development of the shaking table and array system technology in china
Chun-hua Gao, Xiao-bo Yuan, "Development of the Shaking Table and Array System Technology in China", Advances in Civil Engineering, vol. 2019, Article ID 8167684, 10 pages, 2019. https://doi.org/10.1155/2019/8167684
Shaking table is important experimental equipment to carry out antiseismic research. Research, conclusion, comparison, and analysis concerning the developmental history, constructional situation, performance index, control algorithm, and experimental technique of the internal shaking table were reviewed and compared. Such functional parameters as internal shaking tables table-board size, bearing capacity, working frequency, and maximum acceleration were given. Shaking tables constructional status quo and developmental trend were concluded. The advantages and disadvantages of different control algorithms were contrastively analyzed. Typical shaking table test, array system tests, and experimental simulation materials were induced and contrasted. Internal existing shaking table and array system tests structural type, reduced scale, and model-material selection were provided. Analysis and exposition about the developmental tendency of shaking tables enlargement, multiple shaking tables array, full digitalization, and network control were made. The developmental direction, comparison of technical features, and relevant research status quo of shaking table with high-performance were offered. The result can be reference for domestic or overseas shaking tables design and type selection, control technique, and research on experimental technique.
At present, the structural seismic research methods include the pseudostatic test, pseudodynamic test, and shaking table test. The test method of the shaking table test can recreate the structural response and seismic oscillation in the lab accurately and reproduce the whole process of seismic oscillation effect or artificial effect in real time. The development of shaking table provides an accurate and effective way to study structural elastic-plastic seismic response .
Japan and the United States are the first two countries to establish shaking tables in the world. And, China initially built a shaking table in 1960  when Institute of Engineering Mechanics, Chinese Academy of Sciences, built one-way horizontal vibration  with a specimen size of 12m3.3m. So far in China, there are a lot of shaking tables ; some were made in China, some were systematically remodeled from imported parts, and some were totally imported. In recent years, many scholars [8, 9] and Wang et al.  conducted abundant research on the development and control technology of Chinas shaking tables and also got some research achievements. However, such results are mostly summaries of the test technologies or control technologies of shaking table , while there are few summaries concerning the construction history and usage of domestic shaking tables. This paper makes a comprehensive summary of the development and application of domestic shaking tables and array test technologies in terms of the development, control technology, test application, and development trend of shaking table and array system based on current collected information, so as to provide some reference and basis for the construction and development of domestic shaking table.
The development of shaking table in China came relatively late [1, 3, 1114]. It can be roughly divided into four stages. In 1960s, the mechanical shaking table was the main stream with a working frequency of 1Hz40Hz, of which the characteristics of the specimens in low segment are difficult to be controlled [2, 10, 11]. Electrohydraulic shaking table was then rapidly developed with its high frequency. In 1966, departments of machinery and electronics collaborated with each other to build Chinas first exclusive shaking table for national system of defense in three years [2, 10, 13]. Thereafter, many domestic colleges and universities as well as scientific research institutes also begun to conduct researches. For example, Tongji University brought in the 4m4m two-horizontal dimensional identically dynamic electrohydraulic shaking table developed by American MTS, which has been transformed into three- to six-degree-of-freedom identically dynamic shaking table . At the beginning of the 70s, the research on shaking table in China was continuously carried out and quickly developed. Our country also started to develop one-way electrohydraulic servo shaking table but rarely hooked into multiaxis shaking table . And, foreign shaking tables were introduced only when the test was demanding, so the introduction quantity of shaking tables was sharply decreased. Domestic institutes that conduct researches on shaking table mainly include China Academy of Building Research, Xian Jiaotong University, HIT (Harbin Institute of Technology), Institute of Engineering Mechanics, and Tianshui Hongshan Testing Machine Co., Ltd. [21, 22]. The shaking table construction situation in China is shown in Table 1.
The work frequency of electrohydraulic shaking table in the early stage of our country was about 50Hz. At present, at home and abroad, the work frequency of high-thrust shaking table with over 50t can reach more than 1000Hz. For instance, the work frequency of Y2T.10c shaking table developed by 303 Research Institute of China Aviation Industry Corporation is as high as 1000Hz, and the wide band random vibration control precision is 2.0dB  within the frequency range of 20Hz1000Hz.
