Spot welding quality detection method and detection system based on ultrasonic phased array flaw detection

Abstract

The invention discloses a spot welding quality detection method and system based on ultrasonic phased array flaw detection. Comprising the following steps: obtaining structure data and a spot welding detection thermodynamic diagram of a preprocessed spot welding object; according to the structural data, calculating to obtain a quality discrimination threshold value of the selected quality discrimination parameter; when the index data corresponding to the spot welding detection thermodynamic diagram all meet the quality judgment threshold value, it is judged that the spot welding quality is qualified; and otherwise, determining that the spot welding quality is unqualified, and inputting the corresponding spot welding detection thermodynamic diagram into the trained spot welding defect classification model to obtain a corresponding spot welding defect classification result. Wherein the spot welding detection thermodynamic diagram is obtained by detecting the preprocessed spot welding object through an ultrasonic phased array flaw detection device based on a phased array focusing principle. The quality of spot welding objects can be quickly and accurately judged and classified, the working efficiency of field personnel is improved, and the missing detection rate and the false detection rate of detection personnel are reduced.

Description

Technical Field

[0001] The invention relates to a spot welding quality detection method and a detection system based on ultrasonic phased array flaw detection, belonging to the technical field of spot welding quality detection. Background Art

[0002] Urban rail vehicles involve a stainless steel resistance spot welding structure vehicle. The quality of the spot welding structure determines the load-bearing capacity and service safety of the vehicle structure. Therefore, conducting welding quality inspection on the spot welding structure can ensure the integrity, reliability, safety and usability of the welding structure.

[0003] In addition to the requirements for welding technology and welding process, welding quality inspection is also an important part of welding structure quality management. The existing online inspection of spot welding of rail transit vehicles mostly uses visual inspection and chisel inspection as the detection means. However, ultrasonic detection means need to select different probe sizes according to different weld spot sizes, and ordinary ultrasonic probes are not convenient for handheld operation in the narrow space of some actual vehicle structures during use. Visual inspection cannot detect the size of the weld core. Chisel inspection can determine the virtual weld, but the efficiency is very low and it is not suitable for application in some places. Secondly, in the field of rail transit, when vehicles enter the frame for overhaul, it is necessary to quickly inspect the spot welding structure and identify the defect type of the spot welding structure, which is especially important for the status evaluation and further maintenance of the vehicle body structure. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a spot welding quality detection method and detection system based on ultrasonic phased array flaw detection. On the one hand, the ultrasonic phased array flaw detection device realizes the rapid and accurate detection of the weld nugget size of the spot welding object; on the other hand, through the quality discrimination threshold and the spot welding defect classification model, it can realize the rapid and intelligent discrimination and classification of the quality of the spot welding object, improve the work efficiency of the on-site personnel, and reduce the missed detection and false detection rate of the detection personnel.

[0005] To achieve the above object, the present invention is implemented by adopting the following technical solutions:

[0006] On the one hand, the present invention discloses a spot welding quality detection method based on ultrasonic phased array flaw detection, comprising the following steps:

[0007] Obtaining the structural data of the pre-processed spot welding object and the spot welding detection thermal map;

[0008] Calculating a quality discrimination threshold of a selected quality discrimination parameter according to the structural data;

[0009] In response to the indicator data corresponding to the spot welding detection thermogram satisfying the quality discrimination threshold, the spot welding quality of the spot welding object is determined to be qualified; otherwise, the spot welding quality of the spot welding object is determined to be unqualified, and the corresponding spot welding detection thermogram is input into the trained spot welding defect classification model to obtain the corresponding spot welding defect classification result;

[0010] Among them, the spot welding detection thermal map is obtained by detecting the pre-treated spot welding object through an ultrasonic phased array flaw detection device; the ultrasonic phased array flaw detection device is equipped with multiple ultrasonic chips, which realizes the measurement of the spot welding object based on the phased array focusing principle.

[0011] Furthermore, the pre-processing step includes:

[0012] Get the original spot welding object;

[0013] The original spot welding object is subjected to paint removal and decontamination treatment to ensure that the surface of the spot welding object is smooth, thereby obtaining a pre-treated spot welding object.

