Metaheuristic based optimization of tuned mass dampers on single degree of freedom structures subjected to near fault vibrations

Abstract

Near fault ground motions excitations have specific characteristics comparing to regular earthquake excitations. Near fault ground motions contain directivity pulses and flint steps in different directions and these excitations are the reason of more damages than regular excitations for structures. A successful method to reduce structural vibrations is the usage of tuned mass dampers. By using optimally tuned mass dampers, it will be possible to reduce vibrations resulting from earthquake excitations. In the present study, the optimization of tuned mass dampers are done for near fault excitations. During optimization, 6 different pulse like excitations are used. Three of these excitations are directivity pulses while the other ones are flint steps. The periods of excitations are 1.5s, 2.0s and 2.5s since near fault pulses have long period and big peak ground velocity around 200 m/s. The optimization objectives are related to maximum displacement of structure in time domain, the maximum stroke limitation of tuned mass damper and transfer function of the structure in frequency domain analyses. The iterative optimization process uses both time and frequency domain analyses of the structure. Three different metaheuristic algorithms are used in the methodology. These methods are harmony search algorithm, teaching learning based optimization and flower pollination algorithm which are inspired from musical performances, education process and reproduction of flowering plants, respectively. As the numerical investigation, three different single degree of freedom structures with periods 1.5s, 2.0s and 2.5s are investigated for optimum mass, period and damping ratio of a tuned mass damper positioned on the structure.