A tightening quality early warning method based on intelligent analysis of tightening curve

Abstract

This invention relates to a tightening quality early warning method based on intelligent analysis of tightening curves. The method includes: collecting a large number of tightening curves from the production line to establish a defect-tightening curve feature database; analyzing the collected curves: real-time acquisition of production line tightening data; extracting feature values ​​from the collected tightening curves and classifying them; calibrating defects in the classified curves and outputting defect types; data storage; statistical analysis of production results and issuing early warnings for defects; comprehensive summary and analysis of the stored data; prediction of production line tightening trends and early warning of tightening defects. This invention establishes different identification ranges and methods for defect curve feature values ​​for each defect type, ensuring the accuracy of defect identification. By acquiring and analyzing tightening curves in real time, it can significantly improve production efficiency, reduce labor costs, and accurately control production quality. This invention also stores the analyzed data in real time and performs trend judgment and defect early warning.

Description

Technical Field

[0001] This invention belongs to the field of automotive assembly technology, specifically relating to a tightening quality early warning method based on intelligent analysis of tightening curves. Background Technology

[0002] With the advancement of intelligent manufacturing, "intelligent manufacturing" and "smart factories" have become major transformation goals for manufacturing enterprises. Replacing manual operations with intelligent methods can significantly reduce the frequency of human intervention in the production process, thereby greatly improving production efficiency and driving product quality upgrades.

[0003] Currently, in automobile production and assembly, bolt tightening curves can be collected. However, these curves are still analyzed manually, which is inefficient and the accuracy of the results relies heavily on the operator's experience, introducing significant subjectivity. Therefore, there is an urgent need for an online monitoring and analysis technology for bolt tightening curves. Summary of the Invention

[0004] The purpose of this invention is to provide a tightening quality early warning method based on intelligent analysis of tightening curves, in order to solve the problem that bolt tightening curves in automobile production and assembly rely on manual analysis and lack online monitoring and analysis.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A tightening quality early warning method based on intelligent analysis of tightening curves includes the following steps:

[0007] A. Collect a large amount of data on the tightening curves of the production line and establish a defect-tightening curve feature database;

[0008] B. Analyze the collected curves:

[0009] B1. Real-time acquisition of production line tightening data;

[0010] B2. Extract feature values ​​from the collected tightening curves and classify them;

[0011] B3. Perform defect calibration on the classified curves and output the defect type;

[0012] B4. Data storage;

[0013] C. Analyze production results and issue early warnings for defects:

[0014] C1. Conduct a comprehensive summary and analysis of the stored data;

[0015] C2. Predict the tightening trend of the production line and provide early warning of tightening defects;

[0016] D. Repeat steps B1-B4, and repeat steps C1-C2 according to production needs.

[0017] Further, step A specifically involves: collecting tightening data output from the production line, comprehensively considering the currently known defect types, manually classifying and calibrating the collected data based on production experience, establishing a unique feature value extraction range and method for each type of defect, obtaining a one-to-one correspondence between defect and tightening curve, obtaining a tightening qualified feature curve, and establishing a defect-tightening curve feature database.

[0018] Further, in step B1, the collected data includes key tightening information such as production model, connection structure, connection length, tightening speed, connection status, connection position, and tightening curve.

[0019] Further, step B2 specifically involves: matching the collected data with the database curves in real time using a feature value matching algorithm; firstly, determining whether the collected tightening curve matches the qualified tightening curves in the feature database; directly storing the qualified tightening curves; and classifying the unqualified tightening curves by defect type and distribution trend.

[0020] Furthermore, they are divided into three categories: high correlation between defect and tightening curve, low correlation between defect and tightening curve, and no correlation between defect and tightening curve.

[0021] Further, step B3 specifically includes:

[0022] B31. Determine whether there is a characteristic value correlation between the tightening curve and the defect. If there is no correlation between the tightening curve and the defect, perform manual calibration. If there is a high correlation between the tightening curve and the defect, the defect type can be directly determined through the correlation, and the defect calibration can be performed directly.

[0023] B32. If the correlation between the tightening curve and the defect is low, and the defect type cannot be directly determined by the correlation, then the curve is subjected to secondary feature value extraction. The extracted feature values ​​are then compared with the feature curves in the database again. If the defect type can be determined by the extracted feature values, then the defect is calibrated. If the defect type cannot be determined by the extracted feature values, then the defect is manually calibrated.