[0014] Further, the pre-treated spot welding object includes a first stainless steel plate and a second stainless steel plate, and a spot welding area is provided between a lower surface of one end of the first stainless steel plate and an upper surface of one end of the second stainless steel plate;

[0015] The structural data includes the thickness of the first stainless steel plate and the thickness of the second stainless steel plate;

[0016] The quality discrimination parameters include the minimum core diameter, the maximum indentation depth, the minimum indentation depth, the maximum remaining plate thickness and the minimum remaining plate thickness. The quality discrimination thresholds of the quality discrimination parameters are:

[0017] The minimum core diameter The expression is: ;

[0018] The maximum indentation depth The expression is: ;

[0019] The minimum indentation depth The expression is: ;

[0020] The maximum remaining plate thickness The expression is: ;

[0021] The minimum remaining plate thickness The expression is: ;

[0022] In the formula, Indicates the thickness of the first stainless steel plate; Indicates the thickness of the second stainless steel plate.

[0023] Furthermore, the indicator data corresponding to the spot welding detection thermogram includes the detection of the core diameter , Detect indentation depth And detect the remaining plate thickness ;

[0024] In response to the indicator data corresponding to the spot welding detection thermogram satisfying the quality discrimination threshold, the method includes:

[0025] Any detection core diameter that satisfies the spot welding detection thermal diagram Greater than or equal to the minimum core diameter , ;

[0026] And the detection indentation depth corresponding to the spot welding detection thermal map is satisfied Greater than or equal to the minimum indentation depth , and less than or equal to the maximum indentation depth , ;

[0027] And the detection remaining plate thickness corresponding to the spot welding detection thermal map is satisfied Greater than or equal to the minimum remaining plate thickness , and less than or equal to the maximum remaining plate thickness , ;

[0028] The spot welding quality of the spot welding object is judged to be qualified.

[0029] Furthermore, the steps for obtaining the spot welding detection thermal map are as follows:

[0030] After applying coupling agent on the upper surface of the stainless steel plate corresponding to the spot welding area of ​​the pre-treated spot welding object, an ultrasonic phased array flaw detection device is placed just above the spot welding area for scanning to obtain a spot welding detection thermal map;

[0031] Wherein, the ultrasonic phased array flaw detection device scans a single spot weld at a speed of not less than 5S.

[0032] Furthermore, the ultrasonic phased array flaw detection device comprises a data cable, a shell and a probe, the shell is movably connected to the probe, the shell is in an L-shaped structure, and a handheld anti-slip groove is provided on the shell;

[0033] The probe is integrated with a plurality of ultrasonic chips, and the peripheral ultrasonic chips are symmetrically arranged with respect to the central ultrasonic chip to realize the focusing function of the phased array; the data cable runs through the shell and is respectively connected to each ultrasonic chip, and each ultrasonic chip corresponds to an ultrasonic control channel.

[0034] Furthermore, the ultrasonic generating frequency of the ultrasonic chip is 14 MHz~20 MHz; the size of the ultrasonic chip is 1.30*1.30 mm.

[0035] Furthermore, the training method of the spot welding defect classification model is as follows:

[0036] Obtain multiple trained spot welding inspection heat maps with spot welding defect labels;

[0037] For each training spot welding detection heat map, denoising, contrast enhancement and cropping are performed respectively to obtain the processed training spot welding detection heat map, and then the training set is constructed;

[0038] Based on the training set, the pre-built spot welding defect classification model is trained until a preset training termination condition is met to obtain a trained spot welding defect classification model;

[0039] Among them, the spot welding defect label includes the first spot welding quality defect, the second spot welding quality defect and the third spot welding quality defect; the first spot welding quality defect includes weld desoldering; the second spot welding quality defect includes bubbles, slag inclusions, and point defects; the third quality defect includes weak welds.

[0040] Furthermore, the spot welding defect classification model includes:

[0041] CNN convolutional neural network, used to extract features from the input spot welding detection thermal map to obtain feature data, wherein the feature data includes edge, texture, and shape data;

[0042] The SVM support vector machine is used to map the feature data to a high-dimensional space, find the best hyperplane for classification, and obtain the corresponding spot welding defect classification results.

