Parameter identification in constitutive models with a posteriori error control and sensitivity assessment

Licentiate thesis, 2005

In this thesis a framework for the calibration of constitutive models is presented. The framework involves the formulation of an optimization problem which is solved using strategies based on, essentially, Newton's method. The introduction of an additional costate field in order to incorporate the state equation has the advantage that error control in an arbitrary "goal" quantity is formally straightforward. The errors analyzed are those that arise from the Finite Element discretization and from the uncertainties in the experimental data. Among the applications, nonlinear viscoelasticity (Paper B) and moisture transport in wood (Paper C) are discussed in more detail.