[0024] B33. Manually calibrate tightening curves that are not correlated with defects and tightening curves and whose defect types cannot be determined by secondary feature values. After calibration, determine whether the tightening curves are valid data based on the actual defect types on the production line. Remove invalid data and add valid data to the defect-tightening curve feature database. Provide real-time alarms for tightening positions where defects occur and predict possible future failure modes based on defect types.

[0025] Further, step B4 specifically involves storing the analyzed data according to various classification methods.

[0026] Further, step C1 specifically involves summarizing the collected data to obtain key data files such as the trend of vehicle tightening pass rate, the trend of workstation pass rate, and defect type statistics.

[0027] Further, step C2 specifically involves: predicting the tightening pass rate of the production line, predicting the frequency of defects, issuing early warnings for workstations prone to defects, and predicting the types of defects that are likely to occur.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] This invention establishes different defect curve feature value identification ranges and methods for each defect type through feature value extraction, ensuring the accuracy of defect identification in terms of technology. By collecting and analyzing tightening curves in real time, it can significantly improve production efficiency, reduce labor costs, and accurately control production quality. This invention also stores the analyzed data in real time, performs trend judgment and defect early warning, can record production data, and provide guidance for production line improvement. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 Tightening quality early warning flowchart based on intelligent analysis of tightening curve;

[0032] Figure 2 Tighten the data analysis and processing flowchart. Detailed Implementation

[0033] The present invention will be further described below with reference to embodiments:

[0034] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0035] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0036] This invention establishes a defect-tightening curve feature database, quickly matches the collected data with the database, accurately determines the defect type corresponding to the non-conforming curve, and provides real-time alarms for defects. The tightening quality early warning technology based on intelligent tightening curve analysis can effectively improve the efficiency of tightening curve analysis, ensure the accuracy and repeatability of the analysis process, predict the tightening defect trend on the production line, provide timely early warnings for tightening defects, ensure the pass rate of tightening on the production line, and improve the tightening quality and reliability of the final assembly tightening process.

[0037] This invention provides a tightening quality early warning method based on intelligent analysis of tightening curves, comprising the following steps:

[0038] 1. Collect a large number of tightening curves from the production line and establish a defect-tightening curve feature database;

[0039] 2. Real-time collection of tightening data from the production line;

[0040] 3. Extract feature values ​​from the collected tightening curves and classify them;

[0041] 4. Perform defect calibration on the classified curves and output the defect type;

[0042] 5. Store the analyzed data using a multi-classification method;

[0043] 6. Conduct a comprehensive summary and analysis of the stored data;

[0044] 7. Predict the tightening trend of the production line and provide early warning of tightening defects;

[0045] 8. Repeat steps 2-5, and repeat steps 6-7 according to production needs.

[0046] This invention is applicable to scenarios involving automatic analysis and processing of tightening data. The analyzed data can quickly match tightening defect types and provide trend judgments and defect warnings for production lines. This intelligent tightening curve analysis-based tightening quality warning method mainly consists of three parts. For example... Figure 1 As shown, the steps are: first, to establish a defect-tightening curve feature database; second, to analyze the collected curves; and third, to statistically analyze the production results and issue early warnings for defects.

[0047] Part 1: Establishment of the Defect-Tightening Curve Feature Database: A large amount of tightening data output from the production line is collected. Based on the currently known defect types and production experience, the collected data is manually classified and calibrated. A unique feature value extraction range and method are established for each type of defect. The one-to-one correspondence between defect and tightening curve is obtained, and a tightening qualified feature curve is obtained, thus establishing the defect-tightening curve feature database.

[0048] Part Two: Curve Data Acquisition and Analysis, the process is as follows: Figure 2 As shown.

[0049] Data Acquisition: The acquired data is transmitted to the data analysis system in real time. The acquired data includes key tightening information such as production vehicle model, connection structure, connection length, tightening speed, connection status, connection position, and tightening curve.

[0050] Data screening: The collected data is matched with the database curves in real time using the feature value matching algorithm. First, it is determined whether the collected tightening curves match the qualified tightening curves in the feature database. The qualified tightening curves are directly stored in the data. Among the unqualified tightening curves, they are divided into three categories based on the defect type and distribution trend: high correlation between defect and tightening curve, low correlation between defect and tightening curve, and no correlation between defect and tightening curve.

[0051] Data processing: First, determine whether there is a correlation between the unqualified tightening curve and the defect in terms of characteristic values. If there is no correlation between the tightening curve and the defect, manual calibration is performed. If there is a high correlation between the tightening curve and the defect, the defect type can be directly determined through the correlation, and defect calibration is performed directly.

