Loughborough University
Browse

Heterogeneous evolution of pore structure during loess collapse: insights from X-ray micro-computed tomography

Download (2.15 MB)
journal contribution
posted on 2021-03-31, 09:24 authored by B Yu, W Fan, Tom DijkstraTom Dijkstra, YN Wei, LS Deng
The pore structure is an important feature of loess fabric and significantly influences the hydraulic properties. The evolution of pore structure during loess collapse has been extensively investigated but some aspects are still not well understood. In this study, the 3D pore structure evolution within the same loess sample was characterized using X-ray micro-computed tomography in combination with an in-house collapse test set-up. The loess pore space consists of two distinct pore categories: constricted pores and free pores. The former can be assessed only through smaller pores while the latter are generally channel-like in shape. We extracted constricted pores and free pores from the tomography images and our results show that they behave differently during the collapse. The constricted pores were not destroyed during the collapse and many of them maintain well their sizes and shapes. On the other hand, the channel-like free pores were typically closed up and had a more important contribution to the overall porosity reduction. Our results demonstrate that the pore structure evolution associated with loess collapse is not only related to pore size, but also closely linked to pore morphology (i.e. constricted/free pores). The outcome of this work provides new insights into the loess collapse process and serves to better understand the underlying mechanism of fabric collapse.

Funding

National Natural Science of China (grand numbers: 41902281 and 41630634)

History

School

  • Architecture, Building and Civil Engineering

Published in

CATENA

Volume

201

Publisher

Elsevier

Version

  • AM (Accepted Manuscript)

Rights holder

© Elsevier

Publisher statement

This paper was accepted for publication in the journal CATENA and the definitive published version is available at https://doi.org/10.1016/j.catena.2021.105206.

Acceptance date

2021-01-23

Publication date

2021-02-08

Copyright date

2021

ISSN

0341-8162

Language

  • en

Depositor

Dr Tom Dijkstra. Deposit date: 28 March 2021

Article number

105206

Usage metrics

    Loughborough Publications

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC