For power tool applications such as chainsaws, both runtime and continuous power must be maximized.
This work introduces a methodology for optimizing the microstructure of solid-state batteries tailored to power tool applications, which demand significantly higher power density compared to many other fields such as automotive for example. A network-based impedance model, adapted from Braun [1] and parameterized using data from Kato [2], serves as the foundation for this approach.
Using Kato’s cell chemistry as a reference, a Latin Hypercube Sampling (LHS) is used to generate approximately 100.000 virtual cells with varied microstructural parameters within processing-constrained boundaries (e.g., minimal practical particle size). For each configuration, the impedance of the cell is transformed into the time domain to calculate energy and power densities.
A Multilayer Perceptron (MLP) Neural Network is trained on those cells’ parameters and the corresponding energy and power density to predict performance metrics. An optimization algorithm leverages this model to calculate a parameter set that maximizes energy density under a user-defined minimum power density. This ensures optimal runtime for specific power tool applications, such as chainsaws or hedge trimmers, while meeting their stringent power requirements. The minimum power requirement is determined from the necessary electrical power output. This approach enables users to identify optimal cell parameters in a single calculation, eliminating the need for multiple iterative simulations. This significantly reduces development time during the early design phase. Furthermore, the methodology not only enhances overall cell performance but also facilitates the identification of the most suitable parameter set for a given cell chemistry and any targeted application.
References

[1] Braun, Philipp, Elektrische Charakterisierung und Modellierung von Festkörperbatterien. Karlsruhe: KIT Scientific Publishing, 2019
[2] Kato, Yuki et al., High-power all-solid-state batteries using sulfide superionic conductors. s.l.: Nature Energy, 2016