[0043] On the other hand, the present invention discloses a spot welding quality inspection system based on ultrasonic phased array flaw detection, which is applicable to the above-mentioned spot welding quality inspection method based on ultrasonic phased array flaw detection, comprising:

[0044] A data acquisition module is used to acquire the structural data of the spot welding object after preprocessing and the spot welding detection thermal map;

[0045] A discrimination threshold module, used for calculating a quality discrimination threshold of a selected quality discrimination parameter according to the structural data;

[0046] A defect classification module is used for judging that the spot welding quality of the spot welding object is qualified in response to the index data corresponding to the spot welding detection thermogram satisfying the quality discrimination threshold; otherwise, judging that the spot welding quality of the spot welding object is unqualified, and inputting the corresponding spot welding detection thermogram into the trained spot welding defect classification model to obtain the corresponding spot welding defect classification result;

[0047] Among them, the spot welding detection thermal map is obtained by detecting the pre-treated spot welding object through an ultrasonic phased array flaw detection device; the ultrasonic phased array flaw detection device is equipped with multiple ultrasonic chips, which realizes the measurement of the spot welding object based on the phased array focusing principle.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The spot welding quality detection method and detection system based on ultrasonic phased array flaw detection of the present invention, firstly, based on the ultrasonic phased array flaw detection device based on the phased array focusing principle to perform measurement, can realize the rapid and accurate detection of the molten core size of the spot welding object; secondly, based on the constructed quality discrimination threshold, can realize the rapid discrimination of the quality of the spot welding object; finally, based on the spot welding defect classification model, can realize the rapid classification of the quality of the spot welding object, normalize the data, improve the work efficiency of the on-site personnel, and reduce the missed detection and false detection rate of the detection personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of a spot welding quality detection method based on ultrasonic phased array flaw detection provided in Example 1;

[0051] Figure 2 is a schematic structural diagram of a spot welding object provided in Example 1;

[0052] Figure 3 is a schematic structural diagram of the ultrasonic phased array flaw detection device provided in Example 1;

[0053] Figure 4 is a schematic diagram of the phased array spot welding detection principle provided in Example 1;

[0054] Figure 5 It is a spot welding detection thermal diagram of qualified spot welding quality provided in Example 1;

[0055] Figure 6 Embodiment 1 provides a spot welding detection thermal map of the first spot welding quality defect;

[0056] Figure 7 Embodiment 1 provides a spot welding detection thermal diagram of a second spot welding quality defect;

[0057] Figure 8 Embodiment 1 provides a spot welding detection thermal diagram of a third spot welding quality defect;

[0058] In the figure: 1. Shell; 2. Probe; 3. Wedge; 4. Handheld anti-slip groove; 5. Ultrasonic chip; 6. Data cable; 7. First stainless steel plate; 8. Second stainless steel plate; 9. Spot welding area. DETAILED DESCRIPTION

[0059] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0060] Example 1

[0061] This embodiment 1 provides a spot welding quality detection method based on ultrasonic phased array flaw detection, such as Figure 1 As shown, the following steps are included:

[0062] Obtaining the structural data of the pre-processed spot welding object and the spot welding detection thermal map;

[0063] According to the structural data, a quality discrimination threshold of the selected quality discrimination parameter is calculated;

[0064] In response to the indicator data corresponding to the spot welding detection heat map satisfying the quality discrimination threshold, the spot welding quality of the spot welding object is judged to be qualified; otherwise, the spot welding quality of the spot welding object is judged to be unqualified, and the corresponding spot welding detection heat map is input into the trained spot welding defect classification model to obtain the corresponding spot welding defect classification result;

[0065] The spot welding detection thermogram is obtained by detecting the pre-treated spot welding object through an ultrasonic phased array flaw detection device; a plurality of ultrasonic chips 5 are arranged in the ultrasonic phased array flaw detection device, and the spot welding object is measured based on the phased array focusing principle.

[0066] The technical concept of the present invention is as follows: first, based on the ultrasonic phased array flaw detection device, the measurement is performed based on the phased array focusing principle, so that the size of the molten core of the spot welding object can be quickly and accurately detected; second, based on the constructed quality discrimination threshold, the quality of the spot welding object can be quickly discriminated; finally, based on the spot welding defect classification model, the quality of the spot welding object can be quickly classified, the data can be standardized, the work efficiency of the on-site personnel can be improved, and the missed detection and false detection rates of the detection personnel can be reduced.

