Impact of Particle Shape on Microstructure and Performance of Cementitious Composites

Author(s): Huisu Chen, Zhigang Zhu, Lin Liu and Wenxiang Xu
Paper category: Proceedings
Book title: 3rd International RILEM Conference on Microstructure Related Durability of Cementitious Composites
Editor(s): Changwen Miao, Wei Sun, Jiaping Liu, Huisu Chen, Guang Ye and Klaas van Breugel
Print-ISBN: 978-2-35158-188-9
e-ISBN: 978-2-35158-189-6
Publisher: RILEM Publications SARL
Publication year: 2016
Pages: 105-114
Total Pages : 10
Language : English

Abstract: Due to the complexity of constructing non-spherical particle and their packing algorithm, the influence of particle shape on microstructure and macro-properties of cementitious composites is seldom discussed in the literature. In this paper, ITZ layer was successfully constructed around Platonic particles by a numerical technique. Both systematic line sampling(SLS) and normal line sampling(NLS) techniques are employed to statistically investigate the influence of particle shape on the ratio of apparent to actual ITZ thickness (t'/t) around a single particle. Then, a generalized analytical formula of ITZ volume fraction (VITZ) is used to assess the exaggeration of VITZ induced by t'/t. And differential effective medium approximated is employed to present the impact induced by the overestimation of ITZ thickness on the exaggeration of concrete diffusivity. The numerical modelling result demonstrated that the particle shape exerts a pronounced effect on t'/t. The value of t'/t from NLS is lower than that from the SLS, provided all other parameters are kept the same. The exaggerated degree of VITZ can be explicitly expressed as the function of fineness and volume fraction of aggregate, actual ITZ thickness, and particle sphericity. And the numerical simulation result revealed that depending on exact values of aggregate shape, size distribution and volume fraction, the diffusivity ratio of ITZ to cement paste matrix, actual ITZ thickness, the diffusivity of concrete can be overestimated by 19%~34% for SLS algorithm, and 10%~13% for the NLS algorithm.

Online publication : 2016
Publication type : full_text
Public price (Euros) : 0.00