Towards numerical simulation of fatigue damage detection in CFRP laminates by modelling of stochastic vibration response

Abstract

Due to the widespread use of composites in aerospace applications, there is an increasing need for developing Structural Health Monitoring (SHM) systems for composite structures. In this process, simulation models can play a crucial role by contributing to the optimization of the experimental techniques and to understanding of the physical mechanisms. The objective of this work is to numerically simulate fatigue damage detection in laminated composites subjected to random vibration through a white gaussian noise. To this end, the ANSYS FE code has been used. The developed FE model represents an experimental set-up consisting of the composite specimen, the speakers producing the white noise and the measuring equipment. 24 specimens were modelled in total, a healthy one and 23 damaged ones. The white noise was modelled by introducing a normalized sequence of numbers adapted to the free vibration eigenfrequency range. Fatigue damage in the form of matrix cracking and delamination was modelled based on C-Scan images of fatigued specimens by degrading the appropriate elastic properties of the layers in specific groups of elements. Damage of different geometry, size, and type was modelled. The numerical results are in the form Power Spectral Density (PSD) diagrams, natural frequencies, eigenfrequencies and contour plots. The comparison between the numerical and experimental eigenfrequency values show a maximum deviation of 1.7%. The numerical results show a change in the vibration response even in the presence of a small damage, which is an indication of the sensitivity of the method. A significant decrease in natural frequencies of the specimen with the damage accumulation is observed. The present model comprises the first effort towards a fully validated digital twin which will be used for the virtual testing and the optimization of the vibration test.