Feature Identification and Extraction for Battery Aging Estimation in Aircraft Auxiliary Applications

Aircraft applications are increasingly including energy storage, especially batteries, as part of their main and auxiliary power systems. Considering the auxiliary power one, this paper studies the impact of temperature on the identification and extraction of battery features, and their adequacy to be used as indicator to estimate the battery State-of-Health (SOH). For that purpose, four different features were extracted for 6 cells cycled under aircraft auxiliary loads at 2°C and 55°C and were fed into Support Vector Regression (SVR) algorithms, and optimized versions, which are particularly useful for nonlinear applications. Results show the temperature greatly affects the data, requiring more advanced SVR algorithms for weaker features.