Initial Contact Stresses in Abaqus/Standard

Initial stresses for contact are derived from initial stresses that you specify for elements underlying contact surfaces. This involves estimating both normal and shear contact traction such that they are approximately in equilibrium with the element initial stresses. The procedure for estimating initial stresses matches the true distribution very well for cases where underlying element stress distributions result in either uniform or linearly varying contact stresses. For example, linearly varying underlying element stresses result in linearly varying contact stresses on flat contact surfaces. For more complex spatial variations of contact stress, the initial contact stress distribution may only be approximate. In situations where the underlying element stress distribution is an equilibrium stress distribution, nonzero initial contact stresses are meant to ensure that the overall system that includes both contact and user elements is close to an overall equilibrium state with respect to initial loads and boundary conditions. Consequently, an initial static or geostatic equilibration step with no changes to loads and boundary conditions is likely to converge more easily.

In cases where initial element stresses are different across the contact interface, the secondary side of the element stresses is used to estimate the initial contact stresses. These stresses can be visualized in the zero-increment frame of an analysis or after a data check analysis. You can control whether initial contact stresses are to be calculated and reported.

Initial contact stresses are calculated only when the initial contact status is closed. In other cases, contact stresses are reported as zero. It is strongly recommended that you use nodal adjustments (see Contact Initialization for General Contact in Abaqus/Standard and Contact Initialization for Contact Pairs in Abaqus/Standard) to set the initial contact status to closed. In the absence of nodal adjustments, even small variations in the contact gap along contact surfaces can lead to a very noisy distribution of initial contact status and, subsequently, the initial contact stress distribution. Such a noisy initial contact stress distribution may adversely affect convergence.

Initial contact stresses are calculated to be consistent with the mechanical contact properties that you specify for a particular contact region (see About Mechanical Contact Properties). For example, in the case of Coulomb friction the initial contact shear stresses are always capped at the maximum shear stresses that the interface can support based on initial contact normal stresses and the friction coefficient specified in the model definition. In situations where friction properties are changed during the first step (see Frictional Behavior), the friction specification at the model level prevails for initial stress calculations. Similarly, if the contact interface is specified to be frictionless, the initial contact shear stresses are reported as zero even if the underlying element stress distribution may result in nonzero shear tractions to satisfy equilibrium. For normal contact behavior, both tensile and compressive initial contact pressures are permitted if the interface is tied (see Defining Tied Contact in Abaqus/Standard) or no separation (see Contact Pressure-Overclosure Relationships) between contact surfaces is permitted once contact is established. However, for regular contact only compressive initial contact pressures are initialized. When cohesive behavior (see Contact Cohesive Behavior) or debonding (see Crack Propagation Analysis) is specified at the contact interface, initial contact stresses are not supported.

In models with pore pressure degrees of freedom, the initial contact pressures are affected by pore pressure values depending on the specification of pore fluid flow (see Pore Fluid Contact Properties) at the contact interface. If pore fluid flow is allowed across the interface, contact normal pressures are initialized to their effective stress values. However, if there is no pore fluid flow across the interface, contact normal pressures are initialized to their total stress values. The latter may also happen in situations where a contact surface on one side of the contact interface has pore pressure degrees of freedom while the opposing surface does not.