In 2006, Beijing University of Technology built a nine-sub-building block array system with a size of 1m1m, which along with the original 3m3m single-array system composed the 10-subarray system, which can be used to constitute testing systems with any several subarray systems and many optional positions; at the end of 2006, Institute of Electro-Hydraulic Servo Simulation and Test System of Harbin Institute of Technology (HIT) developed successfully the first domestic multiaxis independent intellectual property rights (the hydraulic vibration test system with shaking table system is shown in Figure 1) and got identification, which changed the history of depending on importing shaking tables . In 2012, Jiangsu Suzhou Dongling Vibration Test Instrument Co., Ltd. successfully developed the worlds largest single electromagnetic shaking table test system (http://www.cnki.net/kcms/detail/11.2068.TU.20130124.1608.001.html) with a thrust of 50 tons.
With the 9-subarray system of Beijing University of Technology as an example, this paper introduces the construction situation of shaking table array system. In 2003, The State University of New York built the first set of two-subarray systems. In the same year, the University of Nevada-Reno built the three-subarray system with three movable two-direction shaking tables. The size of the table and the maximum bearing capacity of the shaking table are introduced. The array system (shown in Figure 3) is suitable for experimental research on spindly space structure.
In 2004, Chongqing Jiaotong Institute of our country completed the constitution of the two-subarray system with a specimen size of 6m3m, of which one is fixed and the other is movable (shown in Figure 4). And, in 2008, National Key Laboratory of Bridge Dynamics was established.
In 2011, Beijing University of Technology began to prepare to construct nine-subarray system (shown in Figure 5) and has built 12 sets of actuator building block array systems till 2006, which was increased to 16 sets in 2009 and is now the array system with the largest number of single-array system in the world. Each single shaking table is composed by mesa, 5 connecting rods, a vibrator, and a base. The array system can be made into various combinations by 16 sets of vibrators and connecting rods to conduct varied shaking table array tests with different layouts and forms. The performance indicators of nine-subarray system are shown in Table 2. The system uses four piston pumps to offer oil. The rated oil supply pressure of the seismic simulated shaking table system is the same as the maximum oil supply pressure. In addition to 4 oil pumps, the system also has energy storage to supplement the oil supply when the oil supply of the oil supply pump is insufficient.
There are two main types of traditional shaking table control technology: one is PID control based on displacement control and the other is three-parameter feedback control (also known as the three-state feedback control) synthesized by the displacement, velocity, and acceleration . It is essential for feedback theory to adjust the system after making the right measurement and comparison. In 1950, the PID control method mainly composed of unit P proportion, integral unit I, and differential unit D was developed. The traditional PID control method is simple in control algorithm, good in stability, and high in reliability and thus has been widely applied in the practical engineering. The PID control method is especially suitable for deterministic control system. Yet, as the target signal of shaking table is acceleration signal, high-frequency control performance is poorer when the displacement PID control is adopted, while the mesa cannot be located if acceleration PID is used. Meanwhile, in the process of control, nonlinear behavior exists in every specimen; thus, the effect of traditional PID control is not ideal due to the large waveform distortion [24, 2629]. As the structure sets higher requirement for control accuracy, three-parameter feedback control synthesized by the displacement, velocity, and acceleration was put forward in 1970s (the control principle is shown in Figure 6), which makes up for the narrow frequency band and the inability to realize acceleration control of single displacement control. Acceleration feedback can improve the system damping, and velocity feedback can improve the oil column resonance frequency. Adopting the displacement to control low frequency, speed to control midfrequency, and acceleration to control high frequency plays an important role in improving the dynamic behavior and bandwidth of the system. The introduction of three-parameter control technology greatly improved the playback accuracy of seismic time history, but due to the complexity of transfer function in the system, the correlation of input and output waveform is still not high. Power spectrum emersion control algorithm modifies drive spectrum utilizing system impedance and the deviation of the reference spectrum and the control spectrum, so as to get a relative high consistency of response spectrum and reference spectrum of the system [30, 31]. Power spectrum retrieval principle diagram is shown in Figure 7. This method belongs to the nonparametric method, which has nothing to do with any model parameters. But the matching degree of estimated power spectral density and real power spectral density is very low, so it is an estimation method with low resolution.
Another kind of the parametric estimation method, using the parameterized model, can give a much higher frequency resolution than period gram methods. The power spectrum control method based on the parameter model has high resolution and can improve the system control convergence speed and power spectrum estimation precision, yet it is sensitive to noise with higher computation requirements. Therefore, in the vibration test control, it has not reached practical stage .
