance requirements;

[0008] Performing inverse analysis on the multi-modal defect detection data according to the quality acceptance requirements;

[0009] According to the analysis result, it is judged whether it meets the quality acceptance requirements, if it meets, the quality detection result is judged as qualified, otherwise, the quality detection result is judged as unqualified.

[0010] As a preferred scheme of the aluminum profile machining quality detection method, the multi-modal defect detection data comprises: for each quality defect type of the aluminum profile, using a corresponding detection means to perform defect detection to obtain detection data under each detection means.

[0011] As a preferred scheme of the aluminum profile machining quality detection method, the user quality acceptance requirement comprises: a minimum index of each quality defect detection result and a use requirement of the user for the aluminum profile.

[0012] The quality defect detection result comprises: using the detection data under each detection means to identify the quality defects of the aluminum profile and generate quantifiable parameters of each defect.

[0013] The minimum index is a limit of the quantifiable parameter in a single quality defect detection result of the aluminum profile. If the detection result of any quality defect of the aluminum profile does not meet the limit of the minimum index, the inversion analysis stage is skipped and the quality detection result is directly output as qualified. If the detection result of each quality defect of the aluminum profile meets the limit of the minimum index, inversion analysis is performed according to the use requirement of the user for the aluminum profile and the detection data under each detection means.

[0014] As a preferred scheme of the aluminum profile machining quality detection method, the use requirement of the user for the aluminum profile comprises: a requirement of the user for the aluminum profile to be normally used under a specific use condition.

[0015] The user provides a limit condition under each working condition when setting the use requirement.

[0016] The limit condition comprises: a maximum value of a working condition parameter, which is composed of a maximum pressure and a minimum stress area.

[0017] As a preferred scheme of the aluminum profile machining quality detection method, the inversion analysis of the multi-modal defect detection data comprises: making an assumption that any position of the aluminum profile can be used as an action point under the limit condition.

[0018] Under the assumption, the stress possibly generated at each point is simulated to obtain a limit stress at each point.

[0019] Each detection data is separately inversion analyzed at each point of the aluminum profile to obtain a stress threshold at each point.

[0020] By comparing the limit stress and the stress threshold value at each point, it is judged whether each point can withstand the limit condition; when all points can withstand the limit condition, it is judged that the quality acceptance requirement is met.

[0021] As a preferred scheme of the aluminum profile machining quality detection method, the three-dimensional geometric model of the aluminum profile is discretized in space to construct a three-dimensional volume element grid, and each grid node corresponds to a spatial coordinate point (x, y, z); wherein each grid point is an independent point;

[0022] According to the limit condition, a normal force is applied at any position of the aluminum profile, and the pressure is a quasi-static constant load; a three-dimensional nonlinear mechanics simulation is performed using a finite element solver;

[0023] In the solution result of the force applied at each position, find the maximum equivalent stress point appearing in the entire model; record the stress response values of all other points when the maximum point appears; and construct a three-dimensional stress propagation map with the maximum stress point as the core;

[0024] The limit stress is represented as: {Z1, Z2,..., Z n};

[0025] Wherein, Z n =(σ max，n ,σ(S) n ); Z n represents the limit stress of the nth applied position; n represents the number of applied positions; σ max，n represents the maximum value of the stress at the nth applied position and the coordinate position corresponding to the maximum value; S represents the set of all points;

[0026] σ(S) n ={σ 1，n ,σ 2，n ,...,σ f，n} represents the set of characteristic points at the nth applied position; σ f，n represents the fth coordinate position at the nth applied position and the stress at the coordinate position.

[0027] As a preferred scheme of the aluminum profile machining quality detection method, the inversion analysis includes, for each point, the detection data of each modal, the mean value is taken as the characteristic value T={T1, T2,..., T j} of each point; the characteristic value is taken as input, and the occurrence probability of each defect at each point is output by using Bayesian network;

[0028] Using the typical detection data preselected for each defect, the proportion of comprehensive defects at each point is calculated: the typical detection data preselected for each defect is proportioned according to an arbitrary ratio, so that the proportioning result meets the constraint:

[0029] wherein, is an element embedded in the corresponding Q g , both of which appear simultaneously; D g，i represents the gth defect type, and the characteristic value of the detection data i; T i represents the characteristic value of the detection data i at the point; i represents the index of the detection data type; j represents the number of detection data types; Q g represents the proportion of the gth defect type; G represents the number of defect types; represents the proportion coefficient of the gth defect type, which is not less than 0 and less than represents the maximum proportion coefficient of the gth defect type;

