Nonlinear signal-based control for shake table experiments with sliding masses

Abstract

This study introduces the first application of nonlinear signal-based control (NSBC) to shake table experiments with sliding masses. NSBC utilising a nonlinear signal was specifically developed for nonlinear system control. In shake table experiments with a steel structure, it realised a seismic acceleration record on the table with near 100% accuracy even when the structure displayed nonlinear characteristics due to yielding of its components. Regarding the severity, sliding is stronger than yielding because of the rapid shifts of stick and slip states. Sliding is utilised for structural damage mitigation in seismic isolation systems, which are prominent devices in earthquake engineering. However, when sliding occurs on a shake table, it significantly jeopardises the table control. Thus, this study investigates the performance of NSBC to shake table experiments involving the phenomenon. First, this study examines the effectiveness of NSBC via numerical simulations on a shake table with a sliding interface, demonstrated by the Karnopp model, with a friction coefficient of 0.2. Subsequently, a linear model is designed based on the stability analysis of NSBC to assess stability margins obtained from different models. Experiments are performed by placing masses weighing approximately twice the table weight on sliding interfaces with friction coefficients of 0.2–0.4. In these experiments, NSBC with a reasonable linear model design realised expected acceleration records with sufficiently high accuracies, whereas inversion-based control failed. This study verifies the effectiveness of NSBC for shake table experiments with sliding masses.