A feed-back approach to error control in finite element methods: Application to linear elasticity (Q1372141)

The authors describe a general framework for reliable a posteriori error bounds for finite element solutions in linear elasticity. The key ingredients of this feed-back approach are the representation of the error through a duality argument, the use of the Galerkin orthogonality, and the utilisation of the dual solution in a posteriori error estimates in terms of weights multiplying the local residuals of the solution. The resulting sharp error estimates provide a basis for adaptive construction of almost optimal meshes. This is demonstrated in a practically relevant model situation. The described approach seems to be especially useful if only local error quantities have to be controlled, in contrast to the conventional global error control in energy norm. In a subsequent work the authors promise to apply this method to nonlinear problems in structural mechanics, e.g. to problems in elasto-plasticity theory with material laws of Prandtl-Reuss type.