Development of an aging model to be implemented in a Smart Battery Management System (RE 33/25)

This Deliverable D2.2 introduces the aging model developed based on the outcomes of the feature selection process. This process was defined and implemented as part of Task 2.3, “Development of Feature Selection Algorithm,” and further refined through the activities carried out in Task 2.4, “Aging Model Development and Testing.” The construction of the aging prediction method involved two key phases: Definition and Validation. During the Definition phase, the most suitable algorithm was identified using a subset of the available dataset. In the subsequent Validation phase, the method’s applicability was tested on the remaining portion of the selected dataset. By applying the Differential Voltage technique to charge curves at 0.5 C, the selected features enabled the prediction not only of the cell’s State of Health (SOH), but also of the specific degradation mechanisms affecting the cell under analysis. Throughout the code development process, various techniques were compared to achieve the optimal balance between prediction accuracy and computational efficiency.