Title: Experimental study on the characteristics and strength decay mechanism of moraine soil using deep learning neural network

Authors: B.V. Santhosh Krishna; Kiran Shrimant Kakade; P. Sathiyavathy; Zhiyang Han

Addresses: Department of Computer Science and Engineering, Bangalore Technological Institute, Bengaluru, 560035, Karnataka, India ' Faculty of Management, Symbiosis Institute of Management Studies (SIMS), Symbiosis International (Deemed University), Pune, India ' Department of Computer Science and Engineering, S.A. Engineering College, Poonamallee-Avadi Road, Thiruverkadu, Chennai, Tamil Nadu, 600 077, India ' Agricultural and Animal Husbandry University, Haidian, Beijing, China

Abstract: Southeast Tibet, frequent rainfall, seasonal permafrost distribution is very extensive, freeze-thaw cycle on seasonal permafrost area moraine accumulations and slope stability has a significant impact. In order to investigate the freezing and thawing cycle action and water content on the mechanical properties of moraine soil, to southeast Tibet Linzhi area yigong landslide disaster point moraine soil as the object of study, to carry out different freezing and thawing cycle times and water content under the condition of large-scale straight shear test. The test results show that: moraine soil shear curve is roughly divided into elastic deformation, strain hardening, plastic yield three stages; after experiencing freezing and thawing cycle moraine soil cohesion is a negative exponential function (y = a + b exp(cx)) decrease, the internal friction angle in the range of 38.65~42.9° fluctuations; freezing and thawing cycle destroys the connecting structure of the moraine soil particles.

Keywords: moraine soils; freeze-thaw cycles; water content; shear strength; particle grading; SEM images.

DOI: 10.1504/IJEWM.2026.154487

International Journal of Environment and Waste Management, 2026 Vol.40 No.2, pp.200 - 217

Received: 13 Sep 2024
Accepted: 28 Feb 2025

Published online: 30 Jun 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article