As the core product in the modern energy field, the quality of lithium-batteries is directly related to the safety and reliability of many application scenarios. Electrode coating, as the initial process of lithium-battery manufacturing, plays a decisive role in the performance of subsequent battery cells. DaoAI's 3D AI-AOI online full-inspection solution provides an effective way to solve the quality inspection problems of electrode coating.
Industry background and user scenarios: Against the backdrop of the booming development of new-energy vehicles, energy storage and other fields, the market demand for lithium-batteries has shown an explosive growth. The performance and quality of lithium-batteries directly affect the safety and stability of these applications. Therefore, every link in the manufacturing process is crucial. Electrode coating, as the first key process in lithium-battery manufacturing, involves uniformly applying slurry with a micron-level thickness on copper or aluminum foils. The uniformity of the coating not only directly determines the capacity of subsequent battery cells but also has a profound impact on battery safety. For lithium-battery manufacturers, ensuring the quality of electrode coating is the foundation for guaranteeing the smooth progress of the entire production process and the quality of the final product.
Deep - digging of pain points: Why is it difficult?
From a quantitative perspective, traditional quality-inspection methods face many challenges. In terms of detection accuracy, manual light inspection and traditional 2D vision can only detect obvious apparent abnormalities. They can hardly detect subtle defects at the 50μm level, such as dark spots and material loss. In terms of false-alarm rate, due to the high speed and wide width of the coating line and the metal reflection on the copper foil surface, the false-alarm rate of traditional AOI systems in this scenario can be as high as 30% -40%. In terms of missed-detection rate, because traditional methods have difficulty detecting slight fluctuations in the thickness direction and low-contrast dark spots, the missed-detection rate can reach 10% -15%.
The root causes of these problems lie in the limitations of traditional quality-inspection methods. Manual light inspection is limited by human vision and fatigue. It is difficult for people to maintain a high level of concentration for a long time, and subjective errors are prone to occur in the judgment of minor defects. Traditional 2D vision can only obtain two-dimensional image information on the object surface, and it is unable to accurately quantify the thickness change of the coating. It is even more powerless in detecting defects hidden in the thickness direction. In the complex scenario of high-speed, wide-width and metal reflection, traditional AOI systems are easily interfered by environmental factors, resulting in frequent false alarms and missed detections.
Technical principle
DaoAI's 3D AI-AOI online full-inspection system adopts an advanced technical architecture. The core is a self-developed 3D camera, which collects point-cloud data on the coating surface. The point-cloud data can accurately reflect the three-dimensional shape and thickness information of the coating surface. At the same time, the system is also equipped with a high-resolution 2D camera for obtaining two-dimensional images of the coating surface. By fusing the point-cloud data with the 2D images, the system can achieve all-around detection of the coating.
At the algorithm level, the system uses the APDT positive-sample learning algorithm. This algorithm can establish criteria only with qualified coating samples, without the need for a large number of defect samples. This enables the system to quickly adapt to the color difference and reflection changes of new batches of slurry, shortening the cycle of model change and new-batch launch. Compared with traditional methods, which require a large number of defect samples for training and are difficult to adapt to changes in the production process, DaoAI's solution has stronger adaptability and flexibility.
Typical application scenarios
- Dark - spot detection: Dark spots are usually caused by uneven slurry distribution or impurities in the coating. In the detection of dark spots, 2D high-resolution imaging plays a major role. It can clearly capture the subtle color changes on the coating surface, thereby identifying dark-spot defects. The difficulty lies in the low contrast of dark spots, which are easily overlooked. High - precision imaging technology and advanced image-recognition algorithms are required.
- Scratch detection: Scratches may be caused by equipment or foreign objects during the coating process. 2D imaging can intuitively detect the existence of scratches. By analyzing the features of lines in the image, the position and length of scratches can be determined. The difficulty lies in the fact that some subtle scratches are difficult to clearly show under the influence of surface reflection, and image enhancement processing is required.
- Foil exposure detection: Foil exposure refers to the situation where the coating fails to completely cover the copper or aluminum foil surface. The 2D image can clearly show the exposed part of the foil, so as to determine whether there is a foil-exposure defect. The difficulty lies in accurately distinguishing the normal coating edge from the foil-exposure area, and a precise image-segmentation algorithm is required.
