Strength prediction models for fiber-reinforced soils under effective stress triaxial compression testing

The improvement of shear strength of soils is a common practice applied in numerous civil engineering projects. Geosynthetics like geogrids, geotextiles and fibers contribute significantly in soil strength improvement. Soil reinforcement with randomly oriented fibers has attracted extensive research attention. The shear strength of fiber-reinforced soils is often determined experimentally by performing triaxial compression tests under effective stress conditions. Presented in this paper are easy-to-use models for the strength prediction of fiber-reinforced soils under the aforementioned testing conditions. After an extensive literature review, a database of 436 measurements was created including data for all independent variables relative to the soil and fiber characteristics. This database was then processed and divided into two subsets: the first, consisting of 75% of the measurements, was used to develop the models by performing statistical analyses, whereas the second, with the remaining 25% of the measurements, served as a basis for their validation. The development of the models was achieved through multivariable linear regression analyses performed with suitable statistical software. After an extensive number of trials, three optimal models were derived that satisfy the statistical requirements, exhibiting R 2 determination coefficients greater than 0.95 and satisfactory estimation of the validation measurements at a rate reaching 81%.