Resistance spot welding quality on-line monitoring method based on ultrasonic guided waves

Through ultrasonic waveguide signal acquisition and support vector machine model, real-time quality monitoring of the resistance spot welding process is achieved, and the problem that the existing technology cannot comprehensively and accurately evaluate welding quality is solved, and detection efficiency and accuracy are improved.

CN119077100APending Publication Date: 2024-12-06JILIN UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411464759.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-21
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The prior art cannot comprehensively and accurately evaluate the welding quality of resistive spot welding, especially in the welding process, it is difficult to achieve online monitoring and real-time control of spot welding quality.

Method used

The online monitoring method of resistance spot welding quality based on ultrasonic guides is adopted. By collecting ultrasonic guide signals, extracting characteristic values, reducing dimensionality using principal component analysis method, and using principal component variables as input samples of support vector machine, spot welding melting is constructed Nuclear diameter regression prediction model realizes real-time monitoring of the point welding process.

Benefits of technology

Real-time monitoring of all welding points during the point welding process is achieved, the accuracy and reliability of the monitoring results are ensured, and the detection efficiency is improved, and time and cost savings are achieved compared with non-destructive testing after welding.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119077100A_ABST
    Figure CN119077100A_ABST
Patent Text Reader

Abstract

The invention relates to a resistance spot welding quality online monitoring method based on ultrasonic guided waves, and belongs to the field of resistance spot welding quality nondestructive testing. The method comprises the following steps: firstly, processing an initial guided wave signal to obtain a characteristic quantity related to the nugget size, then carrying out normalization processing on related data, then carrying out dimensionality reduction through a principal component analysis method to obtain k principal component variables, and then adopting a cross validation method to obtain optimal parameters of a model so as to obtain a function model. And finally, k principal component indexes of support vector machine parameters obtained through training are input into the model, and the spot welding nugget diameter size of the next sampling point is obtained. The spot welding nugget on-line monitoring device realizes on-line monitoring of the spot welding nugget, and ensures the quality of the spot welding joint. The calculation precision is greatly improved, the time cost is saved, and the detection efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of nondestructive testing of resistance spot welding quality, and in particular to an online monitoring method for resistance spot welding quality based on ultrasonic guided waves. Background Art

[0002] Resistance spot welding is a type of pressure welding. During welding, the workpiece is clamped by two electrodes. After power is turned on, the current flows through the workpiece to generate a large amount of resistance heat. When the resistance heat reaches a certain level, the contact surface material is heated to a plastic state or melted, thereby forming a weld to connect the plates together. Resistance spot welding has been widely used in rail passenger cars, ships, aerospace and other fields due to its low cost and wide range of applications. During the welding process, the spot welding power-on time is short, the welding current is large, and the molten core is closed and invisible, which is difficult to observe directly and is prone to quality problems. Therefore, it is very important to conduct quality inspection of spot welding joints. Incomplete penetration of spot welding is the most dangerous defect, so meeting the requirements of the spot welding core size is an important condition for becoming a qualified spot welding joint.

[0003] There are two commonly used resistance spot welding detection methods, one is destructive detection and the other is non-destructive detection. Destructive detection can accurately detect the spot welding nugget, but the detection is random and cannot detect all spot welding joints, and it will cause material waste; non-destructive testing includes post-weld non-destructive testing and online monitoring based on the welding process. Post-weld non-destructive testing has slow feedback and cannot realize online detection and control of joint quality. Online monitoring based on the welding process monitors the welding current, electrode pressure, dynamic resistance and other information in real time. Although it can realize spot welding quality assessment, resistance spot welding is a multi-variable coupled non- This method cannot comprehensively and accurately evaluate the welding quality due to its linear and complex process. Therefore, it is necessary to establish an online monitoring and evaluation model for spot welding quality. Summary of the invention

[0004] The purpose of the present invention is to provide an online monitoring method for resistance spot welding quality based on ultrasonic guided waves, which solves the problem that the existing technology cannot comprehensively and accurately evaluate the spot welding quality. The present invention realizes real-time monitoring of all welding points in the spot welding process, ensures the accuracy and reliability of the monitoring results, and ensures the quality requirements of resistance spot welding.

