Offline iterative control method using frequency-splitting to drive double-layer shaking tables

摘要

Large shaking tables with payload capacities exceeding 1000 t can be used to realistically evaluate the seismic performance of engineering structures. The working frequency of large shaking tables is typically limited to 0-30 Hz. This range is not sufficient for testing rigid engineering structures, such as dams and nuclear facilities, because these scaled structures usually have high fundamental frequencies. A new configuration has been proposed, in which a high-frequency shaking table is mounted atop a low-frequency shaking table. Referred to as a double-layer shaking table (DLST), the lower shaking table has a long stoke to reproduce large displacements and velocities in a low-frequency range, and the upper shaking table has a short stroke to reproduce high-frequency accelerations. With this configuration, the large payload capacity of the lower shaking table can be utilized while also achieving a high-frequency excitation. The control of such a complex system, however, is challenging because of strong interactions between the two tables. This paper proposes an offline iterative control method based on frequency-splitting to drive the two tables simultaneously. An input signal with a wide frequency spectrum is first split into two signals, i.e., one containing the low-frequency components and the other containing the high-frequency components. Successfully applying this control technique requires development of the transfer function matrix of the interacting tables by system identification. Numeric simulations and experiments demonstrate that this method can effectively drive a DLST to reproduce a target signal with a wide frequency range.