New insights into the frost heave behavior of coarse grained soils for high-speed railway roadbed: Clustering effect of fines

Abstract

Due to the chemical component and physical property, fines of coarse-grained soil (CGS) are prone to adsorb with each other forming cluster morphology during mixing and compacting process for high-speed railway roadbed construction. The clustering phenomenon of fines in CGS was quantitatively characterized by X-ray computed tomography (X-CT) and its effect on the frost heave behavior was investigated for the first time in this study. For this purpose, greyscale level based digital microstructures were generated by X-CT with a resolution of 20 um/voxel length. The frost heave ratios of samples with fine contents of 0%, 3%, 5%, 8%, 10% and 15% were tested under fully saturated condition in a close freezing system. To overcome the overlapping gray values between fines and aggregates, an image processing method based on a derivative function was introduced to segment the morphological features of the clustering fines from sectional X-CT images. Furthermore, a gray-based method was proposed to describe the fines distribution and to calculate the degree of clustering of fines. It is founded that the distribution of fines is an important factor affecting the frost heave behavior of CGS. When fines content is low, a ‘skeleton gap structure’ is formed, the unexpected frost heave comes from the regional frost expansion in fines cluster, which squeezed adjacent aggregate skeleton, overcoming the interlock force of CGS; When fines content is high and a ‘suspend dense structure’ is formed, fines clusters connect with each other and dominate the frost heave behavior of CGS, resulting in continuous frost heave. Compared with the fines content, clustering ratio shows more clear linear relationship with frost heave ratio because the regional ununiform distribution of fines was considered. The proposed method is regarded more promising as it captures the detailed microstructure of coarse-grained soil, which has the potential to unveil the inner relationship between micro characteristics and macro performance.