Shi-Chen Tu
School of Architecture and Planning, Yunnan University, Kunming 650500

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Strength Prediction and Mechanism Analysis of Fiber-Reinforced Soil Based on RBF Neural Network Yin-Lei Sun; Shi-Chen Tu; Xue-Song Ding; Yang Hu; Xin-Sheng Zhang
Civil Engineering Journal Vol. 12 No. 9 (2026): September
Publisher : Salehan Institute of Higher Education

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.28991/CEJ-2026-012-09-01

Abstract

This study investigates how dry density, moisture content, fiber mass ratio, and fiber length govern the tensile behavior of fiber-reinforced red soil (FRRS). Brazilian splitting tests were conducted on red-soil specimens reinforced with basalt, polypropylene, or polyacrylonitrile fibers. An RBF neural network was applied to model the nonlinear tensile-strength response and quantify the relative importance of the four input factors. Scanning electron microscopy, surface-roughness measurements, strain monitoring, and crack-image analysis were further combined to examine the reinforcement mechanisms. Tensile strength generally increased with dry density but decreased with moisture content. For each fiber type, strength increased and then declined as dosage or length increased, because excessive addition or length impaired fiber dispersion. Moisture content made the largest model-based contribution, followed by dry density, fiber ratio, and fiber length. The strength improvement was mainly associated with fiber-particle frictional interlocking, interfacial stress transfer, crack bridging, and spatial confinement by the fiber network. By integrating four-factor analysis, RBF-based importance ranking, and multiscale observations, this study provides a unified interpretation of tensile reinforcement in uncemented red soil. The reported dosage-length combinations represent the best-performing tested levels rather than statistically verified global optima; therefore, moisture control should be prioritized when applying FRRS in engineering practice.