Title:
Frost Durability and Service Life Prediction of Self-Healing Concrete
Author(s):
Jialuo He, Ayumi Manawadu, Yong Deng, Jie Zhao, and Xianming Shi
Publication:
Materials Journal
Volume:
121
Issue:
5
Appears on pages(s):
23-38
Keywords:
freezing and thawing (F/T); microcapsule; probabilistic damage model; self-healing concrete; service life prediction; three-parameter Weibull distribution
DOI:
10.14359/51742036
Date:
9/1/2024
Abstract:
This laboratory study employed synthesized urea-formaldehyde
(UF) microcapsules and polyvinyl alcohol (PVA) microfibers as a
self-healing system to improve the durability of concrete in cold
climates. The resistance of concrete specimens to rapid freezingand-
thawing (F/T) cycles was evaluated by measuring the change
of relative dynamic modulus of elasticity (RDM) with respect to
the number of F/T cycles. The control specimens (either with or
without PVA microfibers) approached the failure state with a reduction of 38% in RDM after being subjected to 54 F/T cycles, whereas the self-healing specimens (either with or without PVA microfibers) remained in a good state with a reduction of approximately 10 to 15% in RDM after 732 F/T cycles. A polynomial regression model was developed to establish the relationship between the RDM and number of F/T cycles, and a three-parameter Weibull distribution model was employed to conduct the probabilistic damage analysis and characterize the relationship between the number of F/T cycles (N) and the damage level (D) with various reliabilities. The results revealed that the benefits of UF microcapsules and PVA microfibers to the frost durability of concrete diminish once the damage level exceeds a certain high level. Based on the Weibull distribution model, the relationships were established and validated between N and D by comparing the experimental data, the predicted data based on the nonlinear polynomial regression model, and the predicted data based on N-D relationships. The field service life of the self-healing concrete was then predictable at any given reliability.