[0005] The above-mentioned object of the present invention is achieved by the following technical solutions: The online monitoring method of resistance spot welding quality based on ultrasonic guided wave collects ultrasonic guided wave signals to obtain characteristic values ​​of ultrasonic guided wave signals, uses principal component analysis to reduce dimensions to obtain principal component variables, and then uses the principal component variables as input samples of a support vector machine to obtain a regression prediction model for spot welding nugget diameter; the method comprises the following steps: Step 1, performing online detection on the resistance spot welding and collecting ultrasonic guided wave signals; Four sets of ultrasonic guided wave probes are used, which are fixed on the upper plate of the test piece at 45° intervals. The four sets of ultrasonic guided wave probes simultaneously transmit and receive ultrasonic guided wave signals during the spot welding process; an ultrasonic guided wave excitation probe and an ultrasonic guided wave receiving probe constitute a set of ultrasonic guided wave probes; The industrial computer controls the ultrasonic guided wave excitation probe and the ultrasonic guided wave receiving probe to transmit and receive the guided wave signal. During the welding process, the ultrasonic guided wave signal is collected every 0.002s and processed. Step 2: Extract the characteristic quantity of ultrasonic guided wave signal and perform normalization processing: Select five characteristic quantities a of a single ultrasonic guided wave i , that is, the maximum amplitude a 1 、Waveform index a 2 、Pulse index a 3 , margin index a 4 , kurtosis a 5 ; Select n ultrasonic guided wave signals received by each group of ultrasonic guided wave probes every 10 ms, perform feature extraction, and calculate the mean value μ of the five feature quantities of all ultrasonic guided waves extracted from each group i , standard deviation σ i and variance According to formulas (1), (2), and (3), the characteristic matrix of each group of ultrasonic guided wave signals can be obtained, represented by B, as shown in formula (4):

[0006] At this time, the four groups of ultrasonic guided wave signal characteristics of each resistance spot welding weldment are expressed by formula (5):

[0007] Normalization processing includes: Z-score standardization of the original indicator data, that is, mapping the original data to the standard normal distribution, calculating the mean μ and standard deviation σ of the original data set, and converting it through the following formula:

[0008] Step 3: Reduce the dimensionality of the data through principal component analysis: Perform correlation analysis on the normalized feature data to obtain the corresponding correlation coefficient matrix; Perform dimensionality reduction operation on the calculated correlation coefficient matrix, that is, use principal component analysis method to reduce the dimension to obtain the principal component variables; Obtain the eigenvectors and eigenvector cumulative variance calculated by principal component analysis, and use the first eigenvectors whose cumulative contribution rate exceeds 95% of the cumulative variance as the input of the support vector machine for subsequent operations, thus achieving Dimensionality reduction; The data set selected after principal component analysis is , randomly select 80% of the data set to form the training sample The remaining data is used as test samples .

[0009] Step 4: Construct a spot welding nugget diameter prediction model: The weld nugget diameter is used as the spot welding quality evaluation index, and the support vector machine is used to predict the spot welding nugget diameter. The principal component variables are used as input samples, and the spot welding nugget diameter is used as the expected output to calculate the predicted data.

[0010] The beneficial effects of the present invention are as follows: for spot welding quality assessment, an online spot welding quality monitoring method based on ultrasonic guided waves is realized, and the influence on the calculation results is effectively reduced by sampling the parameters. At the same time, by adopting the SVR algorithm, the calculation accuracy is greatly improved compared with other algorithms, and compared with ordinary post-weld non-destructive testing, time cost is saved and detection efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The drawings described herein are used to provide further understanding of the present invention and constitute a part of this application. The illustrative examples of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0012] Figure 1 It is a schematic diagram of the ultrasonic guided wave detection principle of the online monitoring method of resistance spot welding quality based on ultrasonic guided waves of the present invention; Figure 2 It is a schematic diagram of the positions of an ultrasonic guided wave excitation probe and an ultrasonic guided wave receiving probe of the online monitoring method for resistance spot welding quality based on ultrasonic guided waves of the present invention; Figure 3 It is a time domain diagram of ultrasonic guided wave signals of the online monitoring method for resistance spot welding quality based on ultrasonic guided waves of the present invention; Figure 4 It is a maximum amplitude diagram of the ultrasonic guided wave signal of the online monitoring method for resistance spot welding quality based on ultrasonic guided waves of the present invention; Figure 5 A model training diagram of the online monitoring method for resistance spot welding quality based on ultrasonic guided waves of the present invention; Figure 6 The present invention is a flow chart of the online monitoring method of resistance spot welding quality based on ultrasonic guided waves. DETAILED DESCRIPTION