[0030] Get all proportioning sets B e that meet the constraint = {b1, b2,..., b c}; b c represents the cth proportioning, and c represents the number of obtained proportioning, B e represents the proportioning set of the e th point;

[0031] Let: Q e，G represents the proportion of the Gth defect type of the e th point; b e，0 represents the proportioning composed of the output probability of the Bayesian network in the e th point; represents the proportion coefficient embedded in Q e，G ;

[0032] Screen the proportioning set that is most similar to b e，0 , and record it as: Q e，1 : Q e，2 :... : Q e，G ;

[0033] In the screened proportioning, for each element Q e，G , extract the proportion coefficient embedded therein

[0034] Use the pre-fitted function to input the proportion coefficient and the defect type, and output the reduction rate of the maximum bearing stress value;

[0035] Calculate the reduction rate of each node for each defect type, use the screened proportioning as the weight coefficient, perform weighted summation to obtain the actual reduction rate, and use the actual reduction rate to reduce the standard maximum bearing stress value of the node to obtain the stress threshold.

[0036] An aluminum profile machining quality detection system, wherein: a detection unit acquires multi-modal defect detection data and user quality acceptance requirements;

[0037] An analysis unit inversely analyzes the multi-modal defect detection data according to the quality acceptance requirements;

[0038] A judgment unit judges whether the quality detection result is qualified or unqualified according to the analysis result.

[0039] A computer device, comprising a memory and a processor; the memory stores a computer program, wherein the processor implements the steps of the method of any one of the embodiments when executing the computer program.

[0040] A computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement the steps of the method of any one of the embodiments.

[0041] The aluminum profile machining quality detection method provided by the present application can realize collaborative analysis of multi-modal defect data and use working conditions, breaking through the limitations of traditional detection which only relies on surface results or fixed indicators for judgment. By constructing a three-dimensional model and introducing full-field limit working condition simulation, potential failure points of the aluminum profile under any load path are effectively identified, improving the comprehensiveness and reliability of structural safety evaluation. Combined with Bayesian network and matching optimization algorithm, the quantitative inversion of the influence degree of multiple defects is realized, and the dynamic reduction of node mechanical properties according to the defect proportion is realized to ensure that the evaluation result is closer to the actual service state. This method is not only suitable for intelligent screening in batch detection, but also can provide basis for quality prediction and process feedback in complex use scenarios, and has the advantages of high detection precision, strong adaptability, high intelligent degree and the like. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0043] Figure 1 A whole flow chart of an aluminum profile machining quality detection method provided by the present application. DETAILED DESCRIPTION

[0044] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0045] With reference to Figure 1 For an embodiment of the present application, an aluminum profile processing quality detection method is provided, comprising:

[0046] S1: Obtain multi-modal defect detection data and user quality acceptance requirements.

[0047] The multi-modal defect detection data includes, for each quality defect type of the aluminum profile, using a corresponding detection means respectively to perform defect detection, and obtaining detection data under each detection means. For crack type defects, methods such as acoustic emission detection, ultrasonic detection or eddy current detection are adopted to obtain high-frequency acoustic signals, ultrasonic echo time attenuation data or eddy current impedance change values generated when cracks initiate or expand, and then extract crack length, depth and position and other characteristic parameters; for internal porosity type defects, methods such as multi-frequency eddy current detection, X-ray digital imaging or CT scanning are adopted to determine the size, density and spatial distribution of the pores through phase delay characteristics, image gray scale distribution or voxel level hole structure; for surface flatness abnormalities, methods such as structured light projection and three-dimensional laser scanning are adopted to obtain point cloud data of the surface of the aluminum profile, and the warping and concave-convex state is evaluated through calculating local curvature, profile height difference and other information. The original data obtained by each detection means constitutes multi-modal input, providing a data basis for subsequent defect recognition, feature extraction and quality evaluation.

