Abstract When assessing the tensile strength of a cohesive material for a wide range of test geometries, the change of peak loads as a function of size — the well-known size effect — is inconclusive to determine the initial behavior of the softening curve and, in particular, the true value of the tensile strength ft. Two different softening functions, such as linear and trapezoidal, can give the same accurate predictions for the maximum load, but different values of ft must be used. Brazilian splitting tests exhibit the same behavior and are less dependent on the initial shape of the softening curve, so this kind of test is recommended to assess the tensile strength. To know the actual softening curve, one needs additional experimental information, such as the load–displacement curve or the load–CMOD curve.
Read full abstract