- Local material-loss detection: Local material loss will lead to uneven coating thickness. The 3D channel can accurately quantify the degree of local material loss through point-cloud reconstruction of the coating thickness fluctuation. The difficulty lies in accurately extracting the features of the material-loss area from complex point-cloud data, which requires powerful data-analysis and processing capabilities.
- Edge material-loss detection: Edge material loss usually occurs at the edge of the coating. The 3D channel can clearly identify the thickness change at the edge using point-cloud data, thereby detecting edge material-loss defects. The difficulty lies in that the point-cloud data in the edge area is easily interfered with, and filtering and correction processing of the data are required.
Implementation case
A well-known domestic leading enterprise in the power-battery field has a large-scale production and extremely high requirements for the quality of electrode coating. Previously, the enterprise relied on manual light inspection and traditional 2D vision for quality inspection, facing high false-alarm and missed-detection rates, which seriously affected production efficiency and product quality. When introducing DaoAI's 3D AI-AOI online full-inspection system, DaoAI's technical team first conducted a detailed investigation and analysis of the enterprise's production environment and process, and carried out customized deployment of the system according to the actual situation. During the implementation process, professional training was provided to the enterprise's employees to ensure that they could operate and maintain the system proficiently.
DaoAI's solution has brought significant changes to the enterprise, greatly improving the quality and production efficiency of electrode coating.
DaoAI's solution and product
DaoAI's 3D AI-AOI online full-inspection system is a comprehensive solution. The system takes a self-developed 3D camera as the core and combines it with a high-resolution 2D camera to achieve the fusion of point-cloud and 2D images. In terms of software, the APDT positive-sample learning algorithm is used, which can quickly and accurately detect various defects. The system is highly automated and intelligent, capable of real-time online full inspection, and its inspection rhythm fully matches the high-speed coating production line. It can also be customized according to the production needs and process characteristics of different enterprises, providing personalized quality-inspection solutions for enterprises.
Quantitative results
After the implementation of the solution, significant quantitative results have been achieved. In terms of defect detection, the coating process can stably detect dark spots and material-loss defects better than 50μm, and the comprehensive accuracy remains above 97%. In terms of false-alarm rate, compared with traditional methods, the false-alarm rate has been reduced by 30% -40%. In terms of missed-detection rate, the missed-detection rate has been reduced from the original 10% -15% to less than 3%. The online full inspection has replaced the original sampling inspection and manual light inspection, significantly reducing the outflow of defects, providing a more consistent source of electrodes for subsequent winding and formation processes, and effectively improving the quality and efficiency of the entire lithium-battery production process.
FAQ
What problems in lithium-battery electrode coating can DaoAI's 3D AI-AOI online full-inspection solution solve?
DaoAI's solution can intercept thickness abnormalities and surface defects at the source, solving the problems that traditional methods have difficulty in detecting slight fluctuations and dark spots. It can effectively suppress false alarms, reducing the false-alarm rate by 30% -40% and the missed-detection rate to less than 3%. It significantly reduces the outflow of defects and provides highly consistent electrodes for subsequent processes.
What are the features of DaoAI's 3D AI-AOI online full-inspection system?
The system uses a self-developed 3D camera to fuse point-cloud data with 2D images. The 2D channel checks apparent defects, and the 3D channel quantifies the thickness. It uses the APDT positive-sample learning algorithm, does not require a large number of defect samples, can adapt to new batches of slurry, and its full-inspection rhythm matches the high-speed production line. It can be customized according to enterprise requirements.
What is the effect after the implementation of DaoAI's solution?
After the implementation, the coating process can stably detect dark spots and material-loss defects better than 50μm, with a comprehensive accuracy of over 97%. It replaces the original sampling inspection and manual light inspection, significantly reducing the outflow of defects and improving the consistency of electrodes in subsequent processes and production efficiency.
This article was generated by AI. Customer cases are simulated scenarios based on real product capabilities and figures are illustrative; see product pages for official benchmarks.