Microstructure characteristics of nonwoven geotextiles using SEM and CT methods. Issue 2 (April 2023)

• Record Type: Journal Article
• Authors: Li, Ke-Yi; Tang, Xiao-Wu; Zhao, Wen-Fang; Fei, Min-Liang; Chen, Xiu-Liang; Liang, Jia-Xin
• Is Part Of: Geotextiles and geomembranes. Volume 51:Issue 2(2023)
• Page Start: 293
• Page End: 302
• Publication Date: 2023-04
• Subjects: Nonwoven geotextiles -- SEM imaging -- CT imaging -- Image analysis -- Microstructure characteristics; Membranes (Technology) -- Periodicals; Membranes (Technology); Periodicals; Electronic journals; 624.189
• DOI: 10.1016/j.geotexmem.2022.11.003
• Languages: English
• ISSNs: 0266-1144

Abstract

Abstract: Two digital image methods based on scanning electron microscope (SEM) and computed tomography (CT) were proposed to study the microstructural characteristics of staple fibers and continuous filament geotextiles. Two-dimensional (2D) image analysis was developed for SEM images using a machine-learning-based segmentation algorithm. Three-dimensional (3D) image analysis of the CT images was based on 3D reconstruction and a pore network model. The fiber orientation distribution, porosity, pore size distribution (PSD), and characteristic pore size O 95 determined from image analysis were compared with the theoretical equation and bubble point test (BBP) results. It is shown that 2D and 3D image analyses can accurately measure the fiber orientation distribution of the geotextiles. The porosity values obtained using 3D imaging were comparable to theoretical values. The PSD curves obtained in the BBP tests were in good agreement with those obtained using the 3D image method. O 95 sizes of continuous filament geotextiles estimated by 2D image analysis compared well with O 95 sizes obtained by BBP tests, whereas this was not the case for staple fiber geotextiles. The O 95 pore throat sizes of the two nonwoven geotextiles determined by 3D image analysis were comparable to the BBP test-based values and 2D image analysis-based values. Highlights: The applicability and results of 2D and 3D image methods for measuring the microstructure characteristics of nonwoven geotextiles were discussed. A machine learning-based 2D image method without complex sample preparation and parameter tuning was proposed. 2D image analysis have accurate results of O 95 sizes for continuous filament nonwoven geotextiles. 3D image analysis is a good approach for determining the pore characteristic of nonwoven geotextiles.