CN-119360091-B - Screw tightening abnormality identification method, device, equipment and storage medium

CN119360091BCN 119360091 BCN119360091 BCN 119360091BCN-119360091-B

Abstract

The application discloses a screw tightening abnormality identification method, a device, equipment and a storage medium, and belongs to the field of abnormality detection. The method comprises the steps of obtaining tightening data of a screw to be tested in a tightening process, removing abnormal data from the tightening data, determining the data to be tested, carrying out feature extraction on the data to be tested, determining feature data, inputting the feature data into an abnormal recognition model, obtaining a recognition result of the screw to be tested, wherein the recognition result is used for representing whether the screw to be tested is abnormal in the tightening process, and the abnormal recognition model is obtained based on the feature data of a sample screw, a marking result corresponding to the feature data of the sample screw and class training of the sample screw. According to the application, the screw tightening process can be monitored on line through the abnormality identification model, so that the abnormality in the tightening process can be found in time, the effective monitoring of the abnormality in the tightening process is realized, and the influence of screw tightening on the quality of a battery is reduced.

Inventors

GAO YAN
LI ZHENDONG
JI FANG
CHEN JINQIANG

Assignees

欣旺达动力科技股份有限公司

Dates

Publication Date: 20260508
Application Date: 20240929

Claims (10)

1. A method of identifying a screw tightening abnormality, the method comprising: Obtaining tightening data of a screw to be tested in a tightening process; removing abnormal data from the tightening data, and determining data to be tested; extracting features of the data to be detected, and determining feature data; Inputting the characteristic data into an abnormal recognition model, and obtaining a recognition result of the screw to be detected, wherein the recognition result is used for representing whether the screw to be detected is abnormal in the screwing process, and the abnormal recognition model is obtained based on the characteristic data of a sample screw, a marking result corresponding to the characteristic data of the sample screw and the class training of the sample screw; Wherein the category of the sample screw is determined based on tightening data of the sample screw, comprising: Based on the torsion data of the sample screw, determining a torsion mean value, a torsion maximum value, torsion distribution kurtosis and torsion distribution skewness corresponding to the sample screw; determining a first outlier sample screw with abnormal torque mean and/or abnormal torque maximum from all the sample screws based on a triple standard deviation method; Determining second outlier sample screws with abnormal torsion distribution kurtosis and/or abnormal torsion distribution skewness from all the sample screws based on a cluster analysis method; The first and second outlier sample screws are determined to be the negative sample screws, and the remaining sample screws of all sample screws except the first and second outlier sample screws are determined to be the positive sample screws.
2. The method for identifying abnormal screw tightening according to claim 1, wherein the method further comprises, before the obtaining of the tightening data of the screw to be measured in the tightening process: building a framework of the anomaly identification model based on a CART decision tree; Obtaining tightening data of a sample screw; determining a category of the sample screw based on the tightening data of the sample screw, the category of the sample screw including a negative sample screw and a positive sample screw; performing feature extraction and abnormal marking on the tightening data of the sample screw to obtain feature data of the sample screw and a marking result corresponding to the feature data; Training and verifying the abnormal recognition model based on the characteristic data of the sample screw, the marking result corresponding to the characteristic data and the category of the sample screw, and obtaining the trained abnormal recognition model.
3. The method for identifying abnormal screw tightening according to claim 2, wherein the tightening data includes torsion data and angle data, the feature extraction and the abnormal marking are performed on the tightening data of the sample screw to obtain feature data of the sample screw and a marking result corresponding to the feature data, and the method comprises the following steps: extracting feature data of the sample screw based on the torsion data and the angle data of the sample screw, the feature data including a plurality of feature values; and carrying out abnormal marking on each characteristic value, and determining a marking result corresponding to each characteristic value, wherein the marking result comprises normal and abnormal.
4. The screw tightening abnormality identification method according to claim 3, wherein the tightening process includes a screwing stage and a tightening stage, and the plurality of characteristic values include a maximum torque value of the tightening process, a tightening angle difference value of the tightening process, a climbing angle value of the tightening process and a maximum torque value of the screwing stage, a torque average value of the screwing stage, and a torque distribution kurtosis of the tightening stage; The step of carrying out abnormal marking on each characteristic value and determining a marking result corresponding to each characteristic value comprises the following steps: Determining a marking result corresponding to each characteristic value in the maximum torque value in the tightening process, the tightening angle difference in the tightening process and the climbing angle value in the tightening process based on comparison results of the maximum torque value in the tightening process, the tightening angle difference in the tightening process and the climbing angle value in the tightening process with corresponding preset thresholds respectively; Determining a marking result corresponding to each characteristic value in the maximum torsion value of all the screwing stages and the torsion average value of the screwing stages by a triple standard deviation method based on the maximum torsion value of all the screwing stages and the torsion average value of the screwing stages corresponding to all the sample screws; and determining a marking result corresponding to each characteristic value in the torque distribution kurtosis of all the fastening stages by a cluster analysis method based on the torque distribution kurtosis of all the fastening stages.
5. The screw tightening abnormality identification method according to claim 2, characterized in that before training and verifying the abnormality identification model based on the characteristic data of the sample screw, the marking result corresponding to the characteristic data, and the type of the sample screw, the method further comprises: And based on the characteristic data of the negative sample screw and the marking result corresponding to the characteristic data, oversampling the negative sample screw to obtain the characteristic data of a new negative sample screw and the marking result corresponding to the characteristic data, wherein the deviation between the sum of the number of the negative sample screw and the number of the new negative sample screw and the number of the positive sample screw meets a preset condition.
6. The screw tightening abnormality identification method according to claim 2, characterized in that before determining the category of the sample screw based on the tightening data of the sample screw, the method further comprises: and performing data cleaning on the tightening data of the sample screw to remove an abnormal sample from the sample screw.
7. The screw tightening abnormality identification method according to claim 2, wherein the characteristic data includes a plurality of characteristic values, the method further comprising: If the identification result indicates that the screw to be tested is abnormal in the screwing process, acquiring decision path information of the characteristic data in the abnormal identification model; And determining the characteristic value with the marked result being abnormal in the characteristic data based on the decision path information.
8. A screw tightening abnormality identifying device, characterized by comprising: the data acquisition module is used for acquiring tightening data of the screw to be tested in the tightening process; the data cleaning module is used for removing abnormal data from the tightening data and determining data to be tested; The feature extraction module is used for carrying out feature extraction on the data to be detected and determining feature data; The abnormal recognition module is used for inputting the characteristic data into an abnormal recognition model, acquiring a recognition result of the screw to be detected, wherein the recognition result is used for representing whether the screw to be detected is abnormal in the screwing process, and the abnormal recognition model is obtained by training based on the characteristic data of a sample screw, a marking result corresponding to the characteristic data of the sample screw and the type of the sample screw; a model training module for determining a class of the sample screw based on tightening data of the sample screw, comprising: Based on the torsion data of the sample screw, determining a torsion mean value, a torsion maximum value, torsion distribution kurtosis and torsion distribution skewness corresponding to the sample screw; determining a first outlier sample screw with abnormal torque mean and/or abnormal torque maximum from all the sample screws based on a triple standard deviation method; Determining second outlier sample screws with abnormal torsion distribution kurtosis and/or abnormal torsion distribution skewness from all the sample screws based on a cluster analysis method; The first and second outlier sample screws are determined to be the negative sample screws, and the remaining sample screws of all sample screws except the first and second outlier sample screws are determined to be the positive sample screws.
9. An electronic device comprising a memory and a processor; the memory having stored thereon a computer program executable by the processor, the computer program, when executed by the processor, performing the method of any of claims 1-1.
10. A computer-readable storage medium comprising, Stored in the computer readable storage medium are computer program instructions which, when executed by a processor, perform the method according to any one of claims 1-7.

