Aluminum profile section thickness consistency detection method based on X-ray scanning

Abstract

The invention relates to the technical field of aluminum profile processing, in particular to an aluminum profile section thickness consistency detection method based on X-ray scanning, which comprises the following steps: S1, aluminum profile pretreatment: pretreating an aluminum profile to be detected; s2, X-ray total cross-section scanning: obtaining initial thickness distribution image data; s3, multi-source data fusion: integrating with material component data, production process parameters and environmental condition data to generate a comprehensive thickness measurement model; s4, thickness prediction analysis: predicting thickness distribution of different areas; s5, thickness consistency evaluation is conducted, specifically, the thickness consistency of the aluminum profile is evaluated; and S6, tracing and analyzing abnormal reasons: identifying potential roots influencing the thickness consistency. According to the invention, the sensitivity and accuracy of thickness detection are improved, the possibility of false alarm and missing alarm is reduced, the uniformity and stability of the product thickness are ensured, and the reliability of the product is improved.

Description

Technical Field

[0001] The invention relates to the technical field of aluminum profile processing, and in particular to a method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning. Background Art

[0002] As an important industrial material, aluminum profiles are widely used in many fields such as construction, transportation, electronics and aviation. The uniformity of their cross-sectional thickness directly affects the mechanical properties and service life of the product. Therefore, in the production process of aluminum profiles, thickness consistency detection becomes a key link in quality control. In order to ensure product quality, manufacturers often need to accurately detect the cross-sectional thickness of aluminum profiles to identify potential thickness inconsistency problems and make timely adjustments to meet strict product performance requirements.

[0003] At present, common thickness detection methods mostly use technologies such as ultrasonic measurement, laser ranging and mechanical contact. However, these methods have many shortcomings in accuracy, efficiency and applicability. For example, ultrasonic measurement is easily affected by the surface state of the material, laser ranging is difficult to penetrate metal materials, and mechanical contact measurement may affect the surface quality of the workpiece. In addition, existing thickness detection methods usually rely only on a single data source and cannot effectively integrate multiple factors such as material composition, production parameters and environmental conditions, resulting in inaccurate and incomplete detection results, and unable to conduct in-depth analysis of the causes of thickness inconsistency.

[0004] The purpose of the present invention is to provide a method for detecting the thickness consistency of aluminum profile cross sections based on X-ray scanning, so as to comprehensively improve the accuracy and reliability of thickness detection, help manufacturers achieve more efficient quality control, and ensure the product performance and production efficiency of aluminum profiles. Summary of the invention

[0005] The invention provides a method for detecting the consistency of the cross-section thickness of an aluminum profile based on X-ray scanning.

[0006] The method for detecting the consistency of the cross-section thickness of aluminum profiles based on X-ray scanning comprises the following steps:

[0007] S1, aluminum profile pretreatment: pretreatment of the aluminum profile to be tested, including cleaning and removal of surface impurities, oil stains or substances that affect X-ray imaging;

[0008] S2, X-ray full-section scanning: Use X-ray equipment to scan the full section of the pre-treated aluminum profile to obtain preliminary thickness distribution image data;

[0009] S3, multi-source data fusion: Based on the obtained thickness distribution image data, a multi-source data fusion algorithm is applied to integrate it with the material composition data, production process parameters, and environmental condition data to generate a comprehensive thickness measurement model;

[0010] S4, thickness prediction analysis: By analyzing the microstructural characteristics of aluminum profiles, the thickness distribution of different areas is predicted and potential thickness inconsistency issues are identified, including:

[0011] S41, microstructure feature extraction: use imaging technology to analyze aluminum profiles and extract the microstructure features of aluminum profiles, including grain size, phase composition and defect distribution;

[0012] S42, thickness distribution model construction: construct a thickness distribution model based on the extracted microstructure features;

[0013] S43, potential inconsistency identification: using the thickness distribution model to predict the predicted thickness of different areas and compare it with the output of the comprehensive thickness measurement model to identify potential thickness inconsistency problems;

[0014] S5, thickness consistency assessment: Based on the results of thickness distribution in different areas, the thickness consistency of aluminum profiles is assessed and a test report is generated;

[0015] S6, abnormal cause tracing analysis: for the identified thickness inconsistency problem, abnormal cause tracing analysis is carried out, and historical production records, material batches and process parameters are analyzed using data mining technology to identify potential root causes that affect thickness consistency.

