The essential topics of classical and advanced adaptive finite element methods (FEMs) for linear and finite elastic deformations of solids and structures are presented in this chapter from both the mechanical and mathematical point of view.
As the mechanical and mathematical basis, nonlinear and linearized theories of elasticity are derived in a rigorous way, followed by the classical variational principles of elasticity, which are the basis for FEM, derived for classical and advanced Ritz–Galerkin methods, such as the extended finite element method (XFEM) and the meshfree reproducing kernel particle method (RKPM). Next to the discrete (one‐field) Dirichlet minimization principle, the (two‐field) Hellinger–Reissner stationary dual‐mixed principle and the (three‐field) Hu–Washizu stationary mixed principle are presented, including the main features of the associated FEMs.
The main objective of this chapter is the systematic treatment of
a posteriori
error estimators and their applications to both
verification
, that is, efficient mesh and particle adaptivity in order to achieve optimal convergence rates for the discretized solutions with prescribed error tolerances, and
validation
, that is, error‐controlled model adaptivity combined with
verification
.
In the framework of
verification
, linearized and finite elasticity problems are treated, using both energy‐norm and goal‐oriented
a posteriori
error estimators for mesh‐based and meshfree methods. The four basic classes, that is, residual‐type, hierarchical‐type, gradient averaging‐type, and constitutive equation error estimators are presented and applied, for example, to fracture mechanics problems.
A challenging topic in the framework of
validation
is error‐controlled expansive hierarchical modeling of 2D thin‐walled elastic plates and shells to 3D‐elastic continua by dimensional adaptivity of the underlying mathematical models, especially in subdomains with boundary layers and other disturbances where model expansion is necessary for validation.