Adaptive neuro-fuzzy-based vibration approach for structural fault diagnosis

Abstract

In recent times, damage identification based on vibration methods are emerging as common approaches, in which the techniques apply the vibration response of a monitored structure, such as modal frequencies and damping ratios, to evaluate its condition and detect structural damage. The basis of the vibration-based health monitoring method is that when there are alterations in the physical characteristics of a structure, there will also be changes in its vibration properties. This paper proposed a neuro-fuzzy artificial intelligence method, called adaptive neuro-fuzzy inference system (ANFIS), to detect damage using modal properties. To generate the modal characteristics of the structures, experimental study and finite element analysis of I-beams with single damage cases were performed. The results showed that the ANFIS approach was able to detect the magnitude and location of the damage with a significant degree of precision, and notably reduced computational time.