Dynamic Stiffness Prediction in Cracked Cantilever Beams Using Enhanced ANN Techniques

This research paper provides a comprehensive study on how Artificial Neural Networks (ANNs) can be deployed to predict the stiffness characteristics of a cantilever beam with a crack of various depths and positions. The most destructive source of failure is considered to be vibration, so the major focus of this paper will be on how the cracks affect the modal stiffness. This study has various applications, such as airplane wings, bridges, stadiums, and arenas. A common research gap was noticed amongst the existing studies; the position of the cracks in the cantilever wasn’t considered, but this paper discusses how the location of cracks severely affects the dynamic behaviour of the cantilever. This study was done by carrying out modal analysis on a cantilever of the same dimensions with different crack configurations. Various crack dimensions and orientations were analysed to understand the effects of the crack on the dynamic behaviour of the cantilever. From the modal analysis results, we evaluated the natural frequency of the cantilever beams with various crack depths and locations. A decrease in natural frequency was observed as the crack depth increased, from which we can infer that the cantilever will experience resonance at much lower external vibration, which makes the structure unreliable. Cracks near the support markedly lower natural frequency due to maximum shear force and bending there, whereas free-end cracks have a negligible impact compared to a reference cantilever. Simulation results feed an ANN, enabling it to accurately predict dynamic characteristics for any combination of crack depth and position. The developed ANN model achieved high prediction accuracy with a Mean Squared Error (MSE) of less than 1x10^-5 and an R² value exceeding 0.998 on the test dataset, demonstrating its robustness as a tool for structural health monitoring.