The traditional control algorithm is based on the linear model of vibration table and specimen , and the parameters in the process of test are assumed unchanged, but the actual test object is very complex. The components experience elastic-plastic phase and then the failure stage in the process of the test, and the parameters that were assumed to be unchanged turn out to have been changed in the process of test. The change of the parameters influences the accuracy of the input seismic signal, which is the biggest defect in the traditional control technology. From the 1970s to 80s, intelligent control is a new theory and technology with strong control ability and great fault tolerance. The introduction of the adaptive control improved the robustness and control precision of the system, such as adaptive harmonic control theory (AHC), adaptive inverse function control theory (AIC), and the minimum control algorithm (MCS) . At present, the fuzzy control algorithm of the structure control attracted the attention of more and more scholars with its advantages of powerful knowledge expression ability, simple operational method, and the adoption of fuzzy language to describe the dynamic characteristics of the system. As early as 1996, some scholars abroad has carried out the induction and comparison of structural seismic control methods and summarized the advantages and disadvantages of various control methods, particularly expounding that the fuzzy control and neural network control algorithm could better solve the problem of nonlinear. The application of domestic intelligent control algorithm in the engineering structure control is relatively late. In 2000, Ou  and other scholars proposed the control algorithm which can realize fuzzy control according to the control rules and fuzzy subset, which greatly improved the practicability and efficiency of fuzzy control algorithm.
Most of the fuzzy control rules are established based on experience, leading to great difficulty in structure control. In view of this, Wang and Ou , in 2001, put forward the method of extraction, optimization, and generation of fuzzy control rules with the basis of structural vibration fuzzy modeling and genetic algorithm. Qu and Qiu  came up with a kind of active feed forward control method based on adaptive fuzzy logic system method, which better solved the nonlinear control problems of reference signal and external interference in the feedforward control. Wang  for flexible structure completed the application of the fuzzy PID control method in the structural vibration and conducted the active control experimental verification of beam vibration.
The efficiency of fuzzy control depends on the selection of function parameters and the establishment of the fuzzy control rules. Therefore, the adaptive fuzzy control is of great research significance for the nonlinear structure system. Because of the functions of self-adaptation and self-study of artificial neural network, the application of neural network in seismic control in civil engineering began in the 60s, which adopts a simple neural network controller to control the movement of the inverted pendulum, and achieved good effect. In 2003, Mo and Sun  implemented numerical simulation of active vibration control on the beam vibration control model by using genetic algorithm with the minimum energy storage structure as the goal, compared with the exhaustive method, and achieved good control effect. Chen and Gu  carried out simulation research on the application of frequency adaptive control algorithm based on the least square method in the domain of vibration control, and the simulation got the damping effect of about 50db. Li and Mao  achieved evolutionary adaptive filtering algorithm with strong instantaneity and applied it into the vibration control of structures to conduct simulation calculation based on genetic algorithm and moving least mean square algorithm of transient step, and the simulation obtained the damping effect of about 30db.
To solve the limit bearing capacity of shaking table for large structure test, scholars from all over the world conducted a wide variety of researches. The combination of substructure technique and shaking table test is an effective way to solve this problem . Hybrid vibration test divides the structure into test substructure and numerical substructure. Test substructure is the complex part in experiment on shaking table, while numerical substructure is the simple part to carry out numerically simulation. Test substructure can carry out full-scale or large-scale model test, avoiding the influence of the size limit of shaking table with large-scale structure, and thus was widely used in the study of the engineering seismic test. The domestic researchers Chen and Bai  implemented preliminary exploration into structural seismic hybrid test technique on account of the condensation technology. In 2008, Chen and Bai  also embarked on the hybrid vibration test on the hybrid structural system of commercial and residential buildings, of which the bottom commercial district was put into a full-scale experiment on shaking table and other parts were involved in numerical simulation.
In 2007, Mr. Wu Bin from Harbin Institute of Technology applied the center difference method into the change of the acceleration calculation formula in hybrid real-time test which takes consideration of the quality of test substructure and analyzed the stability of the algorithm. The test results show that the stability of the center difference method in real-time substructure application is poorer than that of the standardized center difference method. Such scholars as Yang  in the same year made the numerical simulation analysis on the shaking test substructure test, and the analysis results show that the integral step change is sensitive to the influence of experimental stability. At the same time, he verified the validity of the theoretical research results.
