Abstract

In many standard tests involving composite material, specially those needing tabs at the ends of the sample, configurations of multimaterial corners, with abrupt changes in geometry and material properties, typically appear giving rise to stress fields which achieve very high stresses and theoretically are unbounded. These stress intensifications may promote premature failures, or simply, alterations of the uniform stress state which is expected to occur in the specimen, and the test strength, calculated as a certain load divided by a certain failure area, may not be representative of the real strength of the material under analysis. In the present work three examples are presented, in which, by means of an addecuate selection of the geometrical parameters defining the test configuration which do not affect the mechanical property to be measured, the stress singularities at these multimaterial corners are removed (or at least reduced to a minimum). The three cases are: a) tests in tension and shear (Iosipescu) in bimaterial coupons, b) the off-axis test for the intralaminar shear strength in composite materials and c) a compression test of thick laminates in composite materials. In two of the three cases experimental results are available to corroborate that the removal of the stress singularities gives rise to higher strength values than those obtained in the original test configurations.