Probabilistic model for flexural strength of carbon nanotube reinforced cement-based materials

• Published in: Composite structures 2020-12, Vol.253, p.112748, Article 112748
• Main Authors: Ramezani, Mahyar, Kim, Young Hoon, Sun, Zhihui
• Format: Article
• Language: English
• Subjects: Bayesian methodology; Carbon nanotubes; Cementitious composites; Dispersion; Flexural strength; Interfacial bond strength; Kelly-Tyson model; Probabilistic model
• ISSN: 0263-8223; 1879-1085
• DOI: 10.1016/j.compstruct.2020.112748

The bond between carbon nanotubes (CNTs) and cementitious materials is the key characteristic for predicting the flexural strength. However, CNT dispersion quality may change the bonding mechanism. A probabilistic approach can benefit the deterministic models to capture the uncertainties affecting these characteristics. This study proposes a probabilistic model using the Kelly-Tyson theory to predict the flexural strength of CNT-cement nanocomposites. The proposed model considers the effects of experimental variables on CNT dispersion quality, bonding mechanism and the flexural strength. To this end, a Bayesian methodology is employed to calibrate the unknown model parameters and various sources of uncertainty using extensive test data. The model is then used to identify the optimum ranges of variables to maximize the flexural strength through computing the failure probability which is defined as the probability of not meeting certain strength requirements (herein, 50% increase compared with the control). The model suggests that CNT aspect ratio ranges from 400 to 800 and concentration between 0.08 and 0.18 c-wt% yields the highest flexural strength. Finally, the effect of changes in experimental variables on the probability estimates is examined using sensitivity and importance measures. The analysis reveals that the proposed model can capture the experimentally observed trends with reasonable accuracy. For example, the importance of age increases as CNT concentration increases.