Alessia Cicchino
Early-stage screening of pulsed charging protocols using early-life battery data: feature-based and raw-data learning approaches.
Rel. Davide Papurello. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Energetica E Nucleare, 2026
Abstract
The optimization of pulse fast-charging protocols for lithium-ion batteries (LIBs) requires lengthy and expensive aging tests, which are limited by a vast parameter space. Long-term prediction of cycle life from an early-life cycling window is crucial to accelerate research. However, manual feature extraction from these cycles can lead to the loss of nonlinear electrochemical dynamics, which are critical for understanding stochastic degradation mechanisms. Using experimental online measurement, aging, and Reference Performance Test (RPT) data from LG Chem MJ1 NMC/Gr-SiOx cells subjected to various charging recipes, a progressive predictive modeling study was conducted. Initially, traditional Machine Learning models based on tabular features and Incremental Capacity Analysis (ICA), such as Random Forest, XGBoost, and GPR, were developed.
The optimization of these models allowed for the definition of an innovative Bayesian and multi-criteria probabilistic screening framework
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Informazioni aggiuntive
Corso di laurea
Classe di laurea
Aziende collaboratrici
URI
![]() |
Modifica (riservato agli operatori) |
