Statistical Identifiability of AgingModes in Lithium-ion Batteriesfor Electric Vehicles

Lithium-ion batteries (LIBs) are essential for powering heavy electric vehicles due to their high energy density and efficiency. However, their performance declines over time due to complex aging processes. Accurately understanding and quantifying these effects is crucial for reliable lifetime prediction. This thesis focuses on identifying key aging parameters in LIBs using a model based on Open Circuit Voltage (OCV) data and discusses the sensitivity over different charging ranges. An existing OCV-based model is used to estimate and optimize parameters such as the lithium capacity ratio between electrodes, initial lithium content in the positive electrode, and voltage limits. These are evaluated using capacity-voltage data from both full and partial state-of-charge (SOC) cycles. To assess how well these parameters can be identified, a sensitivity analysis is performed using the Fisher Information Matrix (FIM) and the Cramér-Rao Lower Bound (CRLB) is used to quantify estimation uncertainty.The study also investigates how good parameter identifiability can be achieved even with incomplete OCV data such as the range accessible during typical vehicle operation. It emphasizes the importance of including data from the low SOC region, where the voltage response is particularly informative. These insights contribute significantly to enhancing battery health diagnostics and lifetime modeling for heavy-duty electric vehicle applications, paving the way for more reliable and efficient energy solutions.