[0016] Optionally, the X-ray full-section scan in S2 includes:

[0017] S21, equipment setting: according to the size and characteristics of the aluminum profile, adjust the parameters of the X-ray equipment, including the ray intensity, scanning speed and imaging resolution;

[0018] S22, sample fixing: mounting the pretreated aluminum profile on the X-ray scanning platform;

[0019] S23, full-section scanning: start the X-ray equipment, perform a full-section scan on the aluminum profile, record the attenuation data after the X-ray penetrates the aluminum profile, and generate preliminary thickness distribution image data.

[0020] Optionally, the multi-source data fusion in S3 includes:

[0021] S31, multi-source data acquisition and processing: collect multi-source data of aluminum profiles, including material composition data, production process parameters, and environmental condition data, and process the multi-source data and preliminary thickness distribution image data, including data cleaning, denoising, and normalization;

[0022] S32, fusion algorithm application: Apply the multi-source data fusion algorithm to integrate the processed multi-source data and the preliminary thickness distribution image data to generate a comprehensive thickness measurement model.

[0023] Optionally, the multi-source data collection and processing in S31 includes:

[0024] S311, Multi-source data collection: Collect multi-source data of aluminum profiles through sensors and data recording units, including:

[0025] Material composition data collection: Use spectral analysis technology to obtain the content of each component;

[0026] Production process parameter collection: Real-time monitoring of the processing temperature, extrusion pressure and conveying speed of aluminum profiles during processing through industrial automation systems;

[0027] Environmental condition data collection: Environmental sensors are used to monitor the ambient humidity, ambient temperature and ambient air pressure in the aluminum profile production environment;

[0028] S312, data processing: processing the collected multi-source data and preliminary thickness distribution image data, specifically including:

[0029] Data cleaning: Use Z-score standardization to clean multi-source data and preliminary thickness distribution image data;

[0030] Denoising: Use moving average to reduce random noise in the data;

[0031] Normalization: Use min-max normalization to normalize the data to the same range.

[0032] Optionally, the fusion algorithm application in S32 includes:

[0033] S321, feature weighted calculation: weighting the processed multi-source data and preliminary thickness distribution image data;

[0034] S322, fusion model construction: based on the weighted eigenvalues, a thickness measurement model is constructed using a linear regression model.

[0035] Optionally, the microstructure feature extraction in S41 includes:

[0036] S411, Application of imaging technology: X-ray tomography is used to image the microstructure of aluminum profiles to obtain distribution images of grains, phase composition and defects;

[0037] S412, image feature extraction: Analyze the imaging data through image processing algorithms to extract microstructure features, including:

[0038] Grain size calculation: The imaging data is segmented using the watershed algorithm to separate the grain area from the background area;

[0039] Phase composition analysis: Separate different phases (such as aluminum matrix phase and reinforcement phase) from the image by the area fraction of the phase composition, and calculate the area fraction V of each phase 相 ;

[0040] Defect distribution detection: Use edge detection algorithms to identify and locate defects in the microstructure (such as holes and cracks) and calculate the defect density ρ 缺陷 .

[0041] Optionally, the thickness distribution model in S42 adopts a Gaussian process regression model, and the Gaussian process regression model includes:

[0042] S421, define a composite kernel function with microstructural features: in the composite kernel function, microstructural features (grain size, phase composition ratio and defect density) are used as input features and are assigned different weights to reflect different effects on thickness;

[0043] S422, define the weighted RBF kernel function of microstructure features: the RBF kernel is used to capture the nonlinear relationship in the microstructure features, and introduces the feature weight w in the RBF kernel i And adaptive bandwidth l i , to accommodate the scale differences of different microstructural features;

[0044] S423, defining a weighted linear kernel function of microstructural features: the linear kernel is used to capture the linear relationship in the microstructural features, and feature weights are added to accommodate the impact of different features on the thickness distribution;

[0045] S424, construct a thickness distribution prediction model based on microstructure characteristics: output the predicted thickness distribution value according to the defined composite kernel function

[0046] Optionally, the potential inconsistency identification in S43 includes:

[0047] S431, predicted thickness calculation: Use the thickness distribution model to predict the thickness of different areas of the aluminum profile and obtain the predicted thickness of each detection area

[0048] S432, Standard Thickness Comparison: The predicted thickness of each area With the output of the integrated thickness measurement model T i Compare and calculate thickness deviation ΔT i ;

[0049] S4323, inconsistency identification: by setting the thickness deviation tolerance threshold ∈ and the thickness deviation ΔT i For comparison, if the thickness deviation ΔT iExceeding the thickness deviation tolerance threshold ∈ indicates that there is a potential thickness inconsistency problem in the area.

