Performance of damage detection algorithms for health monitoring of joints in steel frame structures using vibration-based techniques

Abstract

Steel frame structures are widely used for industrial buildings, bridges, off-shore structures, etc., due to their high reliability and less construction time. In these types of structures, beams and columns are often connected by welding or fasteners forming rigid joints. In this study, viability of several damage detection algorithms for health monitoring of joints in planar frame structures using vibration-based techniques is addressed. Planar frames are modelled in the finite element framework with two node beam elements having two degrees of freedom at each node. The joint damage is introduced by reducing the cross section of an element attached to the joint making it semi-rigid. Several damage detection algorithms, with and without baseline information, are applied to mode shapes and frequency response functions (FRFs) of the structures. Synthetically zero mean Gaussian noise of different level is added to the simulated displacement data to examine the effectiveness of the algorithms for identification of joint damage in both the single and two storied planar frames. Experiment is then performed in a laboratory model of a signal storied planer frame. FRF curvature-based damage detection algorithm is found to be most effective among all the algorithms addressed.