Use of genetic algorithms for the identification of optimal system of added viscous dampers for shear-type buildings

This paper illustrates the results obtained using genetic algorithms for the identification of the “optimal” systems of added viscous dampers for the mitigation of the seismic action upon shear-type building structures. The identification of the “optimal” damping systems is here performed according to Holland’s “simple genetic algorithm” taking into consideration shear-type structures and the “average (over all storeys) of the standard deviations of the inter-storey drift angles” of the system to a white noise stochastic process as objective function. In the search the damping matrix is left free, so that (a) classical and non-classical damping systems can be considered at once and (b) no constraint upon the damper placement is imposed. All systems considered in the comparison satisfy a mathematical constraint which imposes that the sum of the damping coefficients of all added viscous dampers is the same (“equal total size” constraint). The dynamic response of the above systems (shear-type structures + added viscous dampers) to both stochastic input and selected earthquake ground motions are here used to evaluate, compare and assess the dissipative properties offered by different damping schemes. The results show that placing dampers so that they connect each floor of the structure to a fixed point (ground or infinitely-stiff adjacent structure) is capable of providing the best overall dissipative properties. The large damping capacities of such damper arrangement can be basically linked to the implementation in building structures of a damping matrix proportional to the mass matrix (mass proportional component of the Rayleigh damping matrix), which had proved, in previous research works by the authors, to lead to very good dissipative performances.