Battery parameter estimation from recorded fleet data

Artikel i vetenskaplig tidskrift, 2016

A method to identify battery parameters for a Li-ION hybrid battery model based on current and voltage measurements performed in a vehicle during 12 months of in use operation is investigated. This is different from previous work where a common approach is to use current pulse de-rating tests, HPPC, hybrid pulse-power capability, which uses predetermined amplitudes and frequencies of current pulses. Equivalent circuit linear models of different complexity were tested and evaluated in order to identify parameter dependencies at different state of charge levels and temperatures. The average accuracy of modelling the DC bus voltage provides a model goodness average higher than 75% for all analyzed individual log files. The models are used in order to estimate energy output from the battery and compared towards how the estimated voltage error propagates throughout a drive cycle. Both single RC equivalent circuit model and R-equivalent circuit model experienced goodness of fit at average 75 % and dual RC equivalent circuit model was less accurate with average 40 %.