[0050] Optionally, the thickness consistency assessment in S5 includes:

[0051] S51, overall thickness deviation calculation: based on the calculated thickness deviation ΔT i , calculate the root mean square error RMSE of the overall thickness deviation to evaluate the overall thickness consistency level of the aluminum profile;

[0052] S52, thickness consistency report generation: Generate a thickness consistency evaluation report based on the RMSE result of the overall thickness deviation. The report includes the overall thickness consistency evaluation (RMSE value) and the conclusion on whether it meets the production quality standards.

[0053] Optionally, the abnormal cause tracing analysis in S6 includes:

[0054] S61, Historical data collection: Collect historical production data related to thickness inconsistency, including material batches, process parameters (such as processing temperature, pressure, speed, etc.) and quality inspection records;

[0055] S62, correlation analysis: using correlation analysis to calculate the correlation coefficient r between each feature and thickness inconsistency;

[0056] S63, abnormal pattern recognition: K-means clustering algorithm is used to cluster abnormal data, samples with similar abnormal characteristics are divided into different categories, and abnormal patterns are identified by minimizing the square error J;

[0057] S64, generation of traceability analysis report: based on the results of correlation analysis and abnormal pattern recognition, a traceability report of abnormal causes is generated, which includes factors affecting thickness consistency, abnormal patterns and their corresponding process conditions.

[0058] Beneficial effects of the present invention:

[0059] The present invention realizes the whole process management of thickness consistency detection from aluminum profile preprocessing, full-section scanning, multi-source data fusion to thickness prediction analysis through an X-ray-based aluminum profile cross-section thickness consistency detection method. By fusing the X-ray scanning results with multi-source data (material composition, production parameters, environmental conditions), a comprehensive thickness measurement model is generated, which not only effectively improves the accuracy and reliability of thickness measurement, but also enables the model to adapt to changes in production conditions, thereby providing solid data support for the quality control of aluminum profiles.

[0060] The present invention, through a method combining microstructure feature extraction with a Gaussian process regression model, constructs a thickness distribution model starting from microstructure features (such as grain size, phase composition ratio, and defect density), accurately predicts the thickness of different regions, and can clearly identify potential thickness inconsistency problems by comparing with the output value of a comprehensive thickness measurement model, and achieve overall thickness consistency evaluation in multiple regions. This method greatly improves the sensitivity and accuracy of thickness detection, reduces the possibility of false alarms and missed alarms, helps to ensure the uniformity and stability of product thickness, and improves product reliability.

[0061] The present invention uses data mining technology to perform correlation analysis and abnormal pattern recognition on historical production records, material batches and process parameters, systematically identifies potential root causes that affect thickness consistency, and generates a traceability analysis report that provides key influencing factors and improvement suggestions, providing a scientific basis for continuous optimization and quality improvement of the production process, thereby improving the overall production quality and economic benefits of aluminum profiles and enhancing the competitiveness of products in the market. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0063] Figure 1 A schematic diagram of a detection method flow chart of an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of thickness prediction analysis according to an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0066] like Figure 1-Figure 2 As shown, the method for detecting the consistency of the cross-section thickness of aluminum profiles based on X-ray scanning includes the following steps:

[0067] S1, aluminum profile pretreatment: pretreatment of the aluminum profile to be tested, including cleaning and removal of surface impurities, oil stains or substances that affect X-ray imaging;

[0068] S2, X-ray full-section scanning: Use X-ray equipment to scan the full section of the pre-treated aluminum profile to obtain preliminary thickness distribution image data;

[0069] S3, multi-source data fusion: Based on the obtained thickness distribution image data, a multi-source data fusion algorithm is applied to integrate it with the material composition data, production process parameters, and environmental condition data to generate a comprehensive thickness measurement model;

[0070] S4, thickness prediction analysis: By analyzing the microstructural characteristics of aluminum profiles, the thickness distribution of different areas is predicted and potential thickness inconsistency issues are identified, including:

[0071] S41, microstructure feature extraction: use imaging technology to analyze aluminum profiles and extract the microstructure features of aluminum profiles, including grain size, phase composition and defect distribution;

[0072] S42, thickness distribution model construction: construct a thickness distribution model based on the extracted microstructure features;

