Lithium battery top cover welding defect detection method and device and lithium battery production system

Abstract

The invention discloses a lithium battery top cover welding defect detection method and device and a lithium battery production system. The lithium battery top cover welding defect detection method comprises the steps that a lithium battery top cover image is collected; and inputting the collected lithium battery top cover image into a lithium battery top cover welding defect detection model trained by adopting the constructed welding defect data set, and outputting a lithium battery top cover welding defect detection result. According to the invention, the efficiency and precision of lithium battery top cover welding defect detection can be effectively improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of lithium battery production, and particularly relates to a method and device for detecting welding defects of a lithium battery top cover and a lithium battery production system. Background Art

[0002] At present, lithium-ion power batteries are widely used in the new energy vehicle industry. The manufacturing of battery cells requires a series of precise and complex processes such as pulp making, coating, slitting, winding, and assembly, and finally the top cover plate is sealed to the battery cell housing through a laser welding process. In the above process, the welding of the battery cell top cover plate, as an important part of the sealing welding process, is related to whether the lithium battery can work safely and stably. However, the welding quality is easily affected by factors such as laser power and product material, and defects such as pinholes, broken welds, explosion points, depressions, and blackening are extremely likely to occur. Severe welding defects will cause safety accidents such as liquid leakage and short circuits. Therefore, the detection of welding quality is extremely important. In the highly automated battery production process, how to quickly and accurately detect welding defects and avoid potential safety hazards of lithium batteries caused thereby is an urgent problem to be solved.

[0003] Currently, the main method for detecting welding defects of battery top covers is manual inspection. However, manual inspection requires picking up and flipping the battery to check whether there are defects in the welding of the top cover on each surface, which is slow and affected by subjective factors, and the detection effect is not ideal. Summary of the Invention

[0004] To solve the deficiencies in the prior art, the present invention provides a method and device for detecting welding defects of a lithium battery top cover and a lithium battery production system, which can effectively improve the efficiency and detection accuracy of detecting welding defects of a lithium battery top cover.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows: In a first aspect, a method for detecting welding defects of a lithium battery top cover is provided, including: collecting an image of the lithium battery top cover; inputting the collected image of the lithium battery top cover into a lithium battery top cover welding defect detection model trained with a constructed welding defect data set, and outputting a detection result of welding defects of the lithium battery top cover.

[0006] By inputting the collected image of the lithium battery top cover into a lithium battery top cover welding defect detection model trained with a constructed welding defect data set and outputting a detection result of welding defects of the lithium battery top cover, the above technical solution can effectively improve the efficiency and detection accuracy of detecting welding defects of a lithium battery top cover.

[0007] Furthermore, the training method of the welding defect detection model for the lithium battery top cover includes: collecting defect pictures of the lithium battery top cover with welding defects; generating corresponding label files according to the defect pictures to produce a welding defect data set; using the produced welding defect data set to train a deep neural network based on YoloV5 to obtain a welding defect detection model for the lithium battery top cover.

[0008] In the above technical solution, the welding defect detection model for the lithium battery top cover is trained by constructing a welding defect data set. Since the data in the data set is the data generated during the actual welding of the current welding equipment, training the model with this data makes the recognition result of the model more targeted and can effectively improve the accuracy of defect recognition.

[0009] Furthermore, generating corresponding label files according to the defect pictures includes: calibrating a picture containing defects, with the width of the picture being w and the height being h; determining the category of the defect and the boundary coordinates of the defect. Assuming the upper left coordinate of the picture is (x0, y0) and the lower right coordinate is (x1, y1); creating a label file image.txt with the same name as the picture using a text editor; writing in image.txt in the following format: i, a, b, c, d; where i is the category of the defect, represents the x coordinate of the center point of the defect, represents the y coordinate of the center point of the defect, represents the width of the defect, represents the height of the defect.

[0010] In the above technical solution, by optimizing the settings of the label file, the defect position can be accurately marked. The marking method meets the actual production needs, is more targeted, and is more helpful for improving the accuracy of defect recognition.

[0011] Furthermore, the welding defect detection result of the lithium battery top cover includes the category of the defect and the defect score.

