Using Deep Learning to Estimate Vibration Comfort of Large‐Scale Shake Table During Operation

Shaking table testing has been regarded as one of the most straightforward experimental approaches to evaluate the seismic response of structures subjected to earthquake ground motions. Therefore, reproducing an acceleration time history accurately becomes crucial for shaking table testing. In this study, a control framework for uniaxial shaking tables is proposed which incorporates a feedback controller into a weighted command shaping controller. It implements through outer-loop control in addition to the conventional existing proportional-integral inner-loop control. The model-based command shaping controller which considers the control-structure interaction can be designed to shape either displacement or acceleration references. The weightings for the shaped displacement and acceleration can be calculated by a linear interpolation algorithm which considers the dominant frequency of the desired acceleration time history as well as the correlation between the displacement and acceleration responses of the shaking table. Accordingly, the weighted combination of the shaped displacement and acceleration generates the control command to the shaking table. On the other hand, the feedback controller deals with the system uncertainty and modeling error. Loop-shaping design method is adopted to synthesize the feedback controller. Finally, the control framework is verified by several shaking table tests with and without a flexible specimen. Experimental results demonstrate the performance and robustness of the proposed control framework for shaking table test systems.

The Shanghai Tower is a newly built 127-story and 632 m high super tall building. As of April 2017, it was ranked as the second tallest building in the world. Its main structural system is a mega-frame-tube-outrigger system with six outrigger trusses along the height. Due to its unique structural configuration, a series of field and laboratory model tests have been conducted to better understand its dynamic characteristics. Before its construction, a scaled model of the tower was tested on a shake table, and the results were used to refine the design of the tower. At the completion of the construction, full-scale ambient vibration tests were performed. A Bayesian method was used to perform operational modal analysis from the shake table and full-scale ambient vibration tests. The most probable value of the modal parameters and the associated posterior uncertainties were calculated using this method. The first eight modes were identified, including three translational modes in each principal direction and two torsional modes. Using these results, the dynamic characteristics and associated uncertainties obtained from the two tests were investigated and compared in this paper. Due to the scaling of the model, there are some discrepancies between the natural frequencies obtained from two different tests, but the identified mode shapes matched very well. Although the structure was designed in a very innovative manner, its dynamic characteristics are similar to regular tall buildings. The results from this investigation provide valuable information for an ongoing condition assessment of this super tall building.

This paper presents a study of shake table testing for aseismic design of structure systems. Two reinforced concrete structure models of different scales were used first to implement and demonstrate the similitude concept using available testing equipment. This concept was then applied to a 33‐story reinforced concrete building's scaled model, using multidirectional simulated seismic excitations. The test results were used to evaluate and modify the original design. Some of these results were also verified using field testing after the building's construction was completed. Comparison of the results between the model test and the full‐scale test shows that testing based on dynamic similitude using shake table are useful although not perfect and can provide supplementary data for decision making in engineering design. Copyright © 2007 John Wiley & Sons, Ltd.

Floor motions in structures are substantially different than the ground motions exciting their foundations. Vibration amplifications of ground motions occur differently in the vertical and horizontal directions in these floors. Excessive motions influence adversely most of building contents and non-structural systems suspended on those floors. Shake table testing allows reproducing true effects of earthquake motions that can challenge a complex specimen, but their reproduction is usually imperfect because of the dynamics of the control system, the shake table controllers and the specimen dynamics, even when basic closed-loop control is used. This paper presents the analytical development and the experimental verification of an ‘openloop’ feed-forward iterative procedure for simulation of desired “floor motions” at any specific location of a test structure. As the first step, a well-known global feedforward procedure using transfer functions is reformulated for uni-axial systems. Subsequently, the same concept is extended to multi-directional systems using a newly developed 6×6 transfer function matrix. The compensation procedure was implemented and experimentally demonstrated, assuming uncoupled system in each axis of a 20ft×20ft×10ft test frame mounted on a shake table at the Structural Engineering and Earthquake Simulation Laboratory at the University at Buffalo. The procedure developed for a coupled system using the transfer function matrix is verified analytically. The developed ‘open-loop’ compensation procedure is limited to linear systems and is affected by the uncertainty and nonlinearity of the systems. Challenges of shake table-structure control observed in this study are also addressed. FLOOR MOTION SIMULATION In the experimental evaluation of non-structural components or in the qualification testing of electric devices, it is often required to produce a target spectrum at a specified location in a structure mounted on the shake table. Floor or roof motions are amplified differently in the vertical and horizontal directions when a structure is subjected to ground excitations as shown in Figure 1. At the top corner of a jth story, horizontal motion is amplified and the amplification can be expressed as the ratio

