Diffusion-limited C-rate as a physics-based constraint to accelerate automated design for lithium-ion batteries

Abstract

Automated design for lithium-ion batteries involves optimizing electrode parameters to meet the performance goals using optimization algorithms, typically requiring thousands of iterations in one design. One possible method to accelerate the optimization is to identify the collapse point on capacity curves versus the discharge C-rate, which is related to the diffusion-limited C-rate (DLC). This work proposes applying DLC as a physics-based constraint in optimization algorithms. DLC constraints limit the initial solution space, improving the efficiency by 19.7% with the 1.5 C constraint. Additionally, we find that under certain DLC constraints, energy density and power density deviate from the generally accepted trade-off relationship. The angle difference of their negative gradient vectors is defined to calibrate the trade-off level. Further, high enough DLC constraints allow simplification of multi-object optimization with DLC constraint in an inequality form to single-object optimization with DLC constraint in an equality form, reducing the calculation time by 97% without significantly affecting the outcome. Finally, we examine the influence of electrolyte volume and particle size on optimizing results. To sum up, DLC constraints reduce the search domain and transform multi-objective problems into single-objective ones under certain conditions, thereby improving the efficiency of the automated design for lithium-ion batteries.