[0073] S43, potential inconsistency identification: using the thickness distribution model to predict the predicted thickness of different areas and compare it with the output of the comprehensive thickness measurement model to identify potential thickness inconsistency problems;

[0074] S5, thickness consistency assessment: Based on the results of thickness distribution in different areas, the thickness consistency of aluminum profiles is assessed and a test report is generated;

[0075] S6, abnormal cause tracing analysis: for the identified thickness inconsistency problem, abnormal cause tracing analysis is carried out, and historical production records, material batches and process parameters are analyzed using data mining technology to identify potential root causes affecting thickness consistency;

[0076] Through the above content, from the pretreatment of aluminum profiles to the thickness consistency assessment, the integration of multi-source data and microstructure analysis has improved the accuracy and reliability of detection, and can effectively identify potential thickness inconsistency problems, provide comprehensive quality control measures, and provide data support for the optimization and improvement of the production process through abnormal cause tracing analysis, thereby improving the overall production quality and economic benefits of aluminum profiles.

[0077] The X-ray full section scan in S2 includes:

[0078] S21, equipment setting: according to the size and characteristics of the aluminum profile, adjust the parameters of the X-ray equipment, including the ray intensity, scanning speed and imaging resolution, to ensure the best scanning effect;

[0079] S22, sample fixation: the pre-treated aluminum profile is mounted on the X-ray scanning platform to ensure that the sample does not move during the scanning process to improve the stability and accuracy of the image;

[0080] S23, full-section scanning: start the X-ray equipment, perform full-section scanning on the aluminum profile, record the attenuation data after the X-ray penetrates the aluminum profile, and generate preliminary thickness distribution image data;

[0081] Through the above, the thickness distribution image data of aluminum profiles can be obtained efficiently and accurately, ensuring the optimal setting of the equipment and the stability of the sample, thereby improving the quality and reliability of imaging, greatly enhancing the quality control capability of aluminum profiles, and helping to identify and correct potential problems in a timely manner during the production process.

[0082] Multi-source data fusion in S3 includes:

[0083] S31, multi-source data acquisition and processing: collect multi-source data of aluminum profiles, including material composition data, production process parameters, and environmental condition data, and process the multi-source data and preliminary thickness distribution image data, including data cleaning, denoising, and normalization;

[0084] S32, fusion algorithm application: applying a multi-source data fusion algorithm to integrate the processed multi-source data and the preliminary thickness distribution image data to generate a comprehensive thickness measurement model;

[0085] Through the above content, comprehensive background information can be provided for thickness measurement, which not only reduces data noise and inconsistency, but also uses fusion algorithms to generate a comprehensive thickness measurement model, thereby more accurately reflecting the actual thickness distribution of aluminum profiles, enhancing quality control capabilities, and helping manufacturers to identify and correct potential problems in a timely manner during the production process, ultimately improving the overall quality and market competitiveness of the product.

[0086] Multi-source data acquisition and processing in S31 include:

[0087] S311, Multi-source data collection: Collect multi-source data of aluminum profiles through sensors and data recording units, including:

[0088] Material composition data collection: Use spectral analysis technology to obtain the content of each component;

[0089] Production process parameter collection: Real-time monitoring of the processing temperature, extrusion pressure and conveying speed of aluminum profiles during processing through industrial automation systems;

[0090] Environmental condition data collection: Environmental sensors are used to monitor the ambient humidity, ambient temperature and ambient air pressure in the aluminum profile production environment;

[0091] S312, data processing: processing the collected multi-source data and preliminary thickness distribution image data, specifically including:

[0092] Data cleaning: Z-score standardization is used to clean multi-source data and preliminary thickness distribution image data, expressed as:

[0093]

[0094] Among them, X is the original data, μ is the data mean, and σ is the standard deviation;

[0095] Denoising: Use the moving average method to reduce random noise in the data, expressed as:

[0096]

[0097] Among them, S n is the smoothed data, N is the window size, X i is the original data point;

[0098] Normalization: Use minimum-maximum normalization to normalize the data to the same range, expressed as:

[0099]

[0100] Among them, X′ is the normalized data, X min and X max are the minimum and maximum values ​​in the data set, respectively;

[0101] Through the above content, the accuracy and comprehensiveness of aluminum profile thickness consistency detection are improved. By acquiring material composition data in real time and monitoring the temperature, pressure and speed during the processing through the industrial automation system, a deep understanding of the aluminum profile processing conditions is ensured. In addition, environmental sensors monitor ambient humidity, temperature and air pressure, providing important background information for the production process. The combination of these data not only helps to identify potential quality problems, but also enhances the product's quality control capabilities and market competitiveness.

