The basis for the performance of the battery cell is already laid during production. Therefore, detailed knowledge of the interaction between process parameters and (intermediate) product characteristics is indispensable. In order to bring these data into a usable context, each value (discrete or continuous) must be uniquely assigned, depending on the type of process, also retrospectively (traceability at process level).
An initial implementation and testing was carried out on a continuous electrode coating process. The recorded data sets were first assigned to the continuously running electrode in order to be able to assign them to individual cells at a later time. In an ongoing project an adapted process FMEA (Failure Mode and Effects Analysis) was used to determine relevant process parameters. Furthermore, relevant disturbance variables were identified within the FMEA. The process parameters were, for example, the wet and dry film thickness of the coating, coating velocity, relative humidity and temperature at different positions in the process, oven temperature stages and several machine parameters. Furthermore, intermediate product characteristics such as the layer thickness before and after drying were determined. In addition, disturbance variables such as relative humidity, ambient temperature or vibrations in the process were recorded and stored. The different data sets are available both discretely and continuously. The challenges was to consolidate the data and implement them in a separate database. Therefore a specific algorithm for space-time resolution was developed.
Regarding to the consistency and plausibility the recorded data are preprocessed using different test algorithms. These algorithms are based on physical limits, statistical tests (e.g. outlier test) and expert knowledge. Using these algorithm it is possible to identify process deviations and detect imperfect areas at an early stage. Continuously applied markers on the current collector linked to the database ensure traceability at all times.