How can we ramp up battery production?

Machine vision can be used in almost all industries, but the major added values of the technology are particularly clear when it comes to battery production. Whether for mobiles, laptops, consumer electronics, or e-mobility, global demand for batteries is rapidly growing. Speed, efficiency, and consistently high quality are therefore all important factors for production. Machine vision offers valuable support in this regard – along every part of the production chain.

Klaus Schrenker, Business Development Manager at MVTec

Whatever their intended purpose, the production of battery cells can be divided into three key process steps: electrode production, cell assembly, and formation and aging. Thanks to machine vision, the processes involved in these steps can be automated and optimized. As such, the technology paves the way for precise workflows and high quality – coupled with increased production speed. This is made possible by sophisticated and robust defect detection methods, which enable appropriate post-processing. The result is greater customer satisfaction, less waste, and fewer resources needed for the manufacturing process.

Machine vision enables precise electrode production

Let’s start with the first step in battery cell production: electrode production. First, both sides of a copper or aluminum foil are coated with an active material. It is important for this coating to be applied precisely, as even the tiniest of defects or impurities would severely impair the quality and performance of the entire battery.

This is where machine vision comes in: The technology inspects the surface, the dimensional accuracy, and the alignment of the coated electrode strips, ensuring optimum quality. It also checks that the coating is uniform, and the electrodes constantly have the correct width. Even any coating residues on the edges can be reliably detected.

The next stage involves adapting the coated electrode foil to the battery cell format. To achieve the dimensions required for the subsequent cell assembly, the foil is divided into “electrode sheets”. When doing this, machine vision algorithms again ensure the highest level of precision and meticulously check the dimensions. The software detects any anomalies such as burrs, impurities, cracks, or unclean cuts. Even unknown defects or defects that are hard to define can be found. When performing surface inspections of the coated electrode foil, suitable lighting even enables machine vision software to reliably detect defects under difficult contrast or lighting conditions.

Cell assembly requires precise electrode foil positioning

Electrode production is followed by cell assembly. The specifics of this differ depending on the cell type: if the electrode foil, comprising an anode, cathode, and separator, is stacked to form prismatic or pouch cells, or winded in the case of cylindrical cells. This way, these components are precisely integrated into the battery housing. After that, the housing is welded and filled with electrolyte through an opening before the cell is completely sealed. Especially when producing prismatic and pouch cells, the stacked electrode sheets have to be aligned with exceptional precision and without any damage.

Reliably avoiding errors during the stacking process

Machine vision again provides support here in the form of sophisticated control mechanisms throughout the entire stacking process. This runs at a high speed of up to one second per sheet, which increases the complexity of the inspection process. Thanks to machine vision software, errors can be reliably avoided when inspecting electrode surfaces and contact terminals, as well as when measuring the cut geometry. Both AI-based deep learning algorithms and classic, rule-based machine vision methods are used. Both approaches are able to detect the widest range of defects that are invisible to the human eye, even in low resolution images. Furthermore, machine vision helps to precisely identify the position of the electrode sheets and to correctly align the stacks. Other potential errors during the stacking process, such as scratched electrode surfaces, contact terminal defects, and improperly positioned anode and cathode foils, are also detected. With the latter, it is important that both are cleanly separated by a separator.

The final step in the cell assembly process is to insert the stack into the cell housing. For the battery to be fully sealed, the seams must be correctly welded. Various machine vision processes can be used to reliably inspect the weld seams and detect potential errors. The technology therefore enables precise welding processes, which are of central importance to the quality of the battery.

A trained deep learning network is able to check the quality of the weld seam and enables non-destructive inline inspection.

Comprehensive tests ensure high-quality battery cells

The third and final process step when producing battery cells is that of formation and aging. During this step, the batteries are charged,  discharged, and subjected to a comprehensive quality inspection. Multiple end-of-line tests, such as visual inspections and surface measurements, are used to ensure optimum product quality. Machine vision again provides valuable support here, for instance by detecting deformed cells, e.g., cylindrical cells that do not have the specified cell diameter or that have other surface damage. This makes it possible to reliably remove any defective cells before they leave the factory and are delivered to customers.

Machine vision is future-proofing the battery sector

To meet rising market demand, batteries have to be produced with ever greater speed and efficiency. Machine vision plays a decisive role in this regard. It helps to optimize and automate the complex battery production process while also ensuring consistently high product quality.