MUSIC‐ANN Analysis for Locating Structural Damages in a Truss‐Type Structure by Means of Vibrations

Abstract

Abstract: This article presents a methodology for damage detection, location, and quantification based on vibration signature analysis and a comprehensive experimental study to assess the utility of the proposed structural health monitoring applied to a five‐bay truss‐type structure. The MUSIC algorithm introduced first by Jiang and Adeli for health monitoring of structures in 2007 is fused with artificial neural networks for an automated result. The developed methodology is based on feeding the amplitude of the natural frequencies as input of an artificial neural network, being the novelty of the proposed methodology its ability to identify, locate, and quantify the severity of damages with precision such as: external and internal corrosion and cracks in an automated monitoring process. Results show the proposed methodology is effective for detecting a healthy structure, a structure with external and internal corrosion, and a structure with crack. Therefore, the proposed fusion of MUSIC‐ANN algorithms can be regarded as a simple, effective, and automated tool without requiring sophisticated equipment, toward establishing a practical and reliable structural health monitoring methodology, which will help to evaluate the condition of the structure, in order to detect damages early and to make the corresponding maintenance decisions in the structures.