Modal identification of localised damage in beams and trusses: experimental and numerical results

Abstract

The paper discusses the possibility of detecting local damages in complex structures typical of civil engineering, as multispan beams and trusses. Namely, it describes a procedure to identify localised cracks in structures in the elastic range of behaviour using only the values of natural frequencies in the intact configuration and in the damaged one evaluated by means of dynamic tests. The error minimisation procedure described in the paper selects the solution within a set of finite element models that simulate a range of positions and levels of damage, by identifying the damaged configuration as the one whose modal frequencies minimise the least-square difference with the measured data. The accuracy of the method is first investigated by applying it to the damage detection of a two-span steel beam, whose modal frequencies were obtained by means of experimental tests. To explore the accuracy of the proposed procedure, numerically simulated data with random noise were also generated for several positions and levels of damage and for different values of the random noise. The procedure was then extended, by means of numerical simulations, to the case of a beam with two localised damages. Finally, the procedure proposed for multispan beams is adapted to the damage identification of plane truss structures.

Fatigue Damage Identification by a Global-Local Integrated Procedure for Truss-Like Steel Bridges

“…In 2012, Beskhyroun et al propose a method based on wavelet transforms using dynamic responses recorded under excitation applied by an actuator, but this procedure shows some limitations under operating conditions [9]; Bowe et al presented in 2015 a method that exploits the analysis of the acceleration of vehicles resulting from the train-track interaction, but this method is not applicable in the presence of noise [10]; in the same year Gonzales and Karoumi studied a strategy based on artifcial neural networks (ANN) but the limits of this methodology lie in modeling the railway bridge as a simply supported beam that does not describe the complexity of the structure [11]. A review of the latest literature is listed in the bibliography [12][13][14][15][16][17][18][19][20][21]. It is known that most of these methods have identifed the existence of damage (the frst level of the SHM process), without fully reaching the goal of level 2 which is the localization of damage.…”