[0048] The user's quality acceptance requirements include the user's minimum indicators for each quality defect detection result and the user's use requirements for the aluminum profile. The quality defect detection result includes using the detection data under each detection means to identify the quality defects of the aluminum profile respectively, and generating quantifiable parameters for each defect; specifically including: for crack type defects, the generated quantifiable parameters include crack length, depth, number or acoustic emission energy value. For flatness defects, the generated quantifiable parameters include local curvature value, maximum height difference value, warping deviation value. For internal porosity defects, the generated quantifiable parameters include maximum pore diameter, pore density per unit area, and eddy current phase delay amplitude change.

[0049] The minimum index is a limit of a quantifiable parameter in a single quality defect detection result of the aluminum profile; if the detection result of any quality defect of the aluminum profile does not meet the limit of the minimum index, the inversion analysis stage is skipped, and the quality detection result is directly output as qualified; if the detection result of each quality defect of the aluminum profile meets the limit of the minimum index, inversion analysis is performed according to the use requirement of the user for the aluminum profile and the detection data under each detection means.

[0050] After the multi-modal defect data is acquired, an independent judgment threshold is set for each type of quality defect to quickly identify serious defect conditions, thereby constructing a front-end judgment threshold in system analysis. If any defect parameter in the detection result does not reach the minimum index, it means that the defect can be ignored or does not constitute an actual safety risk in the current use situation, so there is no need to enter the subsequent complex inversion analysis stage, and the detection process efficiency and adaptability can be improved; when all defects reach the minimum index, i.e., all defect parameters have potential impact risks, the inversion analysis stage is entered, and through three-dimensional simulation and performance inversion, the overall structure safety is further evaluated in detail by combining the user's use requirement. This mechanism realizes "defect pre-screening + grading judgment + intelligent shunting", not only reduces the system calculation burden, but also improves the evaluation pertinence and engineering adaptability, and embodies the dynamic decision-making advantage under data driving.

[0051] The use requirement of the user for the aluminum profile includes the requirement of the user for the aluminum profile to be normally used under specific use conditions. When the user sets the use requirement, the user provides the limit condition under each working condition. The limit condition includes the maximum value of the working condition parameter, which is composed of the maximum pressure and the minimum stress area. The actual use scene requirement of the user for the aluminum profile is included in the quality detection process, so that the detection evaluation result is not only based on the standardized defect data judgment, but also meets the safety requirement of the specific application scene. By constructing the use requirement model from the "working condition" input by the user, the quality evaluation is not limited to the identification of the defects of the material itself, but also focuses on the carrying capacity and service performance in the target application

[0052] S2: performing inversion analysis on the multi-modal defect detection data according to the quality acceptance requirement.

[0053] Performing inversion analysis on the multi-modal defect detection data includes making an assumption that any position of the aluminum profile can be an action point under the limit condition. Under the premise of the assumption, the stress generated at each point is simulated to obtain the limit stress at each point.

[0054] The stress threshold of each point of the aluminum profile is obtained by respectively performing inversion analysis on each detection data.

[0055] By assuming each point of the aluminum profile as a potential limit load point, comprehensive evaluation under the most unfavorable working condition is realized. Compared with the traditional local simulation at a specific loading point, the method obtains the stress response value of each point under the maximum load through full-structure traversal simulation, forms a "limit stress distribution map", and reflects the potential weak points or vulnerable areas. At the same time, in combination with the multi-modal detection data corresponding to each point, the stress bearing capacity (i.e. stress threshold) of the material at the position under the current defect state is back calculated, ensuring that the evaluation result is not only based on theoretical calculation, but also fully reflects the influence of actual defects on the structural strength. Finally, through the point-by-point comparison of the limit stress and the stress threshold, the overall stability of the structure under the limit working condition is judged, realizing the fusion of the "structural integrity + defect tolerance" judgment standard. The design improves the rigor and adaptability of the structure quality evaluation, effectively supporting the reliability screening and risk warning in key application scenarios.

[0056] The three-dimensional geometric model of the aluminum profile is discretized in space to construct a three-dimensional volume element grid, and each grid node corresponds to a spatial coordinate point (x, y, z); wherein each grid point is an independent point.

[0057] According to the limit condition, a normal force is applied at any position of the aluminum profile, and the pressure is a quasi-static constant load; a three-dimensional nonlinear mechanics simulation is performed using a finite element solver. Several preset force application positions can be simulated, or a plurality of force application positions can be randomly extracted from the position center.

[0058] In the solution result of the force applied at each position, the maximum equivalent stress point appearing in the entire model is found; the stress response values of all other points when the maximum point appears are recorded; and a three-dimensional stress propagation map with the maximum stress point as the core is constructed.

