The increasing use of lithium-ion cells in large-scale,long-term applications drives a need for design methods that considers aging and accurate state of health estimation.A common approach is to rely on an empirical ...The increasing use of lithium-ion cells in large-scale,long-term applications drives a need for design methods that considers aging and accurate state of health estimation.A common approach is to rely on an empirical or semiempirical aging model fit to experimental data to estimate the evolution of capacity and power fade.Because aging data are costly to collect,pack designers either use Design of Experiment(DOE)techniques to define a set of efficient tests,or use existing aging data to calibrate aging models.Given the increasing quantity of available aging data,the question arises:how can experimental aging campaigns be quickly compared?However,a methodology for the comparison of sets of aging experiments is not discussed in the literature.As a result,pack designers usually rely on intuition to select between multiple aging studies proposed by DOE techniques or in the literature.This work proposes metrics to quantitatively capture the alignment between a set of aging experiments and a target application.These metrics allow pack designers to quickly compare many sets of aging experiments to evaluate those which have tested conditions relevant to the application.Case studies are presented to illustrate the application of these metrics using aging campaign data from the literature.To validate these metrics,this work examines the relationship between these metric values and aging model validation error for calendar aging data for 18650 NMC battery cells.It is demonstrated that greater metric values correspond to reduced model error for an empirical capacity fade model.展开更多
基金supported by The Center for Automotive Research at The Ohio State University and the Department of Mechanical and Aerospace Engineering at The Ohio State University.
文摘The increasing use of lithium-ion cells in large-scale,long-term applications drives a need for design methods that considers aging and accurate state of health estimation.A common approach is to rely on an empirical or semiempirical aging model fit to experimental data to estimate the evolution of capacity and power fade.Because aging data are costly to collect,pack designers either use Design of Experiment(DOE)techniques to define a set of efficient tests,or use existing aging data to calibrate aging models.Given the increasing quantity of available aging data,the question arises:how can experimental aging campaigns be quickly compared?However,a methodology for the comparison of sets of aging experiments is not discussed in the literature.As a result,pack designers usually rely on intuition to select between multiple aging studies proposed by DOE techniques or in the literature.This work proposes metrics to quantitatively capture the alignment between a set of aging experiments and a target application.These metrics allow pack designers to quickly compare many sets of aging experiments to evaluate those which have tested conditions relevant to the application.Case studies are presented to illustrate the application of these metrics using aging campaign data from the literature.To validate these metrics,this work examines the relationship between these metric values and aging model validation error for calendar aging data for 18650 NMC battery cells.It is demonstrated that greater metric values correspond to reduced model error for an empirical capacity fade model.