[0067] The specific steps are as follows:

[0068] Step 1: Data acquisition.

[0069] 1.1. Pretreatment of spot welding objects.

[0070] The preprocessing steps include:

[0071] Get the original spot welding object;

[0072] The original spot welding object is stripped of paint and decontaminated to ensure that the surface of the spot welding object is smooth, and all rust, spatter and dirt that affect ultrasonic testing are removed to obtain the pre-treated spot welding object.

[0073] 1.2. Acquisition of structural data.

[0074] like Figure 2 As shown, the spot welding object after pretreatment includes a first stainless steel plate 7 and a second stainless steel plate 8 , and a spot welding area 9 is provided between a lower surface of one end of the first stainless steel plate 7 and an upper surface of one end of the second stainless steel plate 8 .

[0075] Before the detection begins, the thickness of the two layers of stainless steel plates is detected using equipment to obtain structural data, which includes the thickness of the first stainless steel plate 7 and the thickness of the second stainless steel plate 8 .

[0076] 1.3. Acquisition of thermal map of spot welding inspection.

[0077] The steps to obtain the spot welding detection thermal map are as follows:

[0078] After applying coupling agent on the upper surface of the stainless steel plate corresponding to the spot welding area 9 of the pre-treated spot welding object, an ultrasonic phased array flaw detection device is placed just above the spot welding area 9 for scanning to obtain a spot welding detection thermal map;

[0079] Among them, the speed of scanning a single spot weld by the ultrasonic phased array flaw detection device is not less than 5S.

[0080] The ultrasonic phased array flaw detection device comprises a data cable 6 , a shell 1 and a probe 2 , wherein the shell 1 is movably connected to the probe 2 .

[0081] A wedge 3 is also provided at the front of the probe 2. The wedge 3 is made of resin material and its size is specifically designed according to the size of the spot welding spot of the rail transit vehicle to cover the welding spots of different sizes and detect the changes in the magnetic field caused by the eddy current during the flaw detection. The wedge 3 is fixed to the housing 1 by bolts to achieve a reliable connection between the probe 2 and the housing 1.

[0082] The shell 1 is designed specifically according to the size of the hand-held part. The shell 1 is L-shaped, which is more convenient when performing flaw detection on rail transit vehicles. A hand-held anti-slip groove 4 is provided on the shell 1 to facilitate hand-held operation.

[0083] A plurality of ultrasonic chips 5 are integrated in the probe 2, and the peripheral ultrasonic chips 5 are symmetrically arranged with respect to the central ultrasonic chip 5 to realize the focusing function of the phased array; the data cable 6 passes through the shell 1 and is respectively connected to each ultrasonic chip 5, and each ultrasonic chip 5 corresponds to an ultrasonic control channel.

[0084] like Figure 3As shown, this embodiment simulates the phased array multi-sensor sound field based on the similarity of the multi-Gaussian acoustic beam model of a single sensor. In order to design an ultrasonic phased array to meet the detection needs of vehicle spot welding structures, namely, detection range and accuracy, according to the characteristics of the vehicle structure and considering the convenience of on-site flaw detection operations, 52 independent ultrasonic chips 5 are integrated into a circular stainless steel shell with a diameter of 20 mm. The ultrasonic chips 5 are closely arranged, and the integration method is shown in FIG. Figure 3 The ultrasonic generation frequency of the ultrasonic chip 5 is 14 MHz~20 MHz, the number of ultrasonic chips 5 is 52, the number of ultrasonic control channels is 52, the size of the ultrasonic chip 5 is 1.30*1.30mm, the shell 1 is made of stainless steel and is designed as a 90° angle probe 2.

[0085] The ultrasonic phased array flaw detection device is convenient for detection personnel to use in a small space, can greatly enhance the working comfort, and is convenient for handheld operation when performing flaw detection on rail transit vehicles. The probe 2 has a detection range and accuracy that meets the requirements of the spot welding structure of the vehicle. The average error of the weld nugget size obtained by the ultrasonic array detection method is controlled within 0.18 mm. The error comparison data table of the detection results is as follows:

[0086] Test result error comparison data table 1

[0087]

[0088] Test result error comparison data table 2

[0089]

[0090] Ultrasonic phased array technology is to control the computer to excite multiple ultrasonic chips 5 with ultrasonic beams of different amplitude and delay characteristics. In order to make the phased sound beams interfere with each other reasonably, the pulse excitation time of each ultrasonic chip 5 of the ultrasonic sensor is slightly different, so the reflected signals received by the ultrasonic chip 5 have a time difference before comprehensive summation.

