Model‐based multi‐metric control of uniaxial shake tables

Abstract

SUMMARY Shake tables provide a direct means by which to evaluate structural performance under earthquake excitation. Because the entire structure is mounted on the base plate and subjected to the ground motion in real time, dynamic effects and rate‐dependent behavior can be accurately represented. Shake table control is not straightforward as the desired signal is an acceleration record, while most actuators operate in displacement feedback for stability. At the same time, the payload is typically large relative to the capacity of the actuator, leading to pronounced control‐structure interaction. Through this interaction, the dynamics of the specimen influence the dynamics of the shake table, which can be problematic when specimens change behavior because of damage or other nonlinearities. Moreover, shake tables are themselves inherently nonlinear, making it difficult to accurately recreate a desired acceleration record over a broad range of amplitudes and frequencies. A model‐based multi‐metric shake table control strategy is proposed to improve tracking of the desired acceleration of a uniaxial shake table, remaining robust to nonlinearities including changes in specimen condition. The proposed strategy is verified for the shake table testing of both linear and nonlinear structures. Copyright © 2013 John Wiley & Sons, Ltd.

Substructure Shake Table Testing with Force Controlled Actuators

“…Since shake tables use hydraulic actuators to produce motion, this compatibility requires high performance acceleration tracking of shake tables. While this is a challenging issue, there has been significant progress in the enhancement of acceleration tracking of shake tables [namely: Stehman and Nakata (2012) and Phillips et. al.…”

Generalised formulation of composite filters and their application to earthquake engineering test systems

“…In Stoten [1], it was found that the corresponding minimum order could be n = 2 or 3, depending upon the CF; hence, the order is now fixed as n = 3 for all CF. Accordingly, the sum of terms on the left-hand side of both (8) and (9) is required to be equal to a third-order monic transfer function defined by:…”

“…Examples of modified DAC strategies, which incorporate methods of circumventing the low frequency control and drift problems, include high fidelity adaptive control [5], acceleration feedforward control with displacement feedback via Kalman filter estimation [6], control augmentation via force feedback [7] and feedforward control with multi-metric feedback [8]. One advantage of DAC is that it allows good fidelity of signal resolution at high frequencies, compared with (say) the resolution provided by a conventional displacement control strategy.…”

Development of high‐performance shake tables using the hierarchical control strategy and nonlinear control techniques

“…The use can be traced back to the first hand-powered table created in the 19th century [1]. These challenges become more problematic when a specimen exhibits nonlinear behavior and its mass is large relative to that of the table [3]. Although the testing capacity has been greatly improved, high-fidelity shake table control remains to be challenging because of several difficulties such as the inherent hydraulic actuator nonlinearity and control-structure interaction (CSI) effects [2].…”

“…A controller that deals with nonlinearity of shake table dynamics and CSI is needed. Moreover, it was difficult to preserve the stability of a shake table acceleration tracking control by acceleration feedback alone [3]; hence, displacement-based shake table control gained popularity. Spencer and Yang [4] adopted a linearized shake table model (a transfer function) to describe the relationship of command signals and measured acceleration.…”