[0102] Fusion algorithm applications in S32 include:

[0103] S321, feature weighted calculation: weight the processed multi-source data and the preliminary thickness distribution image data, expressed as:

[0104]

[0105] Among them, F 融合 is the eigenvalue after fusion, F i is the characteristic value of the i-th data source, n is the number of data sources, w i is the corresponding weight;

[0106] S322, fusion model construction: Based on the weighted eigenvalues, a linear regression model is used to construct a thickness measurement model, which is expressed as:

[0107] T=a·F 融合 +b;

[0108] Where T is the thickness measurement value, a is the regression coefficient, and the response F 融合 The degree of influence on the thickness T, b is a constant term, which represents the offset in the thickness measurement model;

[0109] Through the above content, the contribution of various influencing factors can be fully considered, and the regression model can be effectively mapped to the thickness prediction value to improve the accuracy and reliability of the measurement. Through a data-driven approach, the model can not only adapt to changes under different production conditions, but also output thickness prediction results in real time, which helps to achieve more accurate quality control in the production process and ensure the thickness consistency of aluminum profiles and the stability of product performance.

[0110] Microstructure feature extraction in S41 includes:

[0111] S411, Application of imaging technology: X-ray tomography is used to image the microstructure of aluminum profiles to obtain distribution images of grains, phase composition and defects;

[0112] S412, image feature extraction: Analyze the imaging data through image processing algorithms to extract microstructure features, including:

[0113] Grain size calculation: The watershed algorithm is used to segment the imaging data and separate the grain area from the background area, which is expressed as:

[0114]

[0115] Where D is the equivalent diameter of the grain, A is the area of ​​the grain, and P is the perimeter of the grain;

[0116] Phase composition analysis: Separate different phases (such as aluminum matrix phase and reinforcement phase) from the image by the area fraction of the phase composition, and calculate the area fraction V of each phase 相 , expressed as:

[0117]

[0118] Among them, A 相 is the area of ​​a specific phase, A 总 is the total area of ​​the imaging region;

[0119] Defect distribution detection: Use edge detection algorithms to identify and locate defects in the microstructure (such as holes and cracks) and calculate the defect density ρ缺陷 , expressed as:

[0120]

[0121] Among them, N 缺陷 is the total number of defects, A 总 is the total area of ​​the imaging region;

[0122] Through the above content, the microscopic characteristics of aluminum profiles such as grain size, phase composition and defect distribution can be accurately identified, providing an in-depth analysis of the internal structure of the material, so that the influence of the material microstructure on the thickness can be considered in the thickness consistency detection, which not only improves the accuracy and reliability of thickness measurement, but also helps to identify potential quality problems and ensure the performance and stability of aluminum profile products.

[0123] The thickness distribution model in S42 adopts the Gaussian process regression model, which includes:

[0124] S421, define a composite kernel function with microstructure features: In the composite kernel function, the microstructure features (grain size, phase composition ratio and defect density) are used as input features and are assigned different weights to reflect the different effects on thickness, expressed as:

[0125] k(F 微观 ,F 微观 ′)=w1·k RBF (F 微观 ,F 微观 ′)+w2·k Linear (F 微观 ,F 微观 ′);

[0126] Among them, F 微观 =[D 晶粒 ,V 相 ,ρ 缺陷 ], represents the microstructure feature vector of the current data point, F 微观 ′=[D 晶粒 ′,V 相 ′,ρ 缺陷 ′], represents the microstructure feature vector of another data point, w1 and w2 are the weight coefficients of the kernel function, reflecting the influence of different kernel functions on the thickness distribution, k RBF is the RBF kernel, used to capture nonlinear relationships, k Linear is a linear kernel, used to capture the linear relationship of data;

[0127] S422, define the weighted RBF kernel function of microstructure features: the RBF kernel is used to capture the nonlinear relationship in the microstructure features, and introduces the feature weight w in the RBF kernel i And adaptive bandwidth li , to accommodate the scale differences of different microstructural features, expressed as:

[0128]

