The paper presents the methodology that makes it possible to evaluate computational model and introduce current corrections to it. The methodology ensures proper interpretation of nonlinear results of numerical analyses of thin-walled structures. The suggested methodology is based on carrying out, in parallel to nonlinear numerical analysis, experimental research on some selected crucial zones of loadcarrying structures. Attention is drawn to the determinants concerning the performance of an adequate experiment. The author points out on indicating the role of model tests as a fast and economically justified research instruments practicable when designing thin-walled load-carrying structures.

The presented considerations are illustrated by an example of a structure whose geometrical complexity and ranges of deformation are characteristic for modern solutions applied in the load-carrying structures of airframes. As the representative example, one selected the area of the load-carrying structure that contains an extensive cut-out, in which the highest levels and stress gradients occur in the conditions of torsion evoking the post-buckling states within the permissible loads. The stress distributions within these ranges of deformations were used as the basis for determining the fatigue life of the structure.

The work concerns numerical – experimental studies on pre- and post-buckling of thin-walled, steel, cylindrical shells, with the open section, subjected to constrained torsion. Two geometrically varied structures are considered: an open section cylindrical shell without stiffeners and one that is reinforced by closed section stringers. The shells have five different length to diameter ratios. Numerical simulations were carried out and the neuralgic zone stress distributions in pre- and post-buckling responses, were determined. Torsion experiments were performed and the results were compared to the numerical conclusions, with reasonably high level of agreement. The exactness of the experiment was proven for selected cases, establishing the basis for FEM numerical model estimation.