Assessing nonlinearity in impedance models

Abstract A procedure for assessing nonlinearity in impedance or admittance models which is based on modern nonlinear regression analysis is described. The nonlinearity of a model can be separated into two components: (1) “intrinsic” nonlinearity and (2) “parametric” nonlinearity. The curvature measures of nonlinearity performed by Bates and Watts in 1980 are objective measures of the intrinsic and parametric nonlinearities in a model-data set combination. The methods of quantifying nonlinear behavior provide a standard by which the performance of a model-data set combination can be judged. The nonlinearity of the Randles equivalent circuit is discussed as an example of the general problem.