[0129] Among them, σ 2 is the variance parameter of the kernel function, d is the dimension of the microstructure characteristics (3: grain size, phase composition ratio, defect density), F 微观,i 、F 微观,i ′ is the i-th microstructure eigenvalue, w i is the weight of the i-th microstructure feature, reflecting the influence of this feature on the thickness distribution, l i is the adaptive bandwidth parameter of the i-th microstructure feature, which is used to adjust the scale of different features;

[0130] S423, define the weighted linear kernel function of microstructure features: the linear kernel is used to capture the linear relationship in the microstructure features, and at the same time add feature weights to adapt to the influence of different features on the thickness distribution, expressed as:

[0131]

[0132] Among them, v i is the weight of the i-th microstructural feature, reflecting the influence of this feature in the linear relationship;

[0133] S424, construct a thickness distribution prediction model based on microstructure characteristics: output the predicted thickness distribution value according to the defined composite kernel function It is expressed as:

[0134]

[0135] in, is the predicted thickness value, m(F 微观,i ) is the mean function of the region, α j is the model parameter, k(F 微观,i ,F 微观,j ) is a kernel function used to represent the similarity of the microstructural features of the i-th and j-th regions, and u is the number of reference data points used to predict the thickness of the i-th region;

[0136] Through the above content, the complex pattern of aluminum profile thickness distribution can be captured more accurately. The introduced composite kernel function not only flexibly adapts to the nonlinear influence of different features on thickness, but also reflects the importance of each feature through weight and adaptive bandwidth, thereby greatly improving the accuracy and stability of thickness prediction, helping to identify potential thickness change trends and ensuring the reliability and consistency of final product quality.

[0137] Potential inconsistencies identified in S43 include:

[0138] S431, predicted thickness calculation: Use the thickness distribution model to predict the thickness of different areas of the aluminum profile and obtain the predicted thickness of each detection area

[0139] S432, Standard Thickness Comparison: The predicted thickness of each area With the output of the integrated thickness measurement model T i Compare and calculate thickness deviation ΔT i , expressed as:

[0140]

[0141] S4323, inconsistency identification: by setting the thickness deviation tolerance threshold ∈ and the thickness deviation ΔT i For comparison, if the thickness deviation ΔT i Exceeding the thickness deviation tolerance threshold ∈ indicates that there is a potential thickness inconsistency problem in the area;

[0142] The thickness deviation tolerance threshold ∈ is set based on the statistical analysis of historical data, including:

[0143] Collect data: Collect the deviation data ΔT between the thickness measurement value and the actual standard thickness from the historical inspection data in the production process;

[0144] Calculate the mean and standard deviation: Perform statistical analysis on the deviation data ΔT and calculate the mean μ of the deviation ΔT and standard deviation σ ΔT , expressed as:

[0145]

[0146] Where N is the number of samples, ΔT i is the i-th historical deviation data;

[0147] Set the tolerance threshold: Set the thickness deviation tolerance threshold ∈ to a statistical range of the historical deviation distribution, expressed as:

[0148] ∈=μ ΔT +k·σ ΔT ;

[0149] Among them, k is set to 2 or 3, which is adjusted according to the strictness of the quality requirements. k = 2 corresponds to about 95% of the data within the tolerance range, and k = 3 corresponds to about 99.7%;

[0150] Through the above content, the thickness of different areas can be accurately predicted and compared with the standard thickness value, so as to effectively identify the areas where the thickness deviation exceeds the tolerance range. By setting the tolerance threshold, the potential thickness inconsistency problem can be clearly identified, which not only improves the sensitivity and accuracy of thickness detection, but also reduces the possibility of false alarms and missed alarms, which is helpful for quality control in the production process and ensures the thickness consistency of aluminum profiles and the reliability of product performance.

[0151] Thickness consistency assessment in S5 includes:

[0152] S51, overall thickness deviation calculation: based on the calculated thickness deviation ΔT i , calculate the root mean square error RMSE of the overall thickness deviation to evaluate the overall thickness consistency level of the aluminum profile, expressed as:

[0153]

[0154] Where N is the total number of detection areas;

[0155] S52, thickness consistency report generation: Generate a thickness consistency evaluation report based on the RMSE result of the overall thickness deviation. The report includes the overall thickness consistency evaluation (RMSE value) and the conclusion on whether it meets the production quality standards;

[0156] Through the above content, not only can the overall thickness deviation be quantified and clear thickness consistency evaluation indicators be provided, but also a detailed report can be generated to determine whether the production quality standards are met. The objective evaluation of thickness consistency through the RMSE value can timely identify potential quality problems, provide a scientific basis for quality control and continuous improvement of the production process, and help ensure product reliability and consistency.

