The dynamic instability of a thin-walled carbon-fibre reinforced composite cylindrical shell is studied. The analysis is performed with the Finite Element code, ABAQUS, estimating the dynamic buckling load using the Budiansky–Roth criterion. The influence of the following factors on the dynamic behaviour of the shell is analysed: the shape of pulse loading, the initial geometric imperfection and the pulse duration. It is found that for short load duration, the structure resistance to pulse loading in the form of dynamic buckling load is significantly higher compared to the static buckling load. As load duration increases, the magnitude of the Dynamic Load Factor (DLF), defined as the ratio between the dynamic and static buckling loads, decreases significantly, reaching a value of DLF<1 in the vicinity of the natural frequency of the shell. The results of the numerical analyses indicate a slight increase of the DLF with the increase of the initial geometric imperfection of the shell. The present study highlights the increased sensitivity to the shape of the pulse loading. For triangular and double-triangular pulse shapes, for short load duration, the dynamic buckling load is almost 14 times higher than the static buckling load. Simultaneously, when a trapezoidal pulse shape is applied, the dynamic buckling load is four times greater than the static one.

Two laminated composite shells, one with a conventional straight fiber laminate denoted the classical laminated shell and the second one with a variable angle tow reinforced composite, had been excited and their natural frequencies and mode shapes had been measured and monitored as a function of the axial compression load. Then, the in-situ buckling loads of the two tested specimens were predicted using the Vibration Correlation Technique (VCT) and compared with actual experimental buckling loads and Finite Element buckling predictions, yielding matching, consistent and repeatable results. It was shown that the VCT predicts the actual in-situ buckling loads of laminated composite thin walled cylindrical shells with a high accuracy, yielding 96% and 98.6% of the experimental buckling load, for the classical and variable angle tow composite shells, respectively. These results, although based on only two specimens, join the relatively small data base published in the literature, proving the nondestructive nature of the VCT approach, making it an adequate method for application on thin-walled structures, like shells. In addition, some testing recommendations are presented, to effectively enable the successful application of the VCT for in-situ buckling prediction of the buckling sensitive structures, like composite cylindrical shells.