[0059] The limit stress is represented as: {Z1, Z2,..., Zn}. n}.

[0060] Wherein, Z n = (σ max，n , σ(S) n ) ; Z n represents the limit stress of the nth applied position; n represents the number of applied positions; σ max，n represents the maximum value of the stress and the coordinate position corresponding to the maximum value under the nth applied position. S represents all point position sets. σ(S) n 1，n 2，n f，n σn represents the feature point set under the n-th application position; σ f，n represents the stress size at the f-th coordinate position and the coordinate position under the n-th application position.

[0061] By establishing a three-dimensional discrete model of the aluminum profile structure and simulating loading at multiple possible force application points, a stress propagation map based on the limit working condition is constructed to provide structural level stress risk identification capability for quality detection. By spatially discretizing the aluminum profile model, each grid point has a unique three-dimensional coordinate identity and can serve as an independent loading or response node, ensuring that the stress analysis has position resolution. By using a finite element solver to apply a quasi-static normal constant load at several representative or randomly sampled force application positions, the influence of common or extreme working conditions on the overall structure can be effectively simulated, and the high computational cost of full-point loading can be avoided. By extracting the maximum equivalent stress points and tracking the overall response under each loading condition, a three-dimensional stress propagation map centered on the stress peak value can be obtained. This map reflects the diffusion and concentration of internal stress states under external force application, providing a basis for subsequent judgment of whether each point meets its local stress threshold.

[0062] Further, the inversion analysis includes, for each point, each modal detection data, calculating the mean value as the eigenvalue T = {T1, T2,..., T j} of each point; using the eigenvalue as input, using the Bayesian network to output the occurrence probability of each defect at each point.

[0063] Using the pre-selected typical detection data of each defect (one defect may be reflected in multiple detection data), the comprehensive defect proportion of each point is calculated: the pre-selected typical detection data of each defect is proportioned according to an arbitrary ratio, so that the proportioning result meets the constraint:

[0064] wherein, is the element embedded in the corresponding Q g , both of which appear simultaneously; D g，i represents the eigenvalue of the g-th defect type for detection data i; T i represents the eigenvalue of the point for detection data i; i represents the index of the detection data type; j represents the number of detection data types; Q g represents the proportion of the g-th defect type; G represents the number of defect types; represents the proportion coefficient of the g-th defect type, which is not less than 0 and less than ​​​represents the maximum proportion coefficient of the gth defect type.

[0065] Get all matching ratio set B e ={b1, b2,..., b c}; b c represents the cth ratio, c represents the number of obtained ratios, B e represents the ratio set of the e th point.

[0066] Let: Q e，G represents the proportion of the gth defect type at the e th point; b e，0 represents the ratio of the e th point, which is composed of the output probability of the Bayesian network; represents the proportion coefficient embedded in Q e，G .

[0067] Screen the ratio set closest to b e，0 , denoted as: Q e，1 : Q e，2 :... : Q e，G .

[0068] In the screened ratio, for each element Q e，G , extract the proportion coefficient embedded therein

[0069] On the basis of multi-modal defect detection data, a ratio analysis mechanism capable of reverse inference of defect type proportion is established. Considering that a certain type of defect (such as crack or porosity) may be reflected in multiple detection data, and there is cross-sensitivity between different detection features, the influence degree of the defect cannot be accurately quantified by relying on single modal data only. Therefore, the system pre-sets the mapping samples between typical defects and their feature responses, constructs multiple feasible defect proportion ratio sets based on the feature values of different detection data dimensions, and under the premise of meeting certain constraint conditions (such as the embedded features must appear synchronously and the proportion must be within a reasonable range), the closest proportion result to the true state is screened out by minimizing the distance matching with the probability ratio output by the Bayesian model.

[0070] Finally, the defect proportion coefficient is extracted from the optimal ratio for calculating the reduction effect of each type of defect on the stress bearing capacity of the material, realizing the accurate conversion from “detection features” to “mechanical damage quantification”. This method improves the utilization rate of multi-modal data, solves the uncertainty problem of evaluation caused by mixed and mutual influence of defects in complex scenarios, and enhances the interpretability and reliability of quality detection results.

