Robust stress integration algorithms for implicit elastoviscoplastic finite element analysis of materials with yield-point phenomenon. (January 2019)

• Record Type: Journal Article
• Authors: Kim, Jaehyun; Kim, Do-Nyun
• Is Part Of: International journal of mechanical sciences. Volume 150(2019)
• Page Start: 277
• Page End: 289
• Publication Date: 2019-01
• Subjects: Elastoviscoplasticity -- Yield-point phenomenon -- Stress integration algorithm -- Two-point Newton method -- Constitutive behavior; Mechanical engineering -- Periodicals; Génie mécanique -- Périodiques; Mechanical engineering; Maschinenbau; Mechanik; Zeitschrift; Periodicals; 621.05
• DOI: 10.1016/j.ijmecsci.2018.10.018
• Languages: English
• ISSNs: 0020-7403

Abstract

Highlights: Elastoviscoplastic analysis of materials with yield-point phenomenon is computationally challenging due to its non-convergent characteristics. The conventional one-point Newton method mostly fails to provide a converged solution near the upper yield point in the incremental solution procedure. We propose stress integration algorithms based on the two-point Newton method and demonstrate their robustness in the analysis of materials with yield-point phenomenon. Graphical abstract: Abstract: In order to describe the yield-point phenomenon accurately in computational elastoviscoplastic analysis, several constitutive models have been proposed and studied in the field of metal plasticity. To precisely predict the behavior of materials with this phenomenon, the constitutive model should be able to depict both the yield-point phenomenon and the Bauschinger effect, rendering it difficult to obtain a converged solution in implicit numerical analysis when the conventional one-point Newton method is used. Here, we propose robust stress integration algorithms that can be used effectively in implicit finite element analysis. The two-point Newton method is employed and compared with other stress integration algorithms using the conventional one-point Newton method or the bisection method. Results demonstrate that they can be reliably used to calculate the solutions of YPP problems that cannot be obtained using conventional iterative methods although these algorithms may require longer computational times.