Health Monitoring Method of Differently Tapered Structure Elastically Restrained with a Tip Mass

Cantilever-type longitudinally tapered structures are frequently used in high-rise steel structures. The current paper presents a fault identification technique for differently tapered beams that are elastically restrained and having a tip mass. For this purpose, a method is proposed for identifying the modal parameters of an intact beam by applying continuity and boundary conditions. Then, an equivalent bending rigidity for a beam with a crack is introduced and a characteristic equation is established to estimate the natural frequency change caused by the damage. An experimental study is conducted to verify the presented method. A baseline model is updated for the intact beam before detecting the crack by updating the rotational and translational spring constants. Crack identification is then carried out experimentally based on the neural network technique. The training patterns for the network are composed of the natural frequencies calculated from the derived characteristic equation for cracked beams, along with their corresponding crack sizes and locations. The cracks are identified using the trained neural network, and those are found to be reasonably well identified. The practicality and usability of the presented technique for health monitoring of the differently tapered cantilever-type structures elastically restrained having a tip mass could be thus verified.