An innovative data-driven approach to the design and optimization of battery recycling processes

• A novel methodology for screening of operating conditions is developed. • Over ten thousand process simulations can be automatically generated. • The grade and recovery of all materials are identified in input and output streams. • The large data sets can be used to run robust multiobjective optimization routines. With the growing demand for raw materials to enable the ongoing electrification transition, robust battery recycling technologies will also become necessary to reduce reliance on primary resources and improve sustainability. To boost the recovery of secondary materials, we combined HSC-Sim® recycling process simulations with data science to analyze the flow of Li-ion battery components through the processing stages. Key operating parameters of the process were varied to assess their impact on material recovery and grade of graphite anode (Gr) and nickel-manganese-cobalt cathode (NMC). The resulting data distributions allowed us to establish if the process design was capable of producing desired recovery outcomes, and under which set of conditions optimal performance could be obtained. Materials flow analysis was utilized to guide decision-making and iteratively redesign the recycling process towards better outcomes. In the final stage, multi-objective optimization was deployed to achieve a balance between maximal NMC mass recovery of 66.3% at 95.7% grade and Gr mass recovery of 88.7% with 99.8% grade. This scalable, data-driven framework could replace intuition-led recycling process trials with rational process design to optimize complex device recycling, accelerating the transition towards more sustainable and effective material recycling.