Revealing stress evolution mechanisms in all-solid-state batteries: a non-invasive parameter identification framework for battery design

All-solid-state batteries have garnered attention due to their high safety and energy density. However, the volumetric swell of Si-C anodes during lithiation causes complex interactions among electrochemical reactions, solid-state diffusion, interfacial contact, and mechanical stress, making liquid-cell parameter-identification methods unsuitable. To overcome this issue, an electrochemical–mechanical model was developed within the P2D framework, incorporating chemical strain, a preloading constraint, and an interfacial contact-resistance evolution mechanism to clarify how stress influences chemical potential, interfacial resistance, and terminal voltage. Using Latin hypercube sampling, sensitivity analyses were performed on 26 parameters under multiple conditions to create a hierarchical identification process. A distributed multi-agent, multi-objective optimization algorithm was introduced, assigning each parameter its own fitness function and facilitating information exchange and staged identification through an adjacency matrix and dynamic weighting. Numerical validation demonstrated significant improvements in accuracy and convergence compared to MOPSO and MOGWO. Testing on a sulfide-based ASSB platform confirmed voltage errors within 20 mV and swell force errors within 0.1 MPa. It also reliably reproduced the interfacial resistance and the bathtub-shaped evolution of nonmonotonic high-rate voltage. The approach’s effectiveness in revealing coupling mechanisms and enabling noninvasive parameter identification was confirmed, providing a valuable tool for ASSB state estimation and structural design.