[0091] Phased array spot welding detection principle Figure 4 As shown. During the transmission process, the phased array system obtains a signal stimulated by the signal acquisition system. At this time, the signal is converted into a high-voltage pulse. The delay and width of the pulse are specified by the software-defined focusing law, and it is transmitted through the ultrasonic chip 5 to form a sound beam with a certain focusing depth and a special angle. When the sound beam encounters an obstacle or defect, it is reflected back. In the received signal, the reflected sound wave is received by the ultrasonic chip 5, and is reintegrated into an ultrasonic pulse through the focusing law defined by the phased array system and then received by the signal acquisition system. Therefore, the ultrasonic phased array system, based on the physical interference and superposition of waves and the Huygens principle, has two main characteristics:

[0092] (1) The deflection of the phased array is set by software to adjust the excitation order of each ultrasonic chip 5, so that the excitation time of each chip is different. According to Huygens' principle, a wave is a propagation of a vibration state. Any particle in the wave can be regarded as a new wave source. These waves can form a new wave front through interference, and continue to propagate in this way, thereby achieving the control of the propagation direction of the ultrasonic wave.

[0093] (2) The focusing principle of the phased array is similar to the deflection of the phased array. If the delay time set for each chip makes the chips at both ends symmetrical about the middle chip, the focusing function of the phased array will be realized. The focusing depth is related to the delay time. The shallower the depth, the longer the delay time needs to be set. The software controls the piezoelectric chip and controls the time when each chip emits the sound beam. Through mutual interference, a main wavefront is formed. This wavefront will reflect signals at the edge and defects of the material. Therefore, the ultrasonic phased array system can obtain signals with different focusing depths and different angles to judge defects.

[0094] Compared with the traditional single-transmitter ultrasonic detection, ultrasonic C-scan detection, and ultrasonic array detection, the ultrasonic array synthetic aperture detection technology achieves a focusing effect by controlling the signal of the array probe 2, thereby improving the detection accuracy of the array probe 2, and can realize fast and accurate detection of the spot welding nugget size, and can shorten the detection time of the spot welding nugget to less than 10s, thereby realizing comprehensive shooting or scanning of the spot welding object, and greatly improving the accuracy and stability of image acquisition.

[0095] Step 2: Quality judgment.

[0096] According to the structural data obtained in step 1.2, the quality discrimination threshold of the selected quality discrimination parameter is calculated. The quality discrimination parameters include the minimum nugget diameter, the maximum indentation depth, the minimum indentation depth, the maximum remaining plate thickness and the minimum remaining plate thickness. The quality discrimination thresholds of each quality discrimination parameter are:

[0097] Minimum core diameter The expression is: ;

[0098] Maximum indentation depth The expression is: ;

[0099] Minimum indentation depth The expression is: ;

[0100] Maximum remaining plate thickness The expression is: ;

[0101] Minimum remaining plate thickness The expression is: ;

[0102] In the formula, represents the thickness of the first stainless steel plate 7; represents the thickness of the second stainless steel plate 8 .

[0103] The indicator data corresponding to the spot welding detection thermal map includes the detection of the melt core diameter , Detect indentation depth And detect the remaining plate thickness ;

[0104] The indicator data corresponding to the spot welding detection thermal map all meet the quality judgment threshold, including:

[0105] Any detection core diameter that satisfies the spot welding detection thermal diagram Greater than or equal to the minimum core diameter , ;

[0106] And meet the detection indentation depth corresponding to the spot welding detection thermal map Greater than or equal to the minimum indentation depth , and less than or equal to the maximum indentation depth , ;

[0107] And meet the detection of the remaining plate thickness corresponding to the spot welding detection thermal map Greater than or equal to the minimum remaining plate thickness , and less than or equal to the maximum remaining plate thickness , ;

[0108] The spot welding quality of the spot welding object is judged to be qualified; otherwise, the spot welding quality of the spot welding object is judged to be unqualified, and go to step three, input the corresponding spot welding detection thermal map into the trained spot welding defect classification model to obtain the corresponding spot welding defect classification result.

[0109] Step 3: Defect classification.

[0110] 3.1. Construction of spot welding defect classification model.

[0111] The spot welding defect classification model includes CNN convolutional neural network and SVM support vector machine.

[0112] 3.1.1. CNN convolutional neural network is used to extract features from the input spot welding detection thermal map to obtain feature data, which includes edge, texture, and shape data. CNN convolutional neural network includes an input layer, multiple stacked convolution-pooling layers, a fully connected layer, and an output layer.

[0113] Specifically, the input layer receives the regularized spot welding detection heat map. Images are usually represented in the form of a pixel matrix, where each pixel has the value of one or more color channels, such as the three channels of an RGB image.

[0114] The convolution-pooling layer includes a convolution layer, an activation function, and a pooling layer. The convolution layer performs a convolution operation on the input image by sliding multiple filters, i.e., convolution kernels. Each filter is a small matrix used to detect specific features in the image, such as edges, textures, etc. The convolution operation is the process of performing dot multiplication and summation of the filter with the pixels at the corresponding position on the image to generate a new feature map. Each filter generates a feature map. After the convolution operation, a nonlinear activation function, the Sigmoid function, is applied. The activation function maps the output of the convolution operation to a nonlinear space, increasing the nonlinear expression ability of the model. The pooling layer uses the average pooling rule to calculate the average value of each local area. The pooling layer is located after the convolution layer to reduce the size of the feature map and reduce the computational complexity.

[0115] Through multiple stacked convolution-pooling layers, higher-level features are gradually extracted. As the network goes deeper, the extracted features gradually change from low-level edges, textures, etc. to higher-level shapes, object parts, etc.

[0116] The fully connected layer flattens the output of the convolutional layer and the pooling layer into a one-dimensional vector and performs a linear transformation through a series of weights and biases. The function of the fully connected layer is to integrate the local features extracted previously and output the final classification or recognition result.

[0117] The output layer outputs the corresponding results according to the task type. For classification tasks, the output layer uses the softmax function to convert the output into a probability distribution, indicating the probability that the image belongs to each category.

[0118] 3.1.2. SVM support vector machine is used to map feature data into high-dimensional space, find the best hyperplane for classification, and obtain the corresponding spot welding defect classification results.

[0119] Use the kernel function of the support vector machine (SVM) to map the input feature data to a high-dimensional space and find the best hyperplane for classification. This task uses a linear kernel function and sets the parameters of the SVM, including the penalty factor C and the kernel function parameters. These parameters can be optimized by methods such as cross-validation.

[0120] 3.2. Training of spot welding defect classification model.

[0121] The training method of the spot welding defect classification model is as follows:

[0122] (1) Obtain multiple trained spot welding detection heat maps with spot welding defect labels,

[0123] The spot welding defect labels include the first spot welding quality defect, the second spot welding quality defect and the third spot welding quality defect.

[0124] The first spot welding quality defect includes the desoldering of the solder joint, which is manifested as the detected thickness being smaller than the actual size, such as Figure 6 shown.

[0125] The second spot welding quality defects include bubbles, slag inclusions, and point defects, such as Figure 7 shown.

[0126] The third quality defect includes weak welds, which are manifested as small and irregular weld core diameters. Figure 8 shown.

[0127] The schematic diagram of the spot welding detection thermogram with qualified spot welding quality is as follows Figure 5 As shown, spot welding detection thermal diagram.

[0128] (2) During the image acquisition process, it may be interfered by various noises, which may affect the image quality. Therefore, for each training spot welding detection thermal map, denoising, contrast enhancement and cropping are performed to improve the clarity and detail visibility of the image and reduce the difficulty of model training. The processed training spot welding detection thermal map is obtained, and then the training set is constructed.

[0129] (3) Based on the training set, the pre-built spot welding defect classification model is trained until the preset training termination condition is met, thereby obtaining a trained spot welding defect classification model.

[0130] The training termination condition in this embodiment is not limited here, and can be set to when the number of training times reaches a preset iteration threshold; or can be set to when the classification accuracy reaches a preset accuracy threshold.

[0131] 3.3. Use of spot welding defect classification model.

[0132] Based on the CNN convolutional neural network, feature extraction is performed on the input spot welding detection thermal map to obtain feature data;

[0133] Based on the SVM support vector machine, the feature data is mapped to a high-dimensional space, and the best hyperplane is found for classification to obtain the corresponding spot welding defect classification results. The spot welding defect classification results can be one or more.

[0134] Example 2

[0135] This embodiment 2 provides a spot welding quality detection system based on ultrasonic phased array flaw detection, which is applicable to the spot welding quality detection method based on ultrasonic phased array flaw detection in embodiment 1, including:

[0136] A data acquisition module is used to acquire the structural data of the spot welding object after preprocessing and the spot welding detection thermal map;

[0137] A discrimination threshold module, used for calculating a quality discrimination threshold of a selected quality discrimination parameter according to the structural data;

[0138] The defect classification module is used to judge the spot welding quality of the spot welding object to be qualified if the index data corresponding to the spot welding detection heat map all meet the quality judgment threshold; otherwise, the spot welding quality of the spot welding object is judged to be unqualified, and the corresponding spot welding detection heat map is input into the trained spot welding defect classification model to obtain the corresponding spot welding defect classification result;

[0139] The spot welding detection thermogram is obtained by detecting the pre-treated spot welding object through an ultrasonic phased array flaw detection device; a plurality of ultrasonic chips 5 are arranged in the ultrasonic phased array flaw detection device, and the spot welding object is measured based on the phased array focusing principle.

[0140] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0141] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0142] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0143] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0144] 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 spot welding quality detection method based on ultrasonic phased array flaw detection, characterized in that: The steps include: Obtaining the structural data of the pre-processed spot welding object and the spot welding detection thermal map; Calculating a quality discrimination threshold of a selected quality discrimination parameter according to the structural data; In response to the indicator data corresponding to the spot welding detection thermogram satisfying the quality discrimination threshold, the spot welding quality of the spot welding object is determined to be qualified; otherwise, the spot welding quality of the spot welding object is determined to be unqualified, and the corresponding spot welding detection thermogram is input into the trained spot welding defect classification model to obtain the corresponding spot welding defect classification result; Among them, the spot welding detection thermal map is obtained by detecting the pre-treated spot welding object through an ultrasonic phased array flaw detection device; the ultrasonic phased array flaw detection device is equipped with multiple ultrasonic chips, which realizes the measurement of the spot welding object based on the phased array focusing principle.

2. The spot welding quality detection method based on ultrasonic phased array flaw detection according to claim 1 is characterized in that: The pre-processing steps include: Get the original spot welding object; The original spot welding object is subjected to paint removal and decontamination treatment to ensure that the surface of the spot welding object is smooth, thereby obtaining a pre-treated spot welding object.

3. The spot welding quality detection method based on ultrasonic phased array flaw detection according to claim 1 is characterized in that: The pre-treated spot welding object includes a first stainless steel plate and a second stainless steel plate, and a spot welding area is provided between a lower surface of one end of the first stainless steel plate and an upper surface of one end of the second stainless steel plate; The structural data includes the thickness of the first stainless steel plate and the thickness of the second stainless steel plate; The quality discrimination parameters include the minimum core diameter, the maximum indentation depth, the minimum indentation depth, the maximum remaining plate thickness and the minimum remaining plate thickness. The quality discrimination thresholds of the quality discrimination parameters are: The minimum core diameter The expression is: ; The maximum indentation depth The expression is: ; The minimum indentation depth The expression is: ; The maximum remaining plate thickness The expression is: ; The minimum remaining plate thickness The expression is: ; In the formula, Indicates the thickness of the first stainless steel plate; Indicates the thickness of the second stainless steel plate.

