A novel failure-mechanism-based reinterpretation of the Brazilian test using discontinuity Layout optimization

A simplified numerical formulation of the Brazilian test for intact rock is presented, reinterpreting the test as a failure-mechanism problem analyzed using limit analysis and Discontinuity Layout Optimization (DLO). Although these tests are designed to assess tensile strength, load transmission occurs through highly concentrated compressive contact zones at the platen-specimen interfaces. By relying exclusively on material strength parameters (without explicit fracture modelling or fitting assumptions), the proposed model inherently accounts for the localized crushing and inelastic deformation that precedes tensile splitting, directly predicting the limit load and the transition from contact-controlled behavior to global tensile failure. The Hoek–Brown failure criterion with a Mogi-Walton high-confinement limitation is adopted, so that tensile strength is not introduced as an independent parameter but follows directly from the geological parameter m0 and the uniaxial compressive strength UCS. To validate the model, comprehensive laboratory tests (encompassing uniaxial compression, direct tension, and Brazilian tests, plus stress-sensitive film imprint analyses) were conducted using five distinct lithological types. The validation results demonstrated excellent agreement in most instances, even in ranges where the elastic formula tends to perform inadequately. A key outcome of the approach is that the effective contact width beneath the loading platens is not prescribed a priori; instead, it emerges naturally from the collapse mechanism, showing systematic variations with rock strength consistent with trends reported in the literature. On this basis, a new relationship is proposed to reinterpret Brazilian test results and provide a more reliable tensile strength estimate than the conventional formula.