Effect of Crack-Parallel Compression or Tension on Mode-I Fracture Energy of Quasibrittle Material – As Applied to Concrete

In all widely used fracture test specimens, the compressive or tensile stress parallel to the plane of growing crack is negligible, and thus its effect cannot be revealed. The classical fracture models, including the cohesive crack model, cannot capture any effect of such crack-parallel normal stress and strain, except parametrically, because they do not figure such stress and strain as the basic thermodynamic variable. To capture this, the fracture process zone whose 3D stress and strain state is fully described must be implemented. Here it is shown experimentally, and documented by crack band finite element simulations with microplane model M7, that the crack-parallel normal stresses have a major effect on quasibrittle materials such as concrete. They are shown to cause a major decrease or increase of the Mode I (opening) fracture energy Gf (or fracture toughness KIc). The experiments introduce a modification of the standard three-point bend test, the idea of which is to use plastic pads with a near-perfect yield plateau to first generate compression and a gap at end supports to close later and generate bending. The experiments show and the microplane model confirms that a moderate crack-parallel compression greatly increases Gf (even doubling it), but a higher compression reduces Gf greatly, which represents the case of compression splitting. Through numerical extrapolation, it shows that crack-parallel tension reduces Gf and further that a high compressive or tensile stress normal to the specimen plate has a similar major effect on Gf. While mild parallel stresses arise in shear failure of reinforced concrete beams or slabs and prestressed concrete, high crack-parallel stresses will be impactful in hydraulic fracturing of shale when the effective stress state in the solid phase changes at the presence of a nearby borehole or fluid diffusion.