Investigation of Transverse Cracks with Different Orientations in GFRP Beam Through Modal Data Based ANN Model

Glass fiber reinforced polymer (GFRP) composite structures are extensively utilized across the globe due to their lightweight, corrosion resistance, high specific strength and stiffness. Generally, fatigue failures are common in composite structures such as aircraft structures, mechanical components, windmill structures, etc. The crack initiates and propagates in relative orientation between the crack and loading direction which adversely affects the performance of composite structures. Therefore, it is essential to detect the crack location and orientation to avoid catastrophic failure. This research article explores the investigation of transverse cracks with different orientations in GFRP composite beams using a modal data-based Artificial Neural Network (ANN).

The composite beam laminate is fabricated using vacuum-assisted resin transfer molding with bi-directional GFRP lamina. Crack with consistent depth and triangular shape made on the specimen using a hacksaw. Experimental modal analysis is carried out on four beam specimens with different damage conditions such as without crack and transverse crack with 30, 60, and 90-degree orientations under cantilever boundary conditions. Further, ANN is applied to the modal parameters to predict the frequency response functions (FRFs).

To comprehend the specimen’s behavior for notable changes, modal parameters such as natural frequencies, mode shapes, damping ratios and FRFs are acquired and briefly examined for various experimental cases. Then, FRFs for all four cases are predicted using ANN, and the accuracy of the model is computed.

It is observed that for the fundamental mode, natural frequencies decrease and damping ratios increase respectively with the formation of crack. The predicted FRFs using ANN have agreed well with the experimental FRFs for all different criterion.