Tensile testing is widely used for the mechanical characterization of materials subjected to superplastic deformation. At the same time, it is known that the obtained flow data are affected by specimen geometry. Thus, they characterize the specimen rather than the material. This work provides the numerical analysis aimed to study how the material flow behavior affects the results of tensile tests. The simulations were performed by the finite element method in Abaqus software, utilizing user-defined procedures for calculation of forces acting on the crossheads. The accuracy of tensile testing is evaluated by the difference between the input material flow behavior specified in the simulations and the output one, obtained by the standard ASTM E2448 procedure based on the predicted forces. The results revealed that the accuracy of the superplastic tensile test is affected by the material properties. Even if the material flow behavior follows the Backofen power law, which is invariant for the effective strain, the output stress–strain curves demonstrate significant strain hardening and softening effects. The relation between the basic superplastic characteristics and the tensile test errors is described and analyzed.

Superplasticity is an ability of polycrystalline materials to archive extremely large deformations, which is utilized in advanced forming technologies demanded mainly in the aerospace industry. Design of such technologies needs an information of the material flow behaviour, which describes the relation of the effective stress on the strain and strain rate taking place during deformation. The most popular experimental method for investigation of the flow behaviour of superplastic materials is tensile testing. The procedure of superplastic tensile testing and interpretation of its results is described by several international standards. At the same time, it is known that the due to flow inhomogeneity in the specimen volume the accuracy of such tests may be violated. Moreover, different standards provide different ratio between the width and the height of the gauge area of a specimen. This work provides the numerical analysis aimed to study how the initial specimen geometry affects the results of tensile tests. A computer program implementing finite element method (FEM) was developed to predict the specimen deformation during the test. A flat specimen is discretized using prismatic elements with specific geometrical constraints reducing the degree of freedom to the order of a plane stressed task. The output stress and strain values were calculated as specified in the ASTM E2448 standard. The effect of the gauge length was studied focusing on the output stress strain curves. The results were compared with the experimental results available in the literature.