Anode-Cathode Alignment in Winding and Laminating: 3D X-ray CT Combined with AI Layer-by-Layer Analysis Speeds up by about 90 Times

The anode-cathode alignment degree is crucial for lithium battery safety. Traditional analysis methods have problems such as low efficiency and strong subjectivity. WeLinkirt's DaoAI solution provides an effective way to solve these problems.

Scene: In the current booming new energy battery industry, lithium batteries, as the core energy storage devices, are widely used in electric vehicles, portable electronic devices and other fields. Their safety and performance directly affect the quality of end-products and the user experience. The anode-cathode alignment is the internal bottom line for lithium battery safety. Whether it is the winding process or the laminating process, the relative position between the anode and the cathode must meet the requirement that the anode completely covers the cathode. Once the alignment degree is insufficient, lithium deposition is likely to occur in the exposed area of the cathode edge during charging. Long - term accumulation can pierce the separator and induce an internal short-circuit. Situations such as pole piece bending and misalignment can also change the internal stress distribution, seriously affecting the safety and service life of the battery.

Pain Points: Why is it Difficult?

First, from the time dimension, a leading power battery enterprise previously relied on engineers to manually read CT slices layer by layer. Since the battery cells often have dozens of layers, manually judging the alignment margin layer by layer is both slow and prone to fatigue. The analysis of a single battery cell takes up to hours, which is extremely inefficient for large-scale production lines and cannot meet the sampling scale at the production line level. For example, if a factory produces thousands of battery cells a day and each cell analysis takes several hours, the number of cells that can be analyzed in a day is very limited, seriously affecting the production progress and the timeliness of quality assessment.

Secondly, from the accuracy dimension, manual reading of slices is highly subjective. Different engineers may have different experiences and judgment standards, which makes it difficult to ensure the consistency of the alignment analysis results. For example, for some pole pieces with fuzzy edges, different engineers may give different measurement results of the alignment margin, thus affecting the accurate assessment of battery quality. Moreover, manual reading is prone to visual fatigue, and the probability of missed detection and misjudgment will increase significantly after long-term work.

Finally, from the cost dimension, manual analysis requires a large amount of labor cost. Engineers need to spend a lot of time reading slices and analyzing, and due to low efficiency, more engineers are needed to complete the corresponding work tasks, which undoubtedly increases the operating cost of the enterprise. At the same time, due to the accuracy and efficiency problems of manual analysis, some batteries with quality problems may flow into the market, bringing potential economic losses and brand risks to the enterprise.

Technical Principle

The DaoAI solution combines 3D X-ray CT with AI layer - by - layer analysis. The 3D X-ray CT technology uses X-rays to penetrate the battery cell, obtains the three-dimensional information inside the cell through multi-angle scanning, and then performs image reconstruction, thus presenting the three-dimensional structure inside the cell non-destructively. This imaging method can clearly show the position and shape of the anode and cathode, providing an accurate data basis for subsequent analysis. Compared with the traditional two-dimensional imaging technology, 3D X-ray CT can provide more comprehensive and accurate internal structure information, avoiding the problems of information overlap and loss in two-dimensional images.

After obtaining the three-dimensional structure inside the battery cell, the AI algorithm comes into play. The AI can automatically locate the edges of the anode and cathode layer by layer. Using image recognition and deep learning technology, it analyzes the image of each layer and accurately finds the boundaries of the anode and cathode. Then, the AI quantifies the alignment margin, calculates the distance between the anode and the cathode, and identifies abnormalities such as bending and misalignment. Through learning and training on a large amount of data, the AI algorithm can complete these tasks quickly and accurately, and has high consistency. Compared with the traditional manual slice-reading method, AI analysis is not affected by subjective factors, with higher efficiency and better accuracy.

Typical Application Scenarios

• Anode - cathode alignment detection in the winding process: During the winding process, the alignment degree of the anode and cathode is easily affected by various factors, such as winding tension and the flatness of the pole pieces. Through 3D X-ray CT imaging, the AI can detect the alignment margin of the anode and cathode layer by layer and timely find the problem of insufficient alignment. The difficulty lies in the complex shape of the pole pieces during the winding process. The AI needs to accurately identify the boundaries of different layers of pole pieces to avoid misjudgment.
• Anode - cathode alignment detection in the laminating process: The laminating process requires higher alignment accuracy for each layer of pole pieces. The AI can perform layer - by - layer analysis on the laminated battery cell, quantify the alignment margin, and detect whether there is misalignment of the pole pieces. The difficulty lies in the large number of laminated layers and rich image information. The AI needs to process a large amount of data quickly while ensuring the accuracy of the analysis.
• Pole piece bending detection: Pole piece bending can change the internal stress distribution of the battery, affecting the performance and safety of the battery. 3D X-ray CT can clearly show the bending situation of the pole pieces, and the AI can automatically identify the position and degree of bending. The difficulty lies in the diverse bending forms. The AI needs to accurately classify and judge different types of bending.
• Internal short-circuit hidden danger detection: Insufficient anode-cathode alignment may lead to an internal short-circuit. The AI can predict the potential risk of an internal short-circuit by analyzing the relative position and spacing between the anode and cathode. The difficulty lies in the hidden manifestation of internal short-circuit hidden dangers. The AI needs to extract key information from a large amount of image data to accurately judge the existence of hidden dangers.

Implementation Case

A large-scale power battery enterprise produces thousands of battery cells per day. Before introducing the DaoAI solution, the enterprise used the method of manually reading CT slices layer by layer for anode-cathode alignment analysis. The analysis of a single battery cell took about 3 hours, and due to strong subjectivity, the missed detection rate reached 10% and the false alarm rate was 8%. When the DaoAI solution was launched, the technical team of WeLinkirt first conducted a detailed investigation and analysis of the enterprise's production process and data, and then carried out customized development and debugging of the DaoAI system according to the actual needs of the enterprise. After a period of testing and optimization, the DaoAI solution was officially launched.

The launch of the DaoAI solution has greatly improved the anode-cathode alignment analysis efficiency of the enterprise.

WeLinkirt's Solution and Product

WeLinkirt's DaoAI product perfectly combines 3D X-ray CT with AI layer - by - layer analysis. 3D X-ray CT non-destructively reconstructs the three-dimensional structure inside the battery cell, covering the winding and laminating processes, and provides accurate basic data for AI analysis. The AI automatically locates the edges of the anode and cathode layer by layer, quantifies the alignment margin, and identifies internal structure abnormalities such as bending and misalignment, and automatically summarizes the alignment distribution and extreme values of each battery cell. Engineers change from manual layer - by - layer measurement to the review and judgment of AI results, greatly reducing the workload and improving the analysis throughput.

Quantitative Results: After the implementation of the solution, the analysis efficiency of anode-cathode alignment and bending detection has been increased by about 90 times compared with manual layer - by - layer slice reading. The analysis of a single battery cell has been compressed from hours-level manual analysis to minutes-level, specifically from about 3 hours to about 2 minutes. The sampling scale of CT has been expanded accordingly, and the evaluation of internal alignment quality has shifted from experience-based to data-based and traceable. The missed detection rate has been reduced from 10% to 1%, and the false alarm rate has been reduced from 8% to 2%, effectively improving the accuracy and reliability of battery quality detection.

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.