Analysis of the propagated uncertainty in a identification algorithm applied to a Lead-Acid battery

Individual monitoring of the battery state of charge is especially interesting in applications where several units are used for powering complex systems. In these applications the identification of failures or evidences of failures makes the control more robust, featuring a predictive maintenance of the power system. This article shows the results of a developed system for monitoring parameters that characterize the state of charge of batteries. The system consists of a dedicated hardware and specialized software used to apply a known input signal and measure its response in order to identify the point of minimum-phase (in frequency) of the equivalent impedance of the battery. The work focuses on the analysis of the propagated uncertainty of input variables modifying the final result of frequency that has been experimentally related to the state of charge of a battery type. Among the preliminary results are presented the sensitivities of major uncertainty in the input output beyond the estimated uncertainty of the state of battery charge.