In recent years, the structural styles of shaking table test research were developed from masonry structure, frame structure, tube structure to bridge structure, structures with the consideration of some isolation and damping measures, and structural foundation interaction experiment. The application of shaking table tests on the structure seismic resistance made it possible to establish structure nonlinear model with various structural styles . Many shaking table tests have been carried out in recent years in China, which, according to the testing purpose, can be roughly divided into three categories: the first type is to determine structural earthquake-resistance performance as the test purpose; the second type is to determine the dynamic characteristics of structure, obtain such dynamic parameters as the natural vibration period and damping of structure, seek for weak parts of the structure damage, and provide the basis for super high-rise and supergage designs; the third type is to verify the applicability of certain measure or design theory in the structure. This paper drew a conclusion of typical shaking table tests in recent years in terms of building types, model dimensions, and so on (shown in Table 3).
Shaking table experiment diversifies the structural styles in experiment, makes it possible to establish the nonlinear damage model, and provides a reliable basis for all kinds of structures to establish the corresponding destruction specification. But large span structure tests on bridges, pipes, aqueduct, transmission lines, and so on may produce traveling wave effect under the action of earthquake due to large span, and a single shaking table will not be able to simulate the real response of the whole structure under seismic action. Array system can better solve these problems. For example, the State University of New York-Buffalo did damper damping effect research on Greek Antiliweng Bridge using 2-subarray system; conducted shaking table array test research on two continuous steel plate girder bridge and concrete girder bridge by using the 3-subarray system of University of Nevada. Many domestic scholars also carried out shaking table array test research on different structures of array systems. For instance, in 2008, Gao Wenjun made shaking table array test research of organic glass model on Chongqing Chaotianmen Bridge with the 2-subarray system of Chongqing Traffic Academy; conducted a multipoint shaking table array test research on concrete-filled steel tubes arch bridge with the 9-subarray system in Beijing University of Technology.
According to the size of mesa, shaking tables can be divided into large, medium, and small ones; in general, specimen size less than 2m2m for the small, 6m6m for the medium, and over 10m10m for the large. Due to the size limitation of a small seismic simulation vibration table, it can only do small-scale tests, and there is a certain gap with the prototype test. In the seismic simulation vibration table test of scale model, all parameters are required to meet the similarity principle, but it is difficult to do in practical engineering. For some important structures, especially the important parts of large structures, to accurately reflect the dynamic characteristics of the structure, within the permitted scope of the condition of capital, it is necessary to increase the specimen size and the maximum load as much as possible to eliminate the size effect of the model, so the large full-scale test must be the development trend of shaking table. China Academy of Building Research developed a shaking table with a mesa dimension of 6.1m6.1m and the maximum model load of 80t.
Due to the great investment, high maintenance cost, and test fees as well as long production cycle of large-scale shaking table, infinite increase in size of shaking table is obviously unreasonable, and likewise, it is not possible to fully meet the actual requirements only by increasing the size of shaking table. For large-span structure tests on bridges, pipes, aqueduct, transmission line, and so on array systems composed of many sets of small shaking table can be adopted. Shaking table array can either conduct a single test or make seismic resistance test on the structure of large-scale, multidimensional, multipoint ground motion input with varied combinations according to various needs. Therefore, the array system composed of many sets of small shaking tables must be the development trend of shaking table.
In terms of control mode, power spectral density control was mostly adopted before 1975. After 1975, Huang Haohua and other scholars used the time-history playback control to finish the seismic wave control research in a broad band. In the mid-1990s, digital control and analog control are widely used in the shaking table control, of which digital control is mainly applied in the system signal and compensation and the analog control is the basis for the control, whose control mode is complicated in operation with too much manual adjustment. After 1990s, Fang Zhong and other scholars developed a full digital control technology which has been widely used in the hydraulic servo control system with the rapid development of digital technology. Other than the valve control device and feedback sensor which adopt analog circuits, the rest utilize digital software to fulfill implementation. This control method can make up for some flaws in the analog control with simple test operation, being able to improve the accuracy, reliability, and stability of the system. Full digital control is the inevitable development trend of hydraulic servo system control.
With the appearance of slender and shaped structures and the application of new materials in building engineering, the seismic test methods of structures are put forward with higher and higher requirements. To meet the requirements of actual engineering and seismic research, scholars from all over the world are active in exploration and attempt and put forward some new testing methods. In recent years, countries around the world greatly invest in seismic research. From 2000 to 2004, the United States Science Foundation Committee spent eighty million dollars of research funding on the NEES plan; Europe established a collaborative research system European Network to Reduce Earthquake Risk (ENSRM); South Korea established a virtual structