Survival Analysis of Fatigue Lifetime in Aged Elastomers Using Weibull and Cox Regression Models
Abstract
Fatigue tests conducted on elastomers inherently exhibit significant variability. Aging may induce stiffening and embrittlement, leading to premature failures outside the specimen gauge length and resulting in right‐censored data. Combined with the limited number of specimens, this leads to large uncertainties in fatigue lifetime estimation. This paper presents a survival analysis strategy to propose a probabilistic prediction of elastomers' fatigue lifetime during aging. The approach explicitly accounts for censored data and incorporates aging effects through material descriptors such as crosslink density. A Weibull parametric model and a semi‐parametric Cox regression model with a frailty term are considered. Application to an experimental dataset shows that both models provide consistent predictions within the range supported by the data, while the Cox model offers increased flexibility by handling multiple aging‐related covariates without prescribing a baseline distribution. The proposed framework complements Weibull‐based approaches and provides a practical tool to quantify fatigue failure probabilities under data‐limited aging conditions.
Related
- Burgers dynamics for Poisson point process initial conditions of the Weibull class
- Effect of chirality on active elastomers and solids in three dimensions
- Electromechanical Domain Wall Propagation in Dielectric Elastomers: An Exact Geometric Resolution via Conformal Mapping
- Elastic theory of ideal liquid crystal elastomers
- 4D Printing of Magneto‐Active Soft Elastomers via Efficient Active Mixing Direct Ink Writing
BibTeX
@article{caillat2026survival,
title = {{Survival Analysis of Fatigue Lifetime in Aged Elastomers Using Weibull and Cox Regression Models}},
author = {Maéva Caillat and Hugo Madeira and Pierre-Yves Le Gac and Maelenn Le Gall and Erwan Verron},
journal = {Fatigue and Fracture of Engineering Materials and Structures},
year = {2026},
month = jul,
doi = {10.1111/ffe.70344},
url = {https://hal.science/hal-05678829v1},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Survival Analysis of Fatigue Lifetime in Aged Elastomers Using Weibull and Cox Regression Models
AU - Maéva Caillat
AU - Hugo Madeira
AU - Pierre-Yves Le Gac
AU - Maelenn Le Gall
AU - Erwan Verron
PY - 2026
DA - 2026/07/02
JO - Fatigue and Fracture of Engineering Materials and Structures
DO - 10.1111/ffe.70344
UR - https://hal.science/hal-05678829v1
AB - Fatigue tests conducted on elastomers inherently exhibit significant variability. Aging may induce stiffening and embrittlement, leading to premature failures outside the specimen gauge length and resulting in right‐censored data. Combined with the limited number of specimens, this leads to large uncertainties in fatigue lifetime estimation. This paper presents a survival analysis strategy to propose a probabilistic prediction of elastomers' fatigue lifetime during aging. The approach explicitly accounts for censored data and incorporates aging effects through material descriptors such as crosslink density. A Weibull parametric model and a semi‐parametric Cox regression model with a frailty term are considered. Application to an experimental dataset shows that both models provide consistent predictions within the range supported by the data, while the Cox model offers increased flexibility by handling multiple aging‐related covariates without prescribing a baseline distribution. The proposed framework complements Weibull‐based approaches and provides a practical tool to quantify fatigue failure probabilities under data‐limited aging conditions.
SN - 8756-758X
SN - 1460-2695
ER -