State of Safety Assessment in Electric Vehicle Battery Packs

Abstract

Ensuring safety and guiding first responders in risk mitigation decisions necessitates an accurate evaluation of the State of Safety (SoS) for a damaged electric vehicle (EV). This paper introduces a method for rapidly and precisely assessing the SoS of lithium-ion battery packs that may have been damaged in an accident. Mean-based voltage and temperature residuals are produced for cell groups and deviations are calculated and statistically evaluated. Synthetic anomalous data is generated by injecting model-based anomalies into experimental data to simulate EV accident scenarios. The proposed post-crash safety assessment algorithm is accurate with a low false positive and false negative rates.