Equivalent Circuit Modelling of Hybrid Supercapacitors Through Experimental Spectroscopic Measurements

In this paper, a strategy for the identification of models based on supercapacitors equivalent circuits is proposed. The approach is based on an innovative and efficient encoding procedure, suitable for a generic impedance network. The latter, in combination with an optimization algorithm, is used to investigate the components and topology of the best-fitting network for a given dataset. To demonstrate the generality of the proposed approach, equivalent circuits involving both integer and fractional order elements have been derived. The identification of each model is performed using a cascade optimization algorithm for maximum robustness against local minimum entrapment. The approach is applied to the identification of a 4000 F hybrid supercapacitor, for which data is acquired from an extensive experimental spectroscopy campaign at different states of charge and temperatures. Two models are selected, one for each set, both in excellent agreement with the experimental measurements.