Stéphane Andrieux

Identification of planar cracks by complete overdetermined data: inversion formulae

The problem of determining a crack by overspecified boundary data is considered. When complete data are available on the external boundary, a reciprocity gap concept is introduced. This concept formalizes the comparison of the response of the safe body to the response of the cracked one of the same physical characteristics. If the crack is known (or assumed) to be planar, explicit inversion formulae are derived determining the host plane equation and the length of an emerging crack in two-dimensional (2D) situations. A reciprocity gap functional is designed and exploited to establish a complete identification result. Numerical trials of the identification methods proposed show very good accuracies and insignificant computational costs.

Reciprocity principle and crack identification

In this paper we are concerned with the planar crack identification problem defined by a unique complete elastostatic overdetermined boundary datum. Based on the reciprocity gap principle, we give a direct process for locating the host plane and we establish a new constuctive identifiability result for 3D planar cracks.

On the inverse emergent plane crack problem

This paper deals with the detection of emerging plane cracks, by using boundary measurements. An identifiability result (uniqueness of the solution) is first proved. Then, we look at the stability of this solution with respect to the measurement. A weak stability result is proved, as well as a local Lipshitz stability result for straight cracks, by using domain-derivative techniques.