[0071] Further, a pre-fitted function is used to input the proportion coefficient and the defect type, and output the reduction rate of the maximum bearing stress value. In the design, a quantitative relationship function between the proportion coefficient and the maximum bearing stress of the material is established for each defect type, and a pre-fitted regression model or an interpolation function is used to realize the mapping from the defect proportion to the stress reduction rate. In engineering implementation, the function can be constructed based on a large number of actual tests or finite element simulation data through regression fitting, support vector regression (SVR), polynomial fitting or neural network model, etc. For example, for crack defects, the influence curve can be fitted by different crack lengths and material strength test data; for pore defects, the mapping relationship driven by data can be generated by simulating the weakening effect of different pore diameters on the yield stress.

[0072] In actual use, the system inputs the defect type and the corresponding proportion coefficient as parameters into the function, and quickly outputs the reduction rate of each defect on the mechanical properties of the material. Since the function structure has been pre-fitted through offline training, only parameters need to be substituted for calculation in online use, which has small calculation amount, fast response, and is suitable for embedding in a quality detection system to realize automatic and real-time evaluation. In addition, the function model has scalability, and can be retrained and adaptively updated according to product categories, material types or new detection features, to ensure continuous improvement of analysis accuracy and application scope.

[0073] The actual reduction rate is obtained by calculating each node, weighting and summing the reduction rate of each defect type using the proportion obtained by screening as a weight coefficient. The standard maximum bearing stress value of the node is reduced using the actual reduction rate to obtain the stress threshold value.

[0074] S3: According to the analysis result, it is judged whether it meets the quality acceptance requirement, if it meets, it is judged that the quality detection result is qualified; otherwise, it is judged that the quality detection result is unqualified.

[0075] On the other hand, the embodiment also provides an aluminum profile processing quality detection system, which comprises: a detection unit, which obtains multi-modal defect detection data and user quality acceptance requirements.

[0076] An analysis unit inversely analyzes the multi-modal defect detection data according to the quality acceptance requirements.

[0077] A judgment unit judges whether the analysis result meets the quality acceptance requirement, if it meets, it is judged that the quality detection result is qualified; otherwise, it is judged that the quality detection result is unqualified.

[0078] If the above functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or parts of the present application that make contributions to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0079] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a list of executable instructions for implementing logic functions, which can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor or other system that can fetch the instructions from the instruction execution system, apparatus or device and execute the instructions, or in conjunction with these instruction execution systems, apparatus or devices. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by or in connection with an instruction execution system, apparatus or device, or in conjunction with these instruction execution systems, apparatus or devices.

[0080] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection having one or more wires (electrical devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, because the program can be obtained electronically, for example, by optical scanning of the paper or other medium, followed by editing, interpreting or otherwise processing, if necessary, in other suitable ways to obtain the program electronically, and then storing it in a computer memory.

[0081] It should be understood that portions of the present application can be implemented in hardware, software, firmware, or combinations thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, implementation can be with any or a combination of the following technologies, which are all well known in the art: a discrete logic circuit having logic gates for implementing logic functions upon an application of data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0082] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and all should be covered in the scope of the claims of the present application.

