Authors:
Ali Mansouri; Hosein Ghaffarzadeh; Majid Barghiana; Morteza H. Sadeghi;
Abstract:
This paper presents a novel approach to solve the wave propagation problem based on meshless method and its implementation for wave-based damage identification. Radial point interpolation method (RPIM) is a meshless technique which employs radial basis function (RBF) for shape function construction in the local support domain of the problem. RPIM satisfies the Kronecker delta property, which makes possible the enforcement of essential boundary conditions as easy as in finite element method. This study investigates the suitable parameters of RBF and dimension size of the support domain in longitudinal wave propagation. It is conducted by using minimum error between wave signals of benchmark and RPIM model of a rod. By considering the
best parameters in RPIM modeling, the minimization of discrepancies between the benchmark and RPIM signals leads to identify the damage. The optimization problem was formulated by a new error function and
Imperialist Competitive Algorithm (ICA) in such a way that the position, extent and length of damage are obtained. Various signal-to-noise ratios were concerned in the damage identification process to consider the measurement effect of practical manner.
Keywords:
Damage identification, Wave propagation, Meshless method, Radial basis function, Radial point interpolation, Optimization, Imperialist competitive algorithm