[0012] In the above technical solution, by classifying the detection result into the category of the defect and the defect score, the harm degree of the defect can be effectively hierarchically divided, that is, it can be directly determined whether the defect affects the safe use of the lithium battery according to the score of the defect, realizing an objective evaluation of the welding defect of the lithium battery top cover.

[0013] In the second aspect, a welding defect detection device for the lithium battery top cover is provided, including: a data acquisition module for acquiring images of the lithium battery top cover; a defect detection module for inputting the acquired images of the lithium battery top cover into the trained welding defect detection model for the lithium battery top cover and outputting the welding defect detection result of the lithium battery top cover.

[0014] The above technical solution provides basic support for the industrial application of the battery top cover welding defect detection method through the data acquisition module and the defect detection module, and effectively improves the detection efficiency and detection accuracy of the lithium battery top cover welding defects.

[0015] In a third aspect, a lithium battery production system is provided, including: a defect detection unit, which is configured with the lithium battery top cover welding defect detection device described in the second aspect, and performs defect detection on the lithium battery top covers on the production line through the lithium battery top cover welding defect detection device, and outputs the lithium battery top cover welding defect detection results; a PLC control unit, which is used to transfer the corresponding lithium battery to the OK station or the NG station according to the received lithium battery top cover welding defect detection results.

[0016] The above technical solution realizes the industrial application of the lithium battery top cover welding defect detection through the defect detection unit and the PLC control unit, can eliminate unqualified lithium batteries from the production line according to the detection results, improves the production efficiency of lithium batteries, and improves the qualified rate of lithium batteries.

[0017] Further, the defect detection unit includes: a camera, an optical lens and a spherical light source mounted on a bracket, the camera is used to take pictures of the lithium battery top cover through the optical lens; the spherical light source is arranged between the optical lens and the lithium battery top cover.

[0018] The above technical solution collects images of the lithium battery top cover through the cooperation of the camera and the optical lens, and uses the spherical light source to improve the lighting conditions of the position where the lithium battery top cover is located, so that the collected images are clearer and can more accurately reflect the real situation of the lithium battery top cover, avoiding missed detection and misdetection.

[0019] Further, the defect detection unit further includes: an industrial computer communicatively connected to the camera, the industrial computer is used to receive, process, store the images collected by the camera and the processing results of the images, and output the processing results of the images to a display.

[0020] The above technical solution improves the data processing efficiency and the data accuracy through the industrial computer, and can save the detection data in real time to realize the traceability of the detection data; through the display, the current data processing process can be monitored in real time, which is convenient for management.

[0021] Further, the PLC control unit includes a flipping mechanism, the flipping mechanism is used to clamp and rotate the lithium battery, so that the weld seam of the lithium battery top cover passes through the image acquisition area of the camera in sequence according to a set route.

[0022] In the above technical solution, the PLC control unit controls the flipping mechanism, and the flipping mechanism enables the weld seam of the lithium battery top cover to pass through the image acquisition area of the camera in sequence according to the set route, improving the efficiency of defect detection, ensuring the consistency of data acquisition, liberating human resources, improving the detection efficiency, and ensuring the consistency of detection.

[0023] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: (1) By inputting the collected lithium battery top cover image into the lithium battery top cover welding defect detection model trained with the constructed welding defect data set, and outputting the lithium battery top cover welding defect detection result, the present invention can effectively improve the detection efficiency and detection accuracy of the lithium battery top cover welding defect. (2) Through the defect detection unit and the PLC control unit, the present invention realizes the industrial application of the lithium battery top cover welding defect detection, can remove unqualified lithium batteries from the production line according to the detection result, improves the production efficiency of lithium batteries, and improves the qualified rate of lithium batteries. Description of the Drawings

[0024] Figure 1 is a schematic diagram of the implementation process of a lithium battery top cover welding defect detection method provided by an embodiment of the present invention; Figure 2 is a schematic diagram of the structure of a lithium battery production system provided by an embodiment of the present invention; Figure 2 In the figure: 1. Camera; 2. Optical lens; 3. Spherical light source; 4. Lithium battery; 5. Lithium battery top cover; 6. Industrial control computer; 7. Display. Detailed Embodiments

[0025] The present invention will be further described below with reference to the drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention and should not be used to limit the protection scope of the present invention.