[0157] The abnormal cause tracing analysis in S6 includes:

[0158] S61, Historical data collection: Collect historical production data related to thickness inconsistency, including material batches, process parameters (such as processing temperature, pressure, speed, etc.) and quality inspection records;

[0159] S62, correlation analysis: Correlation analysis is used to calculate the correlation coefficient r between each feature and thickness inconsistency, expressed as:

[0160]

[0161] Where N is the number of samples, X i is the process parameter value of the i-th sample, such as temperature, pressure and other specific production parameters, is the mean of variable X, Y iis the thickness deviation data corresponding to the i-th sample, indicating the thickness inconsistency measured under the production conditions. is the mean of variable Y;

[0162] S63, abnormal pattern recognition: K-means clustering algorithm is used to cluster abnormal data, and samples with similar abnormal characteristics are divided into different categories. The abnormal pattern is identified by minimizing the square error J, which is expressed as:

[0163]

[0164] Among them, K is the number of clusters, C k is the kth cluster, μ k is the centroid of the cluster;

[0165] S64, generation of traceability analysis report: based on the results of correlation analysis and abnormal pattern recognition, a traceability report of abnormal causes is generated, which includes factors affecting thickness consistency, abnormal patterns and their corresponding process conditions;

[0166] Through the above content, the correlation between process parameters and thickness deviation can be quantified, and typical abnormal patterns can be summarized through clustering algorithms, making the causes of inconsistency problems more intuitive and clear. The generated traceability analysis report provides key influencing factors, providing a scientific guidance basis for quality control and process optimization, and helping to improve the overall quality stability and production efficiency of products.

[0167] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0168] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning, characterized in that: The following steps are involved: S1, aluminum profile pretreatment: pretreatment of the aluminum profile to be tested, including cleaning and removal of surface impurities, oil stains or substances that affect X-ray imaging; S2, X-ray full-section scanning: Use X-ray equipment to scan the full section of the pre-treated aluminum profile to obtain preliminary thickness distribution image data; S3, multi-source data fusion: Based on the obtained thickness distribution image data, a multi-source data fusion algorithm is applied to integrate it with the material composition data, production process parameters, and environmental condition data to generate a comprehensive thickness measurement model; S4, thickness prediction analysis: By analyzing the microstructural characteristics of aluminum profiles, the thickness distribution of different areas is predicted and potential thickness inconsistency issues are identified, including: S41, microstructure feature extraction: use imaging technology to analyze aluminum profiles and extract the microstructure features of aluminum profiles, including grain size, phase composition and defect distribution; S42, thickness distribution model construction: construct a thickness distribution model based on the extracted microstructure features; S43, potential inconsistency identification: using the thickness distribution model to predict the predicted thickness of different areas and compare it with the output of the comprehensive thickness measurement model to identify potential thickness inconsistency problems; S5, thickness consistency assessment: Based on the results of thickness distribution in different areas, the thickness consistency of aluminum profiles is assessed and a test report is generated; S6, abnormal cause tracing analysis: for the identified thickness inconsistency problem, abnormal cause tracing analysis is carried out, and historical production records, material batches and process parameters are analyzed using data mining technology to identify potential root causes that affect thickness consistency.

2. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 1, characterized in that: The X-ray full section scan in S2 includes: S21, equipment setting: according to the size and characteristics of the aluminum profile, adjust the parameters of the X-ray equipment, including the ray intensity, scanning speed and imaging resolution; S22, sample fixing: mounting the pretreated aluminum profile on the X-ray scanning platform; S23, full-section scanning: start the X-ray equipment, perform a full-section scan on the aluminum profile, record the attenuation data after the X-ray penetrates the aluminum profile, and generate preliminary thickness distribution image data.

3. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 1, characterized in that: The multi-source data fusion in S3 includes: S31, multi-source data acquisition and processing: collect multi-source data of aluminum profiles, including material composition data, production process parameters, and environmental condition data, and process the multi-source data and preliminary thickness distribution image data, including data cleaning, denoising, and normalization; S32, fusion algorithm application: Apply the multi-source data fusion algorithm to integrate the processed multi-source data and the preliminary thickness distribution image data to generate a comprehensive thickness measurement model.

4. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 3 is characterized in that: The multi-source data collection and processing in S31 includes: S311, Multi-source data collection: Collect multi-source data of aluminum profiles through sensors and data recording units, including: Material composition data collection: Use spectral analysis technology to obtain the content of each component; Production process parameter collection: Real-time monitoring of the processing temperature, extrusion pressure and conveying speed of aluminum profiles during processing through industrial automation systems; Environmental condition data collection: Environmental sensors are used to monitor the ambient humidity, ambient temperature and ambient air pressure in the aluminum profile production environment; S312, data processing: processing the collected multi-source data and preliminary thickness distribution image data, specifically including: Data cleaning: Use Z-score standardization to clean multi-source data and preliminary thickness distribution image data; Denoising: Use moving average to reduce random noise in the data; Normalization: Use min-max normalization to normalize the data to the same range.

5. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 4, characterized in that: The fusion algorithm application in S32 includes: S321, feature weighted calculation: weighting the processed multi-source data and preliminary thickness distribution image data; S322, fusion model construction: based on the weighted eigenvalues, a thickness measurement model is constructed using a linear regression model.

6. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 1, characterized in that: The microstructure feature extraction in S41 includes: S411, Application of imaging technology: X-ray tomography is used to image the microstructure of aluminum profiles to obtain distribution images of grains, phase composition and defects; S412, image feature extraction: Analyze the imaging data through image processing algorithms to extract microstructure features, including: Grain size calculation: The imaging data is segmented using the watershed algorithm to separate the grain area from the background area; Phase composition analysis: Separate different phases from the image by the area fraction of the phase composition and calculate the area fraction V of each phase 相 ; Defect distribution detection: Use edge detection algorithms to identify and locate defects in the microstructure and calculate the defect density ρ 缺陷 .

7. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 6, characterized in that: The thickness distribution model in S42 adopts a Gaussian process regression model, and the Gaussian process regression model includes: S421, defining a composite kernel function with microstructure features: in the composite kernel function, the microstructure features are used as input features and are assigned different weights to reflect different effects on thickness; S422, define the weighted RBF kernel function of microstructure features: the RBF kernel is used to capture the nonlinear relationship in the microstructure features, and introduces the feature weight w in the RBF kernel i And adaptive bandwidth l i , to accommodate the scale differences of different microstructural features; S423, defining a weighted linear kernel function of microstructural features: the linear kernel is used to capture the linear relationship in the microstructural features, and feature weights are added to accommodate the impact of different features on the thickness distribution; S424, construct a thickness distribution prediction model based on microstructure characteristics: output the predicted thickness distribution value according to the defined composite kernel function 8. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 7, characterized in that: The potential inconsistency identification in S43 includes: S431, predicted thickness calculation: Use the thickness distribution model to predict the thickness of different areas of the aluminum profile and obtain the predicted thickness of each detection area S432, Standard Thickness Comparison: The predicted thickness of each area With the output of the integrated thickness measurement model T i Compare and calculate thickness deviation ΔT i ; S4323, inconsistency identification: by setting the thickness deviation tolerance threshold ∈ and the thickness deviation ΔT i For comparison, if the thickness deviation ΔT i Exceeding the thickness deviation tolerance threshold ∈ indicates that there is a potential thickness inconsistency problem in the area.

9. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 8, characterized in that: The thickness consistency assessment in S5 includes: S51, overall thickness deviation calculation: based on the calculated thickness deviation ΔT i , calculate the root mean square error RMSE of the overall thickness deviation to evaluate the overall thickness consistency level of the aluminum profile; S52, thickness consistency report generation: Generate a thickness consistency evaluation report based on the RMSE result of the overall thickness deviation. The report includes the overall thickness consistency evaluation and the conclusion on whether it meets the production quality standards.

10. The method for detecting the consistency of aluminum profile cross-section thickness based on X-ray scanning according to claim 1, characterized in that: The abnormal cause tracing analysis in S6 includes: S61, Historical Data Collection: Collect historical production data related to thickness inconsistency, including material batches, process parameters and quality inspection records; S62, correlation analysis: using correlation analysis to calculate the correlation coefficient r between each feature and thickness inconsistency; S63, abnormal pattern recognition: K-means clustering algorithm is used to cluster abnormal data, samples with similar abnormal characteristics are divided into different categories, and abnormal patterns are identified by minimizing the square error J; S64, generation of traceability analysis report: based on the results of correlation analysis and abnormal pattern recognition, a traceability report of abnormal causes is generated, which includes factors affecting thickness consistency, abnormal patterns and their corresponding process conditions.

Images