Effects of autogenous shrinkage microcracks on UHPC: Insights from a machine learning based crack quantification approach

Xiaolan Zeng, Qian Deng, Shaohua Li (Corresponding author), Hongbo Gao, Qingliang Yu (Corresponding author)

Research output: Contribution to journalArticleAcademicpeer-review

4 Citations (Scopus)
62 Downloads (Pure)

Abstract

Owing to its high pozzolanic reactivity and contribution to packing density, silica fume is almost an indispensable part of UHPC, but it brings potentially serious problem of high autogenous shrinkage. Nanocellulose (NC) is very effective in controlling shrinkage but would also result in different microstructure, whose impact is not clear. This study aims to explore the compensatory effect of NC on high shrinkage of UHPC. An image process method based on machine learning and stereological methods is proposed to quantify the autogenous shrinkage induced microcracks. Results show that the addition of NC reduces the crack width and area by 57.45∼70.55% and 63.2–83.8%, respectively. The MIP analysis reveals that the incorporation of NC introduces a larger proportion of pores. In terms of mechanical properties, the higher content of pores brought by NC has a negative effect on compressive strength, however, the enhancement of flexural strength by NC can reach 66.02%. Excellent correlations between 0 and 50 nm porosity and compressive strength, crack density and flexural strength are observed with R2 of 0.94 and 0.98 respectively. This study provides a theoretical basis for the potential control of porosity and cracks of UHPC to meet the different mechanical performance requirements of components.

Original languageEnglish
Article number136400
Number of pages17
JournalConstruction and Building Materials
Volume428
DOIs
Publication statusPublished - 17 May 2024

Bibliographical note

Publisher Copyright:
© 2024 The Authors

Keywords

  • Autogenous shrinkage
  • Cracks quantification
  • Image process method
  • Nanocellulose
  • Silica fume
  • Ultra-high performance concrete

Fingerprint

Dive into the research topics of 'Effects of autogenous shrinkage microcracks on UHPC: Insights from a machine learning based crack quantification approach'. Together they form a unique fingerprint.

Cite this