The technology of real-time hybrid simulation (RTHS), which is one of the methods to test the performance of train operation on bridge in recent years, has strict real-time requirements. Combined with the fixed-point iteration (FPI) algorithm, this paper developed an off-line iterative hybrid simulation (OIHS) method, avoiding real-time synchronization between numerical and physical parts in RTHS to reproduce train–track–bridge coupling-vibration (TTBV) of high-speed railway in laboratory. In OIHS, the physical train model was installed on the shake table, while the response of the numerical track–bridge structure was simulated. To improve the test reproduction accuracy of the TTBV in high-speed railway, based on the shake table inherent off-line iterative control which is reproduce the commands issued to the shake table, an outer loop off-line iterative controller was designed to form a double-layer off-line iterative control framework to correct the response error between the subsystems. Compared with traditional finite element analysis, the accuracy and efficiency of the OIHS method were validated by numerical simulation and shake table test. The results demonstrated that the OIHS method can solve the dynamic convergence problem of the train–track–bridge on high-speed railway accurately and efficiently.

Facility Performance Indexes and Rapid Test Feasibility Evaluation Method of Shaking Tables

This paper is concerned with the verification of mathematical modeling of the container cranes under earthquake loadings with shake table test results. Comparison of the shake table tests with the theoretical studies has an important role in the estimation of the seismic behavior of the engineering structures. For this purpose, a new shake table and mathematical model were developed. Firstly, a new physical model is directly fixed on the shake table and the seismic response of the container crane model against the past earthquake ground motion was measured. Secondly, a four degrees-of-freedom mathematical model is developed to understand the dynamic behaviour of cranes under the seismic loadings. The results of the verification study indicate that the developed mathematical model reasonably represents the dynamic behaviour of the crane structure both in time and frequency domains. The mathematical model can be used in active-passive vibration control studies to decrease structural vibrations on container cranes.

SummaryThis paper presents an experimental investigation on the seismic response of medical equipment supported on wheels and/or casters. Two pieces of equipment were tested: a large ultrasound machine and a cart carrying smaller medical equipment. In the first phase, the resistance of the wheels and casters of the equipment was characterized through a controlled‐displacement procedure on the shake table. In the second phase, an extensive shake table test program was carried out to investigate the seismic response of the equipment. The input signals for the shake table tests included floor motions of a four‐story steel braced‐frame hospital designed to satisfy seismic requirements of a site in the Los Angeles area. The results of 96 shake table tests reported in this study include the seismic performance of the equipment under both unlocked and locked conditions, located on various floor levels of the building. It was observed that engaging the casters' locking mechanism does not necessarily decrease the relative displacement. The displacement response was sensitive to the excitation intensity and the orientation of the equipment with respect to the input excitation. Based on the experimental observations, appropriate structural engineering demand parameters associated with the relative displacement and relative velocity demands of the equipment are proposed and used to develop conditional probability curves. Finally, in an effort to extend the results of this experimental study to similar equipment on wheels/casters, the performance of a simple numerical model in predicting the peak seismic demands is evaluated. Copyright © 2016 John Wiley & Sons, Ltd.

Extensive damage to buildings and infrastructure observed during past earthquakes resulting from the liquefaction of shallow saturated soil deposits underneath structures has demonstrated the necessity for further research in the area of liquefaction-induced ground movement effects. This study explores utilizing helical piles as a countermeasure to reduce liquefaction-induced foundation settlement and investigates their seismic performance in liquefiable grounds. Two large-scale shake table test series, one without any mitigation measures and one using helical piles, were conducted using the shake table facility at the University of California, San Diego. During each test series, the soil and superstructure models were extensively instrumented and subjected to two consistently applied shaking sequences. The model ground included a shallow liquefiable layer aimed at replicating the subsurface ground conditions observed in the past earthquakes in New Zealand, Japan, and Turkey. Liquefaction-induced foundation settlement mechanisms are broadly categorized as follows: (1) shear-induced, (2) volumetric-induced, and (3) ejecta-induced. In the first test series (referred to as the Baseline test hereafter), all these three components were realistically reproduced, while in the second test series (referred to as the Helical Pile test hereafter) the volumetric and ejecta-induced mechanisms were mainly mitigated, resulting in significant reductions in the foundation settlement. Results from the first test series (i.e., Baseline test) indicated that the flow velocity due to the hydraulic transient gradient displayed an upward flow in the loose layer, which explains the observed sand ejecta. This series of shake table tests resulted in an average total foundation settlement of 28 cm and 42.7 cm during two shaking sequences. The measured foundation settlements were compared to the estimated foundation settlement obtained from Liu and Dobry [1997] and Bray and Macedo’s [2017] simplified procedures. The observed foundation settlements generally were higher than the estimated values. In the second large-scale test series, an identical test setup to the first test series was used except for a group of four helical piles were attached to the shallow foundation to mitigate liquefaction-induced settlements. In this series of tests, a reduced excess pore-water pressure generation around the group of helical piles was observed and is mainly attributed to the increased relative density around their zone of influence as a result of installation. The foundation supported on helical piles underwent almost no differential settlement and tilt. A significant reduction in the total foundation settlement was achieved during the Helical Pile test series compared to the Baseline test series. This shake table project is the first experimental study that reproduced all the key mechanisms mentioned above including the effects of sediment ejecta, which have not been captured in prior experimental studies. In addition, this series of large-scale shake table tests provides a unique benchmark for the calibration of numerical models and simplified procedures to reliably estimate liquefaction-induced building settlements. Although this study introduced helical piles as a reliable and highly efficient measure to mitigate liquefaction-induced foundation tilt and settlement, the proper design nd application of helical piles in seismic areas still need thorough investigation due to possible amplified superstructure response.

