Modeling Inelasticity With SimScale: Plasticity and Creep
The simplest form of analysis in solid mechanics is linear static analysis. Here the deformations in the material are assumed to be small and Hooke’s law (or a linear relation between stress and strain) can be assumed. However, as the load increases, the material behavior (also known as the stress-strain relationship) becomes nonlinear, demonstrating effects like plasticity and creep.
There are several types of nonlinearities, such as material, geometric, and boundary nonlinearities. Thin structures undergo large rotations and displacements in spite of the material satisfying Hooke’s law; these are known as geometric nonlinearities. Alternatively, when certain aspects of contact phenomenon are involved, nonlinearity exists in the boundary conditions (How to Set Up Boundary Conditions in your Simulation?). However, the most common type of nonlinearity is the material nonlinearity observed as a result of the nonlinearity in the stress-strain behavior.
The most challenging problems are, of course, those involving a combination of the above-mentioned nonlinearities and possibly others. For example, in a car crash simulation, the displacements are large and the stress-strain behavior is nonlinear. Additionally, inertia and contact (for example, with other bodies and self-contact) must be considered during the simulation.
Material Nonlinearity: Plasticity
Material nonlinearity is related to aspects of nonlinear elasticity and inelastic phenomenon like plasticity, creep, and others. At the point of small stresses, both materials follow a linear stress-strain behavior. Upon an increase in stress, the stress-strain relation is nonlinear, but there is no distinction made between loading and unloading, except for the sign in nonlinear elastic materials. Upon complete unloading, no residual strains are observed.
In contrast, in elastoplastic materials, as the stress increases beyond a threshold (known as yield limit), the stress-strain behavior becomes increasingly nonlinear. Upon unloading, the elastoplastic material leads to a new branch of the stress-strain curve where the material behaves like elastic again. However, upon complete unloading, a residual strain (known as plastic strain) remains. This is described as plasticity behavior observed in the material.
In reality, the curved part observed in elastoplastic materials is quite complicated. For this reason, a number of approximations—known as hardening models—are made to describe this region. There are several models that can be applied to large-scale simulations to replicate the plastic effect. The most common hardening rules are isotropic, kinematic, and mixed. Several complicated hardening rules and plasticity models have evolved over the recent decades. Nevertheless, it is important to understand the physical cause that leads to plasticity, and this varies according to the material. For example, when existing in metals, it is the movement of dislocations, while in polymers, it is due to chain rearrangements.
In the presence of hardening, both the elastic and plastic strain continuously increase beyond the yield limit. In other words, the strains increase corresponding with the increasing stresses but at a lower rate than below the yield limit. Surprisingly, in the presence of softening effects, the strains still continue to increase in spite of a decrease in stress. The difference between hardening and softening is illustrated in Figure 2.
As shown in Figure 2, strain hardening is demonstrated through a linear hardening model where strains increase with increasing stresses—though at a lower rate than below the yield point. In contrast, strain softening shows that strains continue to increase even when stresses are decreasing.
SimScale’s CEO David Heiny tests the capabilities of the platform to solve a real-life engineering problem. Fill in the form and watch this free webinar to learn more!
Important Terms in Plasticity
Consider a ductile material being subjected to a uniaxial tensile loading. As shown in Figure 3, as the load increases from Pt. 1 to Pt. 2, the material behaves in an elastic manner. In other words, upon unloading, the material follows the same stress-strain curve it followed during loading.
However, upon loading beyond Pt. 2, the material no longer behaves elastically. In other words, it follows a different path upon unloading. This is shown as unloading from Pt. 4. The point at which the material becomes inelastic is termed as the “Yield Point” and the corresponding stress as “Yield Stress”.
At any point, the stress is a symmetric second-order tensor or is represented by a symmetric matrix. This implies that there are six independent values. These stresses are defined in the global coordinate system. Now, if a different coordinate system is selected, the stresses need to be transformed using a transformation matrix.
If a transformation matrix is defined such that resulting stress matrix is diagonal, then the diagonal elements are the principal stresses. The three columns define the three vectors that form the new coordinate system.
The physical meaning is that the three planes on which these three vectors are normals are only subject to tensile loading—and not shear. The tensile loads on the three planes are provided by the three principal stresses.
Von Mises Stress
As we have already discussed, the stress tensor has six independent values. But as previously demonstrated, there is only one yield stress value obtained for a material. The question now becomes how to choose which stress value needs to be compared with the yield stress to determine if the material has yielded.
In this case, the uniaxial tensile loading was simple and can be treated as a one-dimensional problem. In this instance, only the stress along the axis of loading can be considered for comparison, but this does not hold true in a general case.
There are several criteria—which are available in literature—that can be used as a comparison with the yield stress. SimScale uses the von Mises yield condition. The von Mises stress can be expressed in terms of the stress tensor components as:
This von Mises stress is used as a comparison to check if the material has yielded. If the von Mises stress is larger than the yield stress, then the material can be said to have yielded.
Linear Hardening: Plasticity
At present, SimScale offers usage of isotropic hardening rules to describe plastic effects. Linear hardening means that beyond the yield point, the stress-strain relation is still linear. However, the modulus for loading is different from that of unloading.
As shown in Figure 4, the slope of the stress-strain curve beyond the yield limit is still positive, but it is still less than the original.
SimScale allows for a modeling plastic effect using the linear hardening model. As illustrated in Figure 5, the inputs are Young’s modulus, Poisson ratio, and von Mises stress, or the yield limit. The yield limit can also be provided through the input of experimental stress-strain data, and this is explained in more detail in the SimScale documentation.
Material Nonlinearity: Creep
Creep deformation is a time-dependent deformation. When a constant stress (or force) is applied, the strain initially reaches a particular level. Later on in the process, if the stress is continued to be applied, the strain gradually increases. This happens in spite of the constant force/stress. Such behavior is irreversible and more commonly observed in viscoelastic materials like rubbers and polymeric materials. Creep is not necessarily a damaging phenomenon. For example, creep in concrete allows for reduction of tensile stresses which could have otherwise led to cracks and fracture.
As shown in Figure 6, there are three distinct stages of creep. The primary stage demonstrates an initiation of the creeping process and is relatively slow. In contrast, the tertiary stage indicates the formation of necking and is considerably rapid. The secondary stage is when the material undergoes deformation in a reasonably stable manner and is well understood. SimScale facilitates creep deformation in the primary and secondary stages.
As observed above, to simulate creep phenomenon, a relation needs to be defined regarding how the strains change with time and stress applied. Several phenomenon-based models have been defined on the SimScale platform.
As demonstrated, all the formulations are equally stable. However, choosing the appropriate law depends on the availability of experimental results and which of the required parameters can be obtained.
- Norton is the simplest and depends only on the stress. With Norton, the power law is considered with m = 0 or the dependence on time is not considered.
- Time hardening uses the power law with dependence on both the stress and time considered, as shown in Figure 7.
- Strain hardening is more complicated. In addition to dependence on stresses, an additional dependence of “creep strain” is also considered as:
For more information on the additive decomposition of strains into elastic, plastic, creep, parts, etc. are discussed in more detail in the SimScale documentation.
Overall, SimScale allows modeling plastic and creep effects that are commonly observed in metals, polymers, rubber materials, etc. Each of these options already provides an excellent first-order approximation for modelling these phenomena in a wide variety of materials.
Discover all the simulation features provided by SimScale. Download the document below.