Limit-states and time-dependent mechanical properties of cement-based materials under complex stress-states using axial-torsion test setup

Abstract

Concrete structures are primarily designed to leverage their high compressive strength, with reinforcement placed to resist tensile and shear stresses. Unlike compressive and tensile strength, there is no standardized shear testing method for cement-based materials, leading to uncertainty in determining their exact shear strength. Existing correlation equations estimate shear strength from compressive strength, but their results vary significantly due to the uncontrolled, highly mixed stress-states generated by the test methods used in their development. This study evaluates current shear test configurations for their reliability in inducing shear-dominant stress-states and proposes an Axial-Torsion testing configuration to assess cement-based materials under shear, direct tension, compressive-shear, and tensile-shear stress-states. The Axial-Torsion test configuration and proposed specimen geometry enable direct tensile stress generation, which is rare in cement-based material testing. Conventional split-tensile and modulus of rupture tests provide only indirect tensile strength measurements. Direct tension testing reveals a substantial difference between indirect and actual tensile strength, particularly in UHPCs and Fiber-Reinforced Concretes (FRCs). Additionally, this configuration facilitates controlled mixed stress-states, such as compressive-shear and tensile-shear. Strength tests conducted in this study—including uniaxial compression, axial-torsion, and direct tension—demonstrate a dependency of material limit-state on applied stress-states. For example, as the compressive component increases in compressive-shear stress-states, failure stress rises significantly. Despite this increased capacity, standard concrete design practices remain highly conservative, particularly for UHPCs, FRCs, and prestressed concrete. Implementing the proposed Axial-Torsion testing method can enhance understanding of cement-based material limit-states and mechanical properties, which may in turn support improved accuracy in design practices. Creep in concrete depends on multiple factors, including temperature, relative humidity, and age, but it is also influenced by the applied stress-state. While concrete strength varies significantly with stress-state, current creep prediction models, including ACI 209, consider only compressive creep, occasionally leading to underestimated creep deflections in structures like long-span bridges. This study hypothesizes that creep is a function of the applied stress-state and should be incorporated into existing prediction models for more accurate results. Creep strain data was collected under uniaxial compression, shear, and direct tension, and analyzed using the framework of stress triaxiality. These short-term creep experiments reveal significant differences in shear and tensile creep behavior. The strong correlation between creep coefficients and stress triaxiality suggests the need for modifications to current creep design equations to improve their accuracy by integrating stress-state effects.