4. The spot welding quality detection method based on ultrasonic phased array flaw detection according to claim 3 is characterized in that: The index data corresponding to the spot welding detection thermal map includes the detection of the melt core diameter , Detect indentation depth And detect the remaining plate thickness ; In response to the indicator data corresponding to the spot welding detection thermogram satisfying the quality discrimination threshold, the method includes: Any detection core diameter that satisfies the spot welding detection thermal diagram Greater than or equal to the minimum core diameter , ; And the detection indentation depth corresponding to the spot welding detection thermal map is satisfied Greater than or equal to the minimum indentation depth , and less than or equal to the maximum indentation depth , ; And the detection remaining plate thickness corresponding to the spot welding detection thermal map is satisfied Greater than or equal to the minimum remaining plate thickness , and less than or equal to the maximum remaining plate thickness , ; The spot welding quality of the spot welding object is judged to be qualified.

5. The spot welding quality detection method based on ultrasonic phased array flaw detection according to claim 3 is characterized in that: The steps for obtaining the spot welding detection thermal map are as follows: After applying coupling agent on the upper surface of the stainless steel plate corresponding to the spot welding area of ​​the pre-treated spot welding object, an ultrasonic phased array flaw detection device is placed just above the spot welding area for scanning to obtain a spot welding detection thermal map; Wherein, the ultrasonic phased array flaw detection device scans a single spot weld at a speed of not less than 5S.

6. The spot welding quality detection method based on ultrasonic phased array flaw detection according to claim 1 is characterized in that: The ultrasonic phased array flaw detection device comprises a data cable, a shell and a probe, wherein the shell is movably connected to the probe, the shell is in an L-shaped structure, and a handheld anti-slip groove is provided on the shell; The probe is integrated with a plurality of ultrasonic chips, and the peripheral ultrasonic chips are symmetrically arranged with respect to the central ultrasonic chip to realize the focusing function of the phased array; the data cable runs through the shell and is respectively connected to each ultrasonic chip, and each ultrasonic chip corresponds to an ultrasonic control channel.

7. The spot welding quality inspection method based on ultrasonic phased array flaw detection according to claim 6 is characterized in that: The ultrasonic wave chip has an ultrasonic generating frequency of 14 MHz to 20 MHz; the size of the ultrasonic wave chip is 1.30*1.30 mm.

8. The spot welding quality inspection method based on ultrasonic phased array flaw detection according to claim 1 is characterized in that: The training method of the spot welding defect classification model is as follows: Obtain multiple trained spot welding inspection heat maps with spot welding defect labels; For each training spot welding detection heat map, denoising, contrast enhancement and cropping are performed respectively to obtain the processed training spot welding detection heat map, and then the training set is constructed; Based on the training set, the pre-built spot welding defect classification model is trained until a preset training termination condition is met to obtain a trained spot welding defect classification model; Among them, the spot welding defect label includes the first spot welding quality defect, the second spot welding quality defect and the third spot welding quality defect; the first spot welding quality defect includes weld desoldering; the second spot welding quality defect includes bubbles, slag inclusions, and point defects; the third quality defect includes weak welds.

9. The spot welding quality inspection method based on ultrasonic phased array flaw detection according to claim 8 is characterized in that: The spot welding defect classification model includes: CNN convolutional neural network, used to extract features from the input spot welding detection thermal map to obtain feature data, wherein the feature data includes edge, texture, and shape data; The SVM support vector machine is used to map the feature data to a high-dimensional space, find the best hyperplane for classification, and obtain the corresponding spot welding defect classification results.

10. A spot welding quality inspection system based on ultrasonic phased array flaw detection, applicable to the spot welding quality inspection method based on ultrasonic phased array flaw detection according to any one of claims 1 to 9, characterized in that: include: A data acquisition module is used to acquire the structural data of the spot welding object after preprocessing and the spot welding detection thermal map; A discrimination threshold module, used for calculating a quality discrimination threshold of a selected quality discrimination parameter according to the structural data; A defect classification module is used for judging that the spot welding quality of the spot welding object is qualified in response to the index data corresponding to the spot welding detection thermogram satisfying the quality discrimination threshold; otherwise, judging that the spot welding quality of the spot welding object is unqualified, and inputting the corresponding spot welding detection thermogram into the trained spot welding defect classification model to obtain the corresponding spot welding defect classification result; Among them, the spot welding detection thermal map is obtained by detecting the pre-treated spot welding object through an ultrasonic phased array flaw detection device; the ultrasonic phased array flaw detection device is equipped with multiple ultrasonic chips, which realizes the measurement of the spot welding object based on the phased array focusing principle.

Images