Numerically disturbed shake table experimentation to examine nonlinear signal-based control

Abstract

This study introduces a numerically disturbed experimentation to address the shake table control degradation commonly observed in shake table experiments. This degradation is caused by nonlinear characteristics, such as seismic damage, in the experiments; however, observing such nonlinear characteristics is a major purpose of these experiments. In the proposed numerically disturbed experimentation, a structure is numerically simulated, and its structural responses are fed back as the disturbance signal to the table in the physical domain via real-time interaction. This enables us to examine the control performance of a shake table with a structure, without having to place an actual structure on it. This experimentation is beneficial in cases wherein new control methods are applied for shake table control because the control performance can be examined safely and efficiently under various structural conditions by using numerical simulations. The proposed experimentation was applied to the shake table control examination of nonlinear signal-based control (NSBC), which has a nonlinear signal feedback action for nonlinear structural dynamics, as well as inversion-based control (IBC), which is a common feedforward method. In the numerically disturbed experiments, NSBC accurately realized a seismic acceleration record on the shake table with severe nonlinear characteristics, whereas IBC exhibited control degradation due to nonlinear characteristics. Similar results were obtained using actual shake table experiments with a steel structure. Therefore, the proposed numerically disturbed experimentation can be an alternative to shake table experiments using structures.