[0052] If the tightening curve has a low correlation with the defect and the defect type cannot be directly determined by the correlation, then the curve is subjected to secondary feature value extraction. The extracted feature values ​​are then compared with the feature curves in the database again. If the defect type can be determined by the extracted feature values, then the defect is calibrated. If the defect type cannot be determined by the extracted feature values, then the defect is manually calibrated.

[0053] For tightening curves where there is no correlation between defects and tightening curves, and where the defect type cannot be determined by secondary feature value extraction, manual calibration is performed. After calibration, based on the actual defect type on the production line, it is determined whether the tightening curve is valid data. Invalid data is removed, and valid data is added to the defect-tightening curve feature database. Real-time alarms are set for tightening positions where defects occur, and future failure modes are predicted based on the defect type.

[0054] Data storage: The analyzed data is stored according to various classification methods to enable comprehensive analysis and prediction of the production line.

[0055] The third part involves statistical analysis of production results and issuing early warnings for defects.

[0056] Results Summary: The collected data were summarized to obtain key data files such as the trend of vehicle tightening pass rate, the trend of workstation pass rate, and defect type statistics.

[0057] Production line early warning: Predicts the tightening pass rate of the production line, predicts the frequency of defects, provides early warnings for workstations prone to defects, and provides early warnings for the types of defects that are likely to occur.

[0058] This invention provides an online monitoring and analysis technology for tightening curves, which can monitor tightening quality in real time and issue alarms for defects. Different curve feature value extraction methods and ranges are established for each type of defect to ensure accurate matching of the collected curves. Through the established multi-data classification method, the production line's production status can be analyzed comprehensively, providing a basis for production line upgrades and modifications.

[0059] Other methods for analyzing tightening curves by extracting feature values ​​once, twice, or multiple times, as well as methods for predicting tightening trends on production lines by summarizing the analysis results, are all within the scope of this invention.

[0060] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A tightening quality early warning method based on intelligent analysis of tightening curves, characterized in that, Includes the following steps: A. Collect a large amount of data on the tightening curves of the production line and establish a defect-tightening curve feature database; Collect tightening data from the production line, combine the currently known defect types, manually classify and calibrate the collected data based on production experience, establish a unique feature value extraction range and method for each defect, obtain a one-to-one correspondence between defect and tightening curve, obtain a tightening qualified feature curve, and establish a defect-tightening curve feature database. B. Analyze the collected curves: B1. Real-time acquisition of production line tightening data; the acquired data includes production vehicle model, connection structure, connection length, tightening speed, connection status, connection position, and tightening curve; B2. Extract feature values ​​from the collected tightening curves and classify them; The collected data is matched with the database curves in real time using the feature value matching algorithm. First, it is determined whether the collected tightening curve matches the qualified tightening curves in the feature database. The qualified tightening curves are directly stored in the data. Among the unqualified tightening curves, they are classified according to the defect type and distribution trend. They are divided into three categories: high correlation between defect and tightening curve, low correlation between defect and tightening curve, and no correlation between defect and tightening curve. B3. Perform defect calibration on the classified curves and output the defect type; B31. Determine whether there is a correlation between the curve of unqualified tightening and the characteristic value of the defect. If there is no correlation between the defect-tightening curve, then perform manual calibration. If the defect-tightening curve has a high correlation, then defect calibration can be performed directly. B32. If the defect-tightening curve has a low correlation, then perform a second feature value extraction on the curve, and compare the extracted feature values ​​with the feature curve in the database again. If the defect type can be determined by the extracted feature values, then perform defect calibration. If the defect type cannot be determined by the feature values ​​extracted twice, then the defect is manually labeled. B33. Manually calibrate tightening curves that are not correlated with defects and tightening curves and whose defect types cannot be determined by secondary feature values. After calibration, determine whether the tightening curves are valid data based on the actual defect types on the production line. Remove invalid data and add valid data to the defect-tightening curve feature database. Provide real-time alarms for tightening positions where defects occur and predict possible future failure modes based on defect types. B4. Data storage; The analyzed data will be stored according to various classification methods. C. Analyze production results and issue early warnings for defects: C1. Conduct a comprehensive summary and analysis of the stored data; summarize the collected data to obtain data files including the trend of vehicle tightening pass rate, the trend of workstation pass rate, and defect type statistics; C2. Predict the tightening trend of the production line and provide early warning of tightening defects; Predict the tightening pass rate of the production line, predict the frequency of defects, provide early warnings for workstations prone to defects, and provide early warnings for the types of defects that are prone to occur. D. Repeat steps B1-B4, and repeat steps C1-C2 according to production needs.

Images