Damage identification on spatial Timoshenko arches by means of genetic algorithms

• Published in: Mechanical systems and signal processing Vol. 105; pp. 51 - 67
• Main Authors: Greco, A., D'Urso, D., Cannizzaro, F., Pluchino, A.
• Format: Journal Article
• Language: English
• Published: Berlin Elsevier Ltd 15.05.2018; Elsevier BV
• Subjects: Arches; Boundary conditions; Chromosomes; Damage assessment; Damage detection; Damage identification; Eigenproperties; Fracture mechanics; Genetic algorithms; Natural frequencies; Optimization; Property damage; Studies; Timoshenko model; Damage identification; Timoshenko model; Arches; Natural frequencies; Genetic algorithms
• ISSN: 0888-3270; 1096-1216
• DOI: 10.1016/j.ymssp.2017.11.040

•A procedure for the dynamic damage identification in spatial arches is presented.•Both small and large extents of damage and possible losses of mass are considered.•An eigen-properties based objective function is used in a genetic algorithm procedure.•Applications with different number and extents of damaged parts are presented.•The procedure is tested also in presence of instrumental errors on the measured data. In this paper a procedure for the dynamic identification of damage in spatial Timoshenko arches is presented. The proposed approach is based on the calculation of an arbitrary number of exact eigen-properties of a damaged spatial arch by means of the Wittrick and Williams algorithm. The proposed damage model considers a reduction of the volume in a part of the arch, and is therefore suitable, differently than what is commonly proposed in the main part of the dedicated literature, not only for concentrated cracks but also for diffused damaged zones which may involve a loss of mass. Different damage scenarios can be taken into account with variable location, intensity and extension of the damage as well as number of damaged segments. An optimization procedure, aiming at identifying which damage configuration minimizes the difference between its eigen-properties and a set of measured modal quantities for the structure, is implemented making use of genetic algorithms. In this context, an initial random population of chromosomes, representing different damage distributions along the arch, is forced to evolve towards the fittest solution. Several applications with different, single or multiple, damaged zones and boundary conditions confirm the validity and the applicability of the proposed procedure even in presence of instrumental errors on the measured data.