B0712
Title: Simple heterogeneous step-stress accelerated life testing model for lithium-ion battery aging
Authors: Yao Lu - RWTH Aachen University (Germany) [presenting]
Maria Kateri - RWTH Aachen University (Germany)
Abstract: Accelerated life testing (ALT) is widely implemented to investigate lifetime performance within a comparatively short time period in the lithium-ion battery (LIB) field. With higher-than-usual levels of stress factors, e.g., temperature, charge and discharge voltages, the aging of the tested LIB cells is sped up, inducing failures much earlier than under the user level. Step-stress ALT (SSALT) is a special case of ALT, where the stress level imposed on a unit changes gradually at pre-specified time points during the experiment. Statistical models for SSALT experiments assuming a homogeneous population and a variety of lifetime distributions, under different censoring schemes, have been extensively discussed in the literature. However, SSALT for a heterogeneous population received little attention, especially in the case of unknown group membership. Motivated by heterogeneous LIB cell aging patterns observed in practice, a simple heterogeneous SSALT model for exponential distributed lifetimes is introduced. Heterogeneity is captured through a mixture model approach. An EM algorithm is developed to derive the maximum likelihood estimates of the model's parameters under Type-II censoring and corresponding bootstrap confidence intervals are provided. In case of heterogeneity, the validity of the proposed model and its advantage over the classical SSALT model are demonstrated via simulation studies.
