On the creep brittle rupture of structures

• Autor(a) principal: Gonçalves Filho, Orlando João Agostinho
• Data de Publicação: 1984
• Outros Autores: Instituto de Engenharia Nuclear
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional do IEN
• Texto Completo: http://carpedien.ien.gov.br:8080/handle/ien/1897
• topic: creep brittle rupture; structures components

This work is concerned with the application of the finite element method to the study of creep brittle rupture of structural components. In the formulation material behavior is described by an elastocreep model in which the total strain rates are assumed to be the sum of elastic and creep components. The elastic strain rates are given by Hooke’s law while the creep strain rates and the damage rates are espressed by the multiaxial form of the Kachanov-Rabotnov equations proposed by Leckie and Hayhurst. The incremental equations of motion are derived from the principle of virtual work using an updated Lagrangian formulation which accounts for geometric effects due to large displacements, large rotations and deformation dependent loadings. The finite element incremental equations are developed according to a displacement-based formulation. Isoparametric elements with quadratic shape functions are employed for the domain discretization and simple numerical procedures are developed to deal with the presence of partially and/or fully ruptured elements in the mesh. For integration of the creep strain rate equations a family of implicit time marching schemes is developed which can be regarded as Runge-Kutta methods of second order. The integration of the coupled damage rate equations is performed using a first order predictor-corrector scheme with automatic time step length control. For material nonlinear problems only, a substructuring technique is employed in conjunction with the time integration algorithms. Selected numerical applications are presented and discussed in detail. Comparison with alternative numerical, analytical and/or experimental results is made whenever possible.