抄録
In recent years, it has been common to perform a finite-element analysis focusing on the deformation with plastic strain after uniform elongation. For example, a simulation might be performed of the rupture in a smooth material that receives tensile deformation, or the fracture mechanical parameter around a crack tip might be evaluated. Although these analyses require a true stress-strain curve after uniform elongation as an input, a true stress-strain curve after uniform elongation cannot be obtained using a tensile specimen test conducted under the general procedures. This is because necking, which is strain localization, occurs just after uniform elongation, and the true stress-strain curve cannot be calculated from the nominal stress-strain curve. This study investigated the true stress-strain curve after the uniform elongation of line pipe steels. True stress-strain curves for various materials were obtained by measuring the deformation shape of the necking part during tensile tests. To select an appropriate approximate function for describing the true stress-strain curve after uniform elongation, the errors in the Hollomon (exponential law) , Ludwic, Swift, and Ramberg-Osgood functions were compared. The errors in the Ludwic and Swift functions were smaller than those for the Hollomon and Ramberg-Osgood functions when the curves described by these functions were best-fitted to the measured data. Additionally, the errors that occurred in the extrapolation of the true stress-strain curve were shown. In this extrapolation, the parameters for each function were determined using the best-fit to the relationship before the uniform elongation rather than afterward. The errors in the true stress by the extrapolation had values of 10% or more. Finally, a regression formula was proposed for the true stress-strain curve after uniform elongation. This formula is based on a Swift function, and its parameters can be calculated using only the fundamental material strength parameters, including the yield strength, tensile strength, and uniform elongation. The error in the true stress by this regression formula was about 4%.
