TY - JOUR
AU - Rahmanian, Fuzhan
AU - Vogler, Monika
AU - Wölke, Christian
AU - Yan, Peng
AU - Winter, Martin
AU - Cekic-Laskovic, Isidora
AU - Stein, Helge S.
TI - One‐Shot Active Learning for Globally Optimal Battery Electrolyte Conductivity**
JO - Batteries & supercaps
VL - 5
IS - 10
SN - 2566-6223
CY - Weinheim
PB - Wiley-VCH
M1 - FZJ-2023-05462
SP - e202200228
PY - 2022
AB - Non-aqueous aprotic battery electrolytes need to perform wellover a wide range of temperatures in practical applications.Herein we present a one-shot active learning study to find allconductivity optima, confidence bounds, and relating formulationtrends in the temperature range from 30°C to 60°C. Thisoptimization is enabled by a high-throughput formulation andcharacterization setup guided by one-shot active learningutilizing robust and heavily regularized polynomial regression.Whilst there is an initially good agreement for intermediate andlow temperatures, there is a need for the active learning step toimprove the model for high temperatures. Optimized electrolyteformulations likely correspond to the highest physicallypossible conductivities within this formulation system whencompared to literature data. A thorough error propagationanalysis yields a fidelity assessment of conductivity measurementsand electrolyte formulation.
LB - PUB:(DE-HGF)16
UR - <Go to ISI:>//WOS:000844221600001
DO - DOI:10.1002/batt.202200228
UR - https://juser.fz-juelich.de/record/1019515
ER -