[0013] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0014] See also Figures 1 to 6 As shown, the online monitoring method of resistance spot welding quality based on ultrasonic guided wave of the present invention includes ultrasonic guided wave signal acquisition, feature information extraction, principal component analysis and support vector machine prediction, and creates a prediction model for spot welding nugget diameter. By extracting the characteristic value of ultrasonic guided wave signal, that is, 4 groups of ultrasonic guided wave probes in the process of nugget melting, each group of ultrasonic guided wave probes includes an ultrasonic guided wave excitation probe and an ultrasonic guided wave receiving probe. Ultrasonic guided wave excitation probe 11, ultrasonic guided wave excitation probe 21, ultrasonic guided wave excitation probe 31, ultrasonic guided wave excitation probe 41 emit ultrasonic guided wave signals, and ultrasonic guided wave receiving probe 12, ultrasonic guided wave receiving probe 22, ultrasonic guided wave receiving probe 32, ultrasonic guided wave receiving probe 442 receive the maximum amplitude, waveform index, pulse index, margin index, and kurtosis of the single guided wave signal, and process it to form an eigenvalue matrix, use principal component analysis to reduce the dimension to obtain the principal component variables, and then use the principal component variables as input samples of the support vector machine to obtain the regression prediction model of spot welding nugget diameter. The online monitoring of the resistance spot welding process (spot welding nugget) is realized to evaluate the spot welding quality and ensure the quality of the spot welding joint.

[0015] See also Figures 1 to 6 As shown, the online monitoring method of the resistance welding process based on ultrasonic guided waves of the present invention comprises the following steps: Step 1, performing online detection on the resistance spot welding and collecting ultrasonic guided wave signals; Four sets of ultrasonic guided wave excitation probes and ultrasonic guided wave receiving probes are used, which are fixed on the upper plate of the test piece to be welded at an interval of 45 degrees. The four sets of ultrasonic guided wave excitation probes and ultrasonic guided wave receiving probes simultaneously transmit and receive ultrasonic guided wave signals during the spot welding process. An industrial computer controls the ultrasonic guided wave excitation probe and the ultrasonic guided wave receiving probe to transmit and receive the guided wave signals. During the welding process, the ultrasonic guided wave signal is collected every 0.002s, and the ultrasonic guided wave signal is analyzed in the time domain, frequency domain and time-frequency domain to analyze the relationship between the ultrasonic guided wave signal characteristics and the spot welding nugget size.

[0016] Step 2: Extract the characteristic quantity of the waveguide signal and perform normalization processing: Select five characteristic quantities a of a single ultrasonic guided wave i , that is, the maximum amplitude a 1 、Waveform index a 2 、Pulse index a 3 , margin index a 4 , kurtosis a 5 .

[0017] Select n ultrasonic guided wave signals received by each group of ultrasonic guided wave probes every 10 ms, perform feature extraction, and calculate the mean value μ of the five feature quantities of all ultrasonic guided waves extracted from each group i , standard deviation σ i and variance According to formulas (1), (2), and (3), the characteristic matrix of each group of ultrasonic guided wave signals can be obtained, represented by B, as shown in formula (4):

[0018] At this time, the four groups of ultrasonic guided wave signal characteristics of each resistance spot welding weldment can be expressed by formula (5):

[0019] Normalization processing includes: Z-score standardization of the original indicator data, that is, mapping the original data to the standard normal distribution, calculating the mean μ and standard deviation σ of the original data set, and converting it through the following formula:

[0020] Step 3: Reduce the dimensionality of the data through principal component analysis: Perform correlation analysis on the normalized feature data to obtain the corresponding correlation coefficient matrix; Perform dimensionality reduction operation on the calculated correlation coefficient matrix, that is, use principal component analysis method to reduce the dimension to obtain the principal component variables; Obtain the eigenvectors and eigenvector cumulative variance calculated by principal component analysis, and use the first eigenvectors whose cumulative contribution rate exceeds 95% of the cumulative variance as the input of the support vector machine for subsequent operations, thus achieving Dimensionality reduction; The data set selected after principal component analysis is , select 80% of the data set to form the training sample The remaining data is used as test samples .

[0021] Step 4: Train the model and build a spot welding nugget diameter prediction model: In order to determine the model parameters, such as the kernel function type, penalty parameter C and kernel function parameters, the cross-validation method is used to tune the parameters; the overall training sample set is divided into p subsets, each subset is used as a validation set, and the remaining subsets are combined as training sets to obtain p model evaluation indicators, and the optimal parameter combination is determined based on the evaluation indicators; after determining the parameters of the model, the entire data set is used for training to obtain the final prediction model; during the training process, the standard support vector machine algorithm or ensemble learning method is used to further improve the prediction performance of the model to obtain better prediction results.

