Quality Challenges Driven by Energy and AI Applications
With the rapid growth of new energy and AI applications, battery cell quality inspection has become one of the key technologies affecting industry competitiveness. The exceptional compound annual growth rate (CAGR) of the global battery market demands that battery manufacturers adopt more efficient and precise production models. However, traditional manual inspection methods cannot effectively meet the high-speed, high-volume, and fine-grained quality control requirements, especially when detecting microscopic defects or inspecting internal structures of battery cells. Portwell has joined forces with AI vision solution expert Neurocle to offer a highly competitive smart inspection solution by combining leading AI deep-learning vision inspection software with the high-performance PJAI-200 edge computing system.
Pain Point: Limitations of Traditional Inspection Methods
Defect detection in the battery cell production process is highly complex. It must identify tiny external flaws, such as scratches and dents, and inspect internal structures, like cell alignment and foreign object detection. Traditional methods often rely on manual visual inspection or low-level automated systems, which struggle to achieve consistently high accuracy. Human factors and system stability issues can lead to inconsistent quality, adversely affecting product reliability and production efficiency.
Synergistic Advantages of NEURO-T and NEURO-R
To address these industry challenges, Portwell and Neurocle collaborate by integrating NEURO-T and NEURO-R, responsible for AI model training and real-time edge inference, respectively, providing a complete and highly customizable solution. NEURO-T software employs a no-code, graphical user interface (GUI), enabling users without deep AI expertise to train highly accurate defect detection models. This significantly lowers the barrier to AI adoption and shortens training time. NEURO-R ensures that the trained models can perform efficient inference on edge devices, supporting real-time, large-scale production-line quality inspection requirements.
High-Performance PJAI-200 Edge Computing System
Portwell’s PJAI-200 edge computing system, equipped with the NVIDIA Jetson AGX Orin module, delivers up to 200 TOPS of AI performance. It features 32 GB of 256-bit LPDDR5 memory, ensuring fast and stable execution for complex model inference. In addition, the PJAI-200 offers a rich set of interfaces—including 12 PoE GbE ports, one 10 GbE port, and eight USB 3.2 Type-A ports—allowing easy integration with multiple industrial cameras and sensors to realize a comprehensive vision inspection solution.
Precise Inspection for All Battery Formats
This application is especially suitable for surface and internal defect inspection of pouch-type batteries. The OCR model recognizes text and codes on the battery surface, while the segmentation model accurately identifies minute defects such as scratches and dents, quantitatively measuring defect shape and size. The solution also applies to cylindrical and prismatic cells: through internal X-ray image analysis, it precisely locates and identifies cell alignment and foreign objects, enhancing product safety and quality consistency.
Enhancing Yield and Reducing Costs
By deploying the smart inspection system jointly developed by Portwell and Neurocle, battery manufacturers can effectively boost product yield and significantly reduce risks and costs associated with human factors and equipment stability, ensuring an efficient and continuously stable production environment. Meanwhile, the PJAI-200, with its high computing performance and flexibility, adapts to various battery inspection scenarios, helping enterprises rapidly adjust and scale production capacity to meet ever-changing market demands.
A Key Tool for Industry Upgrades
In the fast-developing new energy battery industry, precise and real-time quality inspection systems have become an indispensable competitive advantage. The smart inspection solution launched by Portwell and Neurocle, through the highly user-friendly AI model training software NEURO-T and the high-performance edge inference platforms NEURO-R and PJAI-200, fully overcomes the limitations of traditional inspection systems, delivering outstanding accuracy and stability. Enterprises adopting this solution can not only quickly improve production quality and efficiency but also effectively reduce operating costs. Supported by Portwell’s global DMS (Design and Manufacturing Services) network and robust ecosystem, they can achieve industry upgrades and business expansion.