Description

Screw tightening abnormality identification method, device, equipment and storage medium Technical Field The application relates to the technical field of abnormality detection, in particular to a screw tightening abnormality identification method, a device, equipment and a storage medium. Background Screws are one of the most common parts in industry, and the tightening quality of each screw directly affects the quality of the product. In the battery manufacturing industry, the screw is also one of important parts for completing the packaging of each group of batteries, has important significance for ensuring the normal charge and discharge of the batteries, safe and reliable operation and the like, and can cause serious safety accidents if the screw tightening quality is problematic. The conventional screw tightening abnormality identification method mainly comprises the step of judging by detecting the tightening torque and the tightening angle after the screw tightening is finished. However, the method can cause the abnormality (such as screw skew, screw sliding, etc.) generated in the tightening process to be unrecognized, thereby affecting the product quality. Disclosure of Invention The embodiment of the application provides a screw tightening abnormality identification method, device, equipment and storage medium, which are used for effectively monitoring abnormality generated in the screw tightening process and reducing the influence of screw tightening on the quality of a battery. In order to solve the technical problems, the embodiment of the application discloses the following technical scheme: in a first aspect, there is provided a screw tightening abnormality identification method, the method including: Obtaining tightening data of a screw to be tested in a tightening process; removing abnormal data from the tightening data, and determining data to be tested; extracting features of the data to be detected, and determining feature data; inputting the characteristic data into an abnormal recognition model to obtain a recognition result of the screw to be tested, wherein the recognition result is used for representing whether the screw to be tested is abnormal in the screwing process, and the abnormal recognition model is obtained based on the characteristic data of the sample screw, a marking result corresponding to the characteristic data of the sample screw and the class training of the sample screw. In some embodiments, the method further comprises, prior to the obtaining of the tightening data of the screw to be measured in the tightening process: building a framework of the anomaly identification model based on a CART decision tree; Obtaining tightening data of a sample screw; determining a category of the sample screw based on the tightening data of the sample screw, the category of the sample screw including a negative sample screw and a positive sample screw; performing feature extraction and abnormal marking on the tightening data of the sample screw to obtain feature data of the sample screw and a marking result corresponding to the feature data; Training and verifying the abnormal recognition model based on the characteristic data of the sample screw, the marking result corresponding to the characteristic data and the category of the sample screw, and obtaining the trained abnormal recognition model. In some embodiments, the tightening data includes torque data, and the determining a category of the sample screw based on the tightening data of the sample screw includes: Determining a torsion mean value, a torsion maximum value, torsion distribution kurtosis and torsion distribution skewness corresponding to the sample screw based on the torsion data of the sample screw; determining a first outlier sample screw with abnormal torque mean and/or abnormal torque maximum from all the sample screws based on a triple standard deviation method; Determining second outlier sample screws with abnormal torsion distribution kurtosis and/or abnormal torsion distribution skewness from all the sample screws based on a cluster analysis method; The first and second outlier sample screws are determined to be the negative sample screws, and the remaining sample screws of all sample screws except the first and second outlier sample screws are determined to be the positive sample screws. In some embodiments, the tightening data includes torsion data and angle data, and the feature extraction and the abnormal marking are performed on the tightening data of the sample screw to obtain feature data of the sample screw and a marking result corresponding to the feature data, including: extracting feature data of the sample screw based on the torsion data and the angle data of the sample screw, the feature data including a plurality of feature values; and carrying out abnormal marking on each characteristic value, and determining a marking result corresponding to each characteristic value, wherein the marking res