Title: Vector evaluated particle swarm optimisation for multi-objective structural system identification
Authors: Cibu K. Varghese; Sunny K. George; K. Shankar
Addresses: Department of Mechanical Engineering, Mar Athanasius College of Engineering, Kothamangalam, India ' Department of Mechanical Engineering, Mar Athanasius College of Engineering, Kothamangalam, India ' Department of Mechanical Engineering, IT Madras, India
Abstract: This paper presents the application of VEPSO - a cross searching multi-objective technique - to an inverse problem for identifying structural parameters and damage detection from time domain responses. The aim here is to minimise both instantaneous vibrational power flow balance and acceleration deviations. Whereas minimisation of acceleration deviations is fairly conventional, minimising the additional parameter of power flow balance is a novel concept used with VEPSO here. Instantaneous substructural power flows are balanced by equating the input power to the dissipated and damping powers and the time rate of change of kinetic and strain energies. Numerical simulations are performed for substructures taken from a lumped mass system for parameter identification and a cantilever beam for structural identification and multiple crack damage detection. In these numerical simulations, noise free and noise contaminated response measurements are considered. The results demonstrate that the VEPSO method can be accurately used for identifying the structural parameters and detecting damages of a system.
Keywords: instantaneous power flow; substructures; system identification; vector evaluated PSO; particle swarm optimisation; VESPO; multi-objective optimisation; mathematical modelling; numerical simulation; structural parameters; crack damage detection; kinetics; strain; cantilever beams.
International Journal of Mathematical Modelling and Numerical Optimisation, 2015 Vol.6 No.1, pp.40 - 71
Received: 18 Jun 2014
Accepted: 24 Sep 2014
Published online: 16 Apr 2015 *