Shear Strength Enhancement Prediction of Sand–Fiber Mixtures Using Genetic Expression Programming

Abstract

A series of direct shear tests were conducted on fine, medium, and coarse sand–fiber mixtures subjected to normal pressures of 100, 200, and 300 kPa. Soils were mixed with 0.1%, 0.2%, and 0.3% fibers 6, 12, and 18 mm in length to assess their effects on shear strength characteristics of samples with 70% relative density. Genetic expression programming (GEP) was also evolved to envisage strengths and strains of sand–fiber mixtures. Test results showed that in general, fiber inclusion improves composite characteristics with fine sand–fiber mixtures showing the greatest enhancements in shear strengths. Inclusion of 6 and 12 mm fibers in coarse sand failed to improve strength characteristics due to the inadequate anchorage length and ineffective interactions with grains. Amid the factors investigated, fiber length and concentration, normal stress, and gradation proved to be the most influential. GEP results revealed high accuracy with coefficients of determination of 0.945 and 0.951 for training and testing phases, respectively, and sensitivity analysis illustrated that interaction is most sensitive to fiber length and the coefficient of curvature.