An overview of control schemes for hydraulic shaking tables

Abstract

Shaking table testing is a common experimental method in earthquake engineering for performance assessment of structures subjected to dynamic excitations. As most shaking tables are driven by servo hydraulic actuators to meet the potentially significant force stroke demand, the review is restricted to hydraulic shaking tables. The purpose of the control systems of hydraulic shaking tables is to reproduce reference signals with low distortion. Accurate control of actuators is vital to the effectiveness of such apparatus. However, the system dynamics of a shaking table and the specimens to be tested on the shaking table are usually very complex and nonlinear. Achieving the control goal can prove to be challenging. A variety of closed- and open-loop control algorithms has been developed to solve different control problems. With the focus placed on the control schemes for hydraulic shaking tables, the paper reviews algorithms that are currently used in the testing industry, as well as those which are the subject of academic and industrial research. It is by no means a complete survey but provides key reference for further development.

Acceleration Harmonics Identification for an Electro-Hydraulic Servo Shaking Table based on a Nonlinear Adaptive Algorithm

“…The most obvious distinction between harmonic identification for electro-hydraulic servo shaking table system and for power system is that the former requires better real-time and precision performance [11][12]. The LMM algorithm, a LMS-like algorithm, uses a more robust "M-estimator" to replace LMS for the purpose of harmonic identification.…”

Experimental evaluation of acceleration waveform replication on electrohydraulic shaking tables

“…Controllers based on variable feedback compensation are widely termed as a TVC. 2,[18][19][20][30][31][32][33][34][35][36] A block diagram of the TVC is shown in Figure 4, 35 where K vf , K df , and K af are three feedback parameters, and K dr , K vr , and K ar are three feed-forward parameters, respectively. Displacement is measured by an LVDT and acceleration is measured by an accelerometer mounted on the table, and velocity is synthesized from the measured displacement and acceleration.…”

“…Parameters inÂðzÞ andBðzÞ can be identified by the RELS algorithm and z g is the time delay for physical realizability ofĜ In order to obtain accurate acceleration tracking performances, an FIMC of the EHST was employed to cancel out dynamic characteristics of shaking tables in the literatures 9,[18][19][20]22,34,35,37,39,44 because the FIMC can expand the frequency bandwidth of the acceleration closed-loop system without changing its stability. 84 Lee et al 38 employed an acceleration inverse transfer function of a shaking table to cancel out its dynamic characteristics.…”

“…Due to potential advantages of the EHST in power density, large forces, high-fidelity position and acceleration tracking accuracy, [16][17][18][19] high durability and stiffness, a rapid and wide frequency output response, and more accurate vibration tests, many investigations have been carried on their academic studies and industrial applications. 2,20 However, dynamic characteristics of acceleration output responses on the EHST would not be intended in its magnitude and phase 21,22 due to inherent nonlinearities and uncertainties of the EHST, such as dynamics of servovalves and hydraulic actuators. 17,[23][24][25][26] Besides, disturbed reaction forces generated by a specimen deteriorate, 27,28 dynamics of the base support, 29 and internal coupling force in the EHST, 16,17,22 and so on, affect acceleration tracking control performances.…”

“…Many compensation approaches for acceleration waveform replication on the EHST have been employed to minimize nonlinear behaviors of electrohydraulic servo systems, such as off-line feed-forward compensation controllers including a three-variable controller (TVC), 2,[18][19][20][30][31][32][33][34][35][36] a feed-forward inverse model controller (FIMC), 2,18,19,22,[34][35][36][37][38][39] and an off-line iterative controller (OIC) 2,20,34,36,[40][41][42][43][44] and online adaptive controllers including adaptive inverse control (AIC), 18,21,22,33,35,37,[44][45][46][47][48] minimal control synthesis (MCS), 2,18,37,39,40,[49][50]…”