[0022] In general, the online monitoring method of the resistance spot welding process based on ultrasonic guided wave of the present invention aims at the problem of spot welding nugget diameter prediction. First, the initial guided wave signal is processed to obtain the characteristic quantity related to the nugget size, and then the relevant data is normalized. Then, the principal component analysis (PCA) is used to reduce the dimension to obtain k principal component variables, and then the cross-validation method is used to obtain the optimal parameters of the model, and then a function model is obtained. Finally, the k principal component indicators of the support vector machine parameters obtained by training are input into the model to obtain the spot welding nugget diameter size of the next sampling point. Example

[0023] This embodiment performs online monitoring of the resistance spot welding process of the SUS301L austenitic stainless steel plate used in the rail passenger car body. The plate specification is 150mm×150mm, the plate thickness combination is 2mm+1mm, and 4 sets of ultrasonic guided wave excitation / receiving probes are used. The center frequency of the excitation probe is 2.5MHz, the excitation incident angle is 30°, and the distance between the guided wave excitation probe and the receiving probe is 150mm. The detection method of the present invention is used to perform online detection of the SUS301L stainless steel resistance spot welding nugget and display and store it. The actual nugget diameter inside the SUS301L stainless steel resistance spot welding joint is measured by metallographic test and compared. The results show that the maximum error between the spot welding nugget diameter detected by the method of the present invention and the measured actual nugget diameter does not exceed 0.1mm.

[0024] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made to the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for online monitoring of resistance spot welding quality based on ultrasonic guided waves, characterized in that: The ultrasonic guided wave signal is collected to obtain the characteristic value of the ultrasonic guided wave signal, the principal component analysis method is used to reduce the dimension to obtain the principal component variable, and then the principal component variable is used as the input sample of the support vector machine to obtain the spot welding nugget diameter regression prediction model; the steps include: Step 1, performing online detection on the resistance spot welding and collecting ultrasonic guided wave signals; Step 2: extracting the characteristic quantity of the ultrasonic guided wave signal and performing normalization processing; Step 3: Reduce the dimensionality of the data through principal component analysis; Step 4, constructing a spot welding nugget diameter prediction model; The extraction of the characteristic quantity of the ultrasonic guided wave signal and the normalization processing are specifically: selecting 5 characteristic quantities a of a single ultrasonic guided wave i , namely, the maximum amplitude a1, waveform index a2, pulse index a3, margin index a4, kurtosis a5. A group of ultrasonic guided wave probes includes an ultrasonic guided wave excitation probe and an ultrasonic guided wave receiving probe; select n ultrasonic guided wave signals received by each group of ultrasonic guided wave probes every 10ms, and perform feature extraction, and calculate the mean value μ of the five feature quantities of all ultrasonic guided waves extracted from each group i , standard deviation σ i and variance According to formulas (1)-(3), the characteristic matrix of each group of ultrasonic guided wave signals is obtained, which is represented by B: ; At this time, the four groups of ultrasonic guided wave signal characteristics of each resistance spot welding weldment are expressed by formula (5): ; Normalization processing: Map the original data to a normal distribution, find the mean μ and standard deviation σ of the original data set, and convert them using formula (6): 。 2. The method for online monitoring of resistance spot welding quality based on ultrasonic guided waves according to claim 1, characterized in that: The online detection of the resistance spot welding described in step 1 and the collection of ultrasonic guided wave signals are specifically: 4 groups of ultrasonic guided wave probes are used, which are fixed on the upper plate of the test piece to be welded at an interval of 45°, and the 4 groups of ultrasonic guided wave probes simultaneously transmit and receive ultrasonic guided wave signals during the spot welding process; an industrial computer controls the ultrasonic guided wave probe to transmit and receive guided wave signals, and during the welding process, the ultrasonic guided wave signal is collected every 0.002s.

3. The method for online monitoring of resistance spot welding quality based on ultrasonic guided waves according to claim 1, characterized in that: The principal component analysis method described in step 3 is used to reduce the dimensionality of the data, specifically: Perform correlation analysis on the normalized feature data to obtain the corresponding correlation coefficient matrix; The calculated correlation coefficient matrix is ​​subjected to dimensionality reduction operation, and the principal component analysis method is used to obtain the principal component variables; Obtain the eigenvectors and eigenvector cumulative variance calculated by principal component analysis, and use the first eigenvectors whose cumulative contribution rate exceeds 95% of the cumulative variance as the input of the support vector machine for subsequent operations to achieve Dimensionality reduction; The data set selected after principal component analysis is , randomly select 80% of the data set to form the training sample The remaining data is used as test samples .

4. The method for online monitoring of resistance spot welding quality based on ultrasonic guided waves according to claim 1, characterized in that: The step 4 of constructing a spot welding nugget diameter prediction model specifically includes: using the spot welding nugget diameter as a spot welding quality evaluation index, using a support vector machine to predict the spot welding nugget diameter, using the principal component variables as input samples, and using the spot welding nugget diameter as the expected output to calculate the prediction data.