Online Identification of VLRA Battery Model Parameters Using Electrochemical Impedance Spectroscopy

This paper introduces the use of a new low-computation cost algorithm combining neural networks with Nelder–Mead simplex method to monitor variations parameters previously selected equivalent circuit calculated from Electrochemical Impedance Spectroscopy (EIS) corresponding series battery aging experiments. These could be correlated in state over time and, therefore, identify or predict degradation patterns failure modes. The authors have benchmarked four different Electrical Equivalent Circuit (EEC) parameter identification algorithms: plain network mapping EIS raw data EEC parameters, Particle Swarm Optimization, Zview, and proposed one. In order improve prediction accuracy network, augmentation has been training error. algorithms compared validated through real obtained six-month test experiment carried out set six commercial 80 Ah VLRA batteries under cycling temperature operation conditions.