Investigation of the Role of Battery Resistance and Overpotential on Performance Models, Power Capability Prognostics, and Aging Behavior

This thesis investigates the behavior of battery resistance and overpotential for the purpose of modeling power capability prognostics and battery aging behavior. For a lead-acid battery, the traditional linear circuit model is inadequate to capture the battery electrode overpotential behavior described by the nonlinear Butler-Volmer equation. A discrete model that incorporates the Butler-Volmer nonlinear behavior is introduced, and its recursive form is developed for online battery monitoring in a Corbin Sparrow electric vehicle. For battery power capability prognostics, the commonly defined State-of-Power has been found to have a high variability within the context of recursive estimation. An equivalent State-of-Function metric, suitable even for nonlinear battery model forms, is proposed with the confidence interval provided by Kalman filter estimation. For lithium batteries, it has been found that, while the Butler-Volmer nonlinear behavior is approximately linear at room temperature, the nonlinear behavior manifests itself at lower temperatures. A discrete battery model with temperature explicitly built into it has been proposed. This discrete battery model with temperature as an input has been adopted in recursive form for online battery power prognostics and State-of-Charge estimation. In addition, an investigation has been conducted into the aging caused by superimposing an AC current signal onto the discharge current waveform, using measured resistance as the battery aging metric. It has been found that, by increasing the RMS value of the discharge waveform, the superimposed AC signal causes statistically significant aging acceleration.