Implementation of real-time hybrid simulation using a large shake table with onboard actuators

Earthquake response of submerged gravity foundation on soft rock

This paper describes the results of the response analysis and the experimental tests on the newly developed earthquake isolation system based on the theory of friction pendulum. In Japan, many types of earthquake isolation device or system have been developed and utilized to proted computers, predsion equipments and important works of art from the severe seismic attacks. Although the performance of isolation of the each every one of the systems is sufficient, cost of the system is considered generally higher than the user expected and the system requires enough surrounding space to obtain adequate isolation performance. When applying the isolation device to the objects such as display cabinets in museum, computer server racks, bookshelves in library and vending machines which are generally placed adjacent to the wall, it is desirable to set them as closer to the behind wall as possible without decreasing their performance of isolation very much. The friction pendulum type isolation device with poly-curvature has been developed to satisfy these requirements, and the response analysis and the shake table tests have been carried out to investigate the performance of the system.

Given sufficient anchorage to the diaphragms, an unreinforced masonry (URM) wall subjected to out‐of‐plane inertial forces will likely develop a horizontal crack at an intermediate height about which the wall will rock as semirigid bodies. The effect of wall slenderness on out‐of‐plane stability has been demonstrated in past studies, but treatment of the effects of diaphragm flexibility and ground motion variability has been limited. This paper presents an experimental study examining the out‐of‐plane stability under seismic loading of URM walls connected to flexible diaphragms. Five full‐scale unreinforced solid clay brick wall specimens spanning one story were subjected to earthquake ground motions using a shake table. The top and bottom of the walls were independently connected to the shake table through coil springs, simulating the flexibility of diaphragms. Variables examined experimentally included diaphragm stiffness and wall height. Both the amount of rocking observed as well as the ground motion scale causing collapse varied significantly with changes in the diaphragm properties. The test results provided data used for validation of a rigid‐body rocking model, enabling an extensive parametric study on wall stability and the development of new assessment guidelines in a companion paper.

Seismic response investigation of 1/20 scale container crane through shake table test and finite element analysis

The ALSC base isolation system was investigated by testing a church model to the scale 1/3.5 on the two-component shaking table at the IZIIS’ Laboratory, Skopje, Macedonia. The investigation was performed within the frames of the European FP6 project PROHITECH, related to seismic protection of historical buildings by innovative and reversible mixed technology systems. This system was originally developed in 2001 and applied on a model of a steel reservoir. The church model was tested twice: with an active base-isolation system and fixed to the base. The effectiveness of the base-isolation system was investigated by the transfer function/floor response spectra technique. Special software for calculation of the transfer functions and floor response spectra based on experimental data, referred as “SPECTRA”, was developed by the authors of this paper. The main objective of the investigations presented in this paper was evaluation of the effectiveness of the ALSC floating sliding base-isolation system by using the Response Spectra approach. Comparison of the Floor Response Spectra was done for two cases a fixed base model and a base-isolated model. The base response spectrum was calculated from the input time history of the Montenegro earthquake compressed 3.5 times (geometry scale of the model). The Transfer Functions and the Floor response Spectra at a particular level were calculated from the Cross Power Spectra (CPS) and the Auto-Power Spectra. The obtained results showed that the effect of the base-isolation system was considerable, with a low transmissibility of energy from shake table to the model.