Periodic Optimal Input Shaping for Maximizing Lithium-Sulfur Battery Parameter Identifiability

This article investigates the problem of optimal periodic cycling for maximizing the identifiability of the unknown parameters of a Lithium-Sulfur (Li-S) battery model, including estimates of the initial values of species masses. This research is motivated by the need for more accurate Li-S battery modeling and diagnostics. Li-S batteries offer higher energy density levels compared to more traditional lithium-ion batteries, making them an attractive option for energy storage applications. However, the monitoring and control of Li-S batteries are challenging because of the complexity of the underlying multistep reaction chain. The existing literature addresses poor battery parameter identifiability through a variety of tools, including optimal input shaping for Fisher information maximization. However, this literature's focus is predominantly on the identifiability of lithium-ion battery model parameters. The main purpose of this study is to optimize Li-S battery Fisher identifiability through optimal input shaping. The study shows that such optimal input shaping indeed improves the accuracy of Li-S parameter estimation significantly. This outcome is demonstrated in simulation. Moreover, an experimental study is conducted showing that the underlying battery model fits laboratory experimental cycling data reasonably well when the optimized test cycle is employed.