Vibration-based delamination evaluation in GFRP composite beams using ANN

Delamination is definitely an important topic in the area of composite structures as it progressively worsens the mechanical performance of fiber-reinforced polymer composite structures in its service period. The detection and severity analysis of delaminations in engineering areas like the aviation industry is vital for safety and economic considerations. The existence of delaminations varies the vibration characteristics such as natural frequencies, mode shapes, etc. of composites and hence this indication can be effectively used for locating and quantifying the delaminations. The changes in vibration characteristics are considered as inputs for the inverse problem to determine the location and size of delaminations. In this paper Artificial Neural Network (ANN) is used for delamination evaluationof glass fiber-reinforced composite beams using natural frequency as typical vibration parameter. The Finite Element Analysis is used for generating the required dataset for ANN. The frequency-based delamination prediction technique is validated by finite element models and experimental modal analysis. The results indicate that the ANN-based back propagation algorithm can predict the location and size of delaminations in composites with good accuracy for numerical natural frequency data but the accuracy is comparitivelyless for experimental natural frequency data.