Claims

1. A method for inspecting the processing quality of aluminum profiles, characterized in that, include: Acquire multimodal defect detection data and user quality acceptance requirements; Based on the aforementioned quality acceptance requirements, inversion analysis is performed on the multimodal defect detection data; Based on the analysis results, determine whether the quality acceptance requirements are met. If they are met, the quality inspection result is deemed qualified. Conversely, the quality inspection result is judged as unqualified. The inversion analysis of the multimodal defect detection data includes making the following assumptions about the aluminum profile: any position of the aluminum profile may be used as a point of action under extreme conditions; Under the aforementioned assumptions, the stress that may occur at each point is simulated to obtain the ultimate stress at each point; At each point on the aluminum profile, inversion analysis is performed on each type of test data to obtain the stress threshold at each analysis point; By comparing the ultimate stress at each point with the stress threshold, it is determined whether each point can withstand the ultimate condition; when all points can withstand the ultimate condition, it is determined that the quality acceptance requirements are met. The three-dimensional geometric model of the aluminum profile is spatially discretized to construct a three-dimensional volume element mesh, where each mesh node corresponds to a spatial coordinate point (x, y, z); each mesh point is treated as an independent point. Based on the aforementioned limit conditions, a normal force is applied at any position on the aluminum profile, with the pressure being a quasi-static constant load; a three-dimensional nonlinear mechanical simulation is performed using a finite element solver. In the solution results of applying force at each location, find the point of maximum equivalent stress in the entire model; record the stress response values ​​of all other points when the maximum stress point appears; construct a three-dimensional stress propagation map with the maximum stress point as the core; The ultimate stress is expressed as: ; in, ; This represents the ultimate stress at the nth application location; n represents the number of application locations. This represents the maximum stress value and the corresponding coordinate position at the nth applied position. ; Represents the set of all points; Let n be the set of feature points at the nth application position; This represents the f-th coordinate position and the magnitude of the stress at the n-th application position; The inversion analysis includes calculating the mean value of the detection data for each mode at each location as the feature value for each location. Using the feature values ​​as input, a Bayesian network is used to output the probability of occurrence of each type of defect at each location. Using typical detection data pre-selected for each type of defect, the overall defect ratio is calculated for each location. Typical detection data pre-selected for each type of defect are then allocated in any proportion to ensure the allocation results meet the following constraints: ; in, It is embedded into the corresponding The elements in the text, both appearing simultaneously; This represents the g-th defect type, and is a feature value for detection data i. This represents the feature value of detection data i at the specified location; i represents the index of the detection data type; j represents the number of detection data types. This represents the percentage of the g-th defect type; G represents the number of defect types. The proportionality coefficient representing the g-th defect type takes a value that is not less than 0 and is less than 1. ; This represents the maximum scaling factor for the g-th defect type; Obtain the set of all matching ratios that satisfy the constraints. ; This represents the c-th proportion, where c represents the quantity of the resulting proportion. This represents the set of allocation ratios for the e-th point; set up: This represents the percentage of the G-th defect type at the e-th point; This represents the proportion of the output probabilities of the Bayesian network at the e-th point; Indicates embedding The proportionality coefficient; In the selection of proportions, with The closest ratio is denoted as: : :...: ; In the selected ratios, for each element Extract the embedded proportional coefficients. ; Using a pre-fitted function, input the scaling factor and defect type, and output the reduction rate to the maximum stress value; For each node, the reduction rate for each defect type is calculated. The weighted summation is performed using the selected ratio as a weighting coefficient to obtain the actual reduction rate. The standard maximum stress value of the node is reduced using the actual reduction rate to obtain the stress threshold.

2. The method for inspecting the processing quality of aluminum profiles as described in claim 1, characterized in that: The multimodal defect detection data includes the use of corresponding detection methods for each type of quality defect in the aluminum profile to obtain detection data for each detection method.

3. The method for inspecting the processing quality of aluminum profiles as described in claim 2, characterized in that: The user's quality acceptance requirements include the minimum indicators for each quality defect detection result and the user's usage requirements for aluminum profiles. The quality defect detection results include identifying quality defects in aluminum profiles using detection data from each detection method, and generating quantifiable parameters for each type of defect. The minimum index refers to the limitation of quantifiable parameters in the detection results of a single quality defect of aluminum profile. If the detection result of any quality defect of aluminum profile does not meet the limitation of the minimum index, the inversion analysis stage is skipped and the quality detection result is directly output as qualified. If the detection result of each quality defect of aluminum profile meets the limitation of the minimum index, the inversion analysis is performed based on the user's usage requirements for aluminum profile and the detection data under each detection method.

4. The method for inspecting the processing quality of aluminum profiles as described in claim 3, characterized in that: The user's requirements for aluminum profiles include the user's requirement that aluminum profiles can be used normally under specific working conditions. Among them, when setting usage requirements, users provide the extreme conditions for each working condition; The limiting conditions include the maximum values ​​of the operating condition parameters, which are determined by two parameters: maximum pressure and minimum force-bearing area.

5. A quality inspection system for aluminum profile processing using the method described in any one of claims 1-4, characterized in that: The detection unit acquires multimodal defect detection data and user quality acceptance requirements; The analysis unit performs inversion analysis on the multimodal defect detection data according to the quality acceptance requirements. The judgment unit determines whether the quality acceptance requirements are met based on the analysis results. If they are met, the quality inspection result is judged to be qualified. Conversely, the quality inspection result is judged as unqualified.

6. A computer device, comprising: Memory and processor; The memory stores a computer program, characterized in that: when the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-4.

Images