Health diagnosis and remaining useful life prognostics of lithium-ion batteries using data-driven methods

Abstract The accurate estimation of state of health (SOH) and a reliable prediction of the remaining useful life (RUL) of Lithium-ion (Li-ion) batteries in hybrid and electrical vehicles are indispensable for safe and lifetime-optimized operation. The SOH is indicated by internal battery parameters like the actual capacity value. Furthermore, this value changes within the battery lifetime, so it has to be monitored on-board the vehicle. In this contribution, a new data-driven approach for embedding diagnosis and prognostics of battery health in alternative power trains is proposed. For the estimation of SOH and RUL, the support vector machine (SVM) as a well-known machine learning method is used. As the estimation of SOH and RUL is highly influenced by environmental and load conditions, the SVM is combined with a new method for training and testing data processing based on load collectives. For this approach, an intensive measurement investigation was carried out on Li-ion power-cells aged to different degrees ensuring a large amount of data.