[0026] Embodiment 1 A lithium battery top cover welding defect detection method includes: collecting a lithium battery top cover image; inputting the collected lithium battery top cover image into a lithium battery top cover welding defect detection model trained with a constructed welding defect data set, and outputting a lithium battery top cover welding defect detection result.

[0027] The main detection process of the present invention is as Figure 1 shown: S1. Collect defect pictures during welding; S2. Generate a label file for each picture to make a welding defect data set; S3. Import the welding defect data set into a deep learning neural network model for training to obtain a lithium battery top cover welding defect detection model; S4. Use the detection model to detect the pictures of production summary to determine whether there are defects in the current lithium battery top cover; S5. Return the detection result to the PLC, and the PLC control mechanism places the corresponding lithium battery at the OK station or the NG station.

[0028] Generate a corresponding label file according to the defect picture, including: calibrating a picture containing defects, the width of the picture is w, and the height of the picture is h; determining the category of the defect and the boundary coordinates of the defect. Assume that the upper left corner coordinates of the picture are (x0, y0), and the lower right corner coordinates are (x1, y1); use a text editor to create a label file image.txt with the same name as the picture; in image.txt, write in the following format: i, a, b, c, d; where i is the category of the defect (for example, if there are 3 types of defects in total, then i is one of 0, 1, 2), represents the x coordinate of the center point of the defect, represents the y coordinate of the center point of the defect, represents the width of the defect, represents the height of the defect, separated by spaces in the middle.

[0029] Import the welding defect data set into the deep learning neural network model for training, including: t1. Collect defect pictures of the lithium battery top covers with welding defects; t2. Generate corresponding label files according to the defect pictures to make a welding defect data set; place the defect pictures and the corresponding label files in the images and labels folders respectively; divide the defect pictures and the corresponding label files, 70% for the training set, 20% for the validation set, and 10% for the test set; t3. Use the made welding defect data set to train a deep neural network based on YoloV5 to obtain a lithium battery top cover welding defect detection model. Use the validation set to adjust the hyperparameters of the model and evaluate the ability of the model, and use the test set to verify the generalization ability of the detection model. Finally, the training is completed to obtain the detection model.

[0030] As Figure 2 shown, the on-line detection end includes an industrial camera 1, an optical lens 2, a spherical light source 3, a fixed bracket, a light source controller, an industrial control computer 6, a display 7, etc. The PLC control end includes a gripper, a control cabinet, etc. The on-line detection end is installed at the station after battery welding. After the lithium battery top cover 5 is welded, the gripper grabs the lithium battery 4 with the welded lithium battery top cover 5 to the fixed fixture directly below the camera 1.

[0031] The camera 1, the optical lens 2 and the spherical light source 3 are installed on the bracket. The camera 1 is used to take pictures of the lithium battery top cover 5 through the optical lens 2; the spherical light source 3 is arranged between the optical lens 2 and the lithium battery top cover 5.

[0032] The entire detection process is as follows: a1. System initialization, connecting the PLC and the industrial camera; a2. The lithium battery is moved under the camera, and the camera takes pictures; a3. Load the trained lithium battery top cover welding defect detection model, detect each picture, and determine whether there are defects in the welding; a4. If there are no defects in the welding, it is released to the next station to continue detecting the next battery; a5. If there are defects in the welding, an NG signal is sent to the PLC to control the gripper to grab the defective lithium battery to the rejection station, and the picture of the defective battery is saved for easy viewing.

[0033] The detection result will output whether there are defects in each picture, and output the category and defect score of the defects; by setting an appropriate threshold, filtering according to the defect score, excluding a small part of the low-score results, so as to achieve the purpose of reducing the overkill rate.

[0034] Embodiment 2 Based on the lithium battery top cover welding defect detection method described in Embodiment 1, this embodiment provides a lithium battery top cover welding defect detection device, including: a data acquisition module for acquiring images of the lithium battery top cover; a defect detection module for inputting the acquired images of the lithium battery top cover into the lithium battery top cover welding defect detection model trained by using the constructed welding defect data set, and outputting the lithium battery top cover welding defect detection result.

[0035] Embodiment 3 Based on the lithium battery top cover welding defect detection method described in Embodiment 1, this embodiment provides a lithium battery production system, including: a defect detection unit configured with the lithium battery top cover welding defect detection device described in claim 5, and performing defect detection on the lithium battery top cover on the production line through the lithium battery top cover welding defect detection device, and outputting the lithium battery top cover welding defect detection result; a PLC control unit for transferring the corresponding lithium battery to the OK station or the NG station according to the received lithium battery top cover welding defect detection result.

[0036] As Figure 2 shown, the defect detection unit includes: a camera 1, an optical lens 2 and a spherical light source 3 mounted on a bracket. The camera 1 is used to take pictures of the lithium battery top cover 5 through the optical lens 2; the spherical light source 3 is arranged between the optical lens 2 and the lithium battery top cover 5, and the lithium battery top cover 5 is mounted on both sides of the lithium battery 4.

[0037] The defect detection unit further includes an industrial control computer 6 communicatively connected to the camera 1. The industrial control computer 6 is configured to receive, process, and store the images collected by the camera 1 and the processing results of the images, and output the processing results of the images to the display 7.

[0038] The PLC control unit includes a flipping mechanism, which is used to clamp and rotate the lithium battery so that the top cover weld of the lithium battery passes through the image acquisition area of the camera 1 in sequence along a set route.

[0039] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A method for detecting welding defects of a lithium battery top cover, characterized in that: include: Collect images of lithium battery top cover; The collected lithium battery top cover image is input into the lithium battery top cover welding defect detection model trained with the constructed welding defect dataset, and the lithium battery top cover welding defect detection result is output.

2. The lithium battery top cover welding defect detection method according to claim 1, characterized in that: The training method of the lithium battery top cover welding defect detection model comprises: Collect defective images of lithium battery top covers with welding defects; Generate corresponding label files based on defect images and create welding defect datasets; The prepared welding defect dataset is used to train a deep neural network based on YoloV5 to obtain a lithium battery top cover welding defect detection model.

3. The lithium battery top cover welding defect detection method according to claim 2, characterized in that: Generate corresponding label files based on defect images, including: Mark a picture with defects, with the width of the picture being w and the height of the picture being h; determine the category of the defect and the boundary coordinates of the defect, assuming that the coordinates of the upper left corner of the picture are (x0, y0) and the coordinates of the lower right corner are (x1, y1); use a text editor to create a label file image.txt with the same name as the picture; in image.txt, write in the following format: i, a, b, c, d; where i is the category of the defect, represents the x-coordinate of the defect center point, Indicates the y coordinate of the defect center point, represents the defect width, Indicates the defect height.

4. The lithium battery top cover welding defect detection method according to claim 3, characterized in that: The lithium battery top cover welding defect detection result includes the defect category and defect score.

5. A lithium battery top cover welding defect detection device, characterized in that: include: A data acquisition module, used to acquire images of lithium battery top covers; The defect detection module is used to input the collected lithium battery top cover image into the lithium battery top cover welding defect detection model trained by the constructed welding defect data set, and output the lithium battery top cover welding defect detection result.

6. A lithium battery production system, characterized in that: include: A defect detection unit, wherein the defect detection unit is equipped with the lithium battery top cover welding defect detection device according to claim 5, and the lithium battery top cover welding defect detection device is used to perform defect detection on the lithium battery top cover on the production line, and output the lithium battery top cover welding defect detection result; The PLC control unit is used to transfer the corresponding lithium battery to an OK station or an NG station according to the received lithium battery top cover welding defect detection result.

7. The lithium battery production system according to claim 6, characterized in that: The defect detection unit comprises: a camera (1), an optical lens (2) and a spherical light source (3) mounted on a bracket; the camera (1) is used to take a picture of a lithium battery top cover (5) through the optical lens (2); and the spherical light source (3) is arranged between the optical lens (2) and the lithium battery top cover (5).

8. The lithium battery production system according to claim 7, characterized in that: The defect detection unit further comprises: an industrial computer (6) connected to the camera (1) for communication, the industrial computer (6) being used to receive, process and store images captured by the camera (1) and image processing results, and output the image processing results to a display (7).

9. The lithium battery production system according to claim 6, characterized in that: The PLC control unit comprises a flipping mechanism, which is used to clamp and rotate the lithium battery so that the weld seam of the lithium battery top cover passes through the image acquisition area of ​​the camera (1) in sequence according to a set route.

Images