Search

Abstract This paper presents a synthesis strategy for impact force localization and identification of framed structures in time-domain using a two-step wavelet-based fitness evaluation scheme in conjunction with genetic and water cycle algorithms. For this purpose, a straightforward approach is developed for sensitivity analysis of accelerations and spatial signal connecting the peak values. The proposed scheme is capable of using diverse scales of wavelet functions at considerably small sampling rates. A decimal genetic algorithm (GA) coding system and a recently developed water cycle algorithm (WCA) are improved to be used for impact localization and identification, respectively. The fitness evaluation is then modified for using the obtained results from sensitivity analysis rather than the acceleration itself. Numerical study is first conducted for a large space frame to evaluate the precision, convergence, and efficiency of the identified results using the proposed GA-WCA strategy to different measurement scenarios and initial estimates of the structural model. For comparison purposes, the robustness of WCA in identification step is thoroughly compared with other three state-of-the-art optimization algorithms i.e., particle swarm optimization (PSO), imperialist competitive algorithm (ICA), and differential evolution (DE). Afterwards, an experimental validation study is carried out on a laboratory scale truss bridge. It is concluded that the computational performance of the proposed method is significantly better than the existing methods with respect to fitness evaluation. Results show that, even for the cases with a rough knowledge on structural parameters, the impact force is successfully identified with an excellent precision. This demonstrates the superiority of WCA strategy in handling the global search over large design space with large number of design variables. It is also concluded that the impact localization step is accomplished very fast, thus providing a near real-time strategy in dealing with large-scaled frame and bridge structures. Highlights • A near real-time method is achieved for impact localization of large systems. • A wavelet-based and multi-species decimal genetic algorithm (DGA) is improved. • Impact identification step is satisfactorily achieved using a water cycle strategy. • An efficient wavelet-based sensitivity approach is proposed for fitness evaluation. • Impact identification of a large-scaled space frame is examined. [ABSTRACT FROM AUTHOR]