Modeling of zinc-air batteries using theoretical and empirical approach

Abstract

Nowadays, an energy storage system (ESS) is attracting widespread research interest. Zinc-air battery (ZAB) is a promising candidate for ESS owing to its high energy density at low cost. Although ZAB has been researched in various aspects, the modeling aspect is still insufficient. Therefore, this work aims to study the modeling of ZAB using both theoretical and empirical approaches. At first, the ZAB system was analyzed using a theoretical approach. The studied system was an integrated system of zinc-air flow battery (ZAFB) and zinc electrolyzer. The zero-dimensional mathematical model was developed in MATLAB and validated against experimental data. The model was used to investigate 3 parameters: potassium hydroxide (KOH) concentration, zincate ion initial concentration and electrolyte flowrate. Hydrogen evolution reaction (HER) was contained in the model therefore the current efficiency based on HER of the system can be calculated. The result showed that increasing KOH concentration improved the discharge energy of the battery, but it also promoted HER. An optimal KOH concentration of 6-7 M was obtained from the simulation result. Increasing zincate ion initial concentration improved the current efficiency of the system as it reduced HER. For electrolyte flowrate, the higher flowrate helped maintain the concentration of the active species in the battery; however, the higher flowrate also provided a negative effect on the battery performance. Next, ZAB dynamic behavior was investigated by empirical modeling. A linear parameter-varying model was proposed to examine the nonlinear behavior of ZAB. The LPV model was created from a set of linear time-invariant models. The data used to identify and validate the model in this study was measured from a homemade refuellable ZAB. As a result, it was found that the LPV model was able to predict the nonlinear behavior of the battery and its performance was comparable to the nonlinear model. Finally, the SOC estimation of ZAB was studied. The LPV model integrated with the extended Kalman filter algorithm was proposed as SOC estimator. The data used to test in this case were obtained from a laboratory-made tri-electrode ZAFB. The tested scenarios include varied tuning parameters and the correctness of the initial SOC guess. The result revealed that the developed estimator was able to estimate true SOC value only when SOC approached zero due to the flat voltage profile of the flow battery. Nevertheless, the SOC estimator was capable of true SOC tracking in the long-term operation with multiple charge-discharge cycles. The results of this research provide a better comprehension and extended knowledge boundary of modeling of ZAB.