Further developments of a dynamic real-time model of a tubular centrifuge fed with multi-component dispersions for application in fractionation for Direct Recycling of lithium-ion batteries.

As the Direct Recycling of Lithium-ion batteries may become imperative, centrifugal separation of electrode materials is studied as a potential key process step. The focus here lies on Lithium Iron Phosphate (LFP) and Carbon Black fractionation in a tubular centrifuge and its digital real-time representation. A real-time model is therefore elaborated, which is an enhancement of the existing compartment approach and relies on mass balances for particle sizes and species. In addition, it requires material parameters for accurate representation of hindered settling and sediment compression. Global sensitivity analysis with Latin-Hypercube One-At-A-Time Sampling confirms the parameters' significance and the respective development over process time. A correspondingly tailored optimization procedure incorporating a Genetic Algorithm is designed in order to derive the searched material parameters from actual process data. Results indicate that Genetic Algorithms are promising, robust tools in this application. Also, the model can be validated with appropriate material parameters. • Real-time compartment model of a tubular centrifuge for multiple solid species. • Settling and sediment analysis for LFP and Carbon Black (carbon, agglomeration). • Time-resolved sensitivity analysis. • Genetic Algorithm finds optimal material parameters for fractionation modeling. • Model with optimized material functions is validated against experimental data. [ABSTRACT FROM AUTHOR]