Virtual Two-Dimensional Electrode Representation Through the Spatial Transformation of Production Data

The production of battery cells is characterized by a multi-process production chain in which the individual processes significantly impact intermediate and product properties. A virtual representation of the produced electrode enables the analysis of cause-and-effect relationships and can be used for quality assurance. While the virtual representation of components or products is already common in other engineering domains, an approach to a holistic virtual representation of the electrode in battery cell production does not yet exist. A high level of data accuracy in multiple dimensions is required to reduce scrap within electrode production and automate processes. Therefore, this work presents a comprehensive data acquisition, preprocessing, and analysis methodology to get a multi-dimensional representation of the electrode properties and process parameters. Multiple data modeling options were analyzed to generate a uniform and complete database of all sensor data. Gaussian process regression and Kriging modeling were used to interpolate the mass loading of electrodes after the coating process so that the holistic virtual representation of the electrode was enabled. The use of industrial edge computing and interpolation of sensor data enables a more accurate, virtual representation of produced electrodes.