Time-domain Spectral Element Method for Impact Identification of Frame Structures using Enhanced GAs

This paper develops an Enhanced Genetic Algorithm (GA) strategy in conjunction with a time-domain Spectral Finite Element Method (SFEM) for impact identification of framed structures. For this purpose, a spatial truss spectral element is proposed for impact response simulation. In this regard, Gauss-Lobatto-Legendre quadrature rules and points configuration are adopted to construct a diagonal matrix and to gain an optimum computational demands. A decimal and a mixed GA coding system are implemented to locate and re-construct the impact features, respectively. An improved GA-based fitness assessment is designed to significantly accelerate the convergence rate of the optimization strategy. The impact identification of two frame structures is investigated taking into account the influence of externally applied loading. It is concluded that, the proposed mixed coding GA strategy effectively overcomes the main drawbacks of classical GA approach. The developed SFEM is superior to conventional FEM because of its high order interpolation and integration rules. For large structures, impact localization is successfully accomplished very fast, which provides the excellent ability in developing an on-line health monitoring system. It is included that, the robustness of the proposed SFEM lies on the considerably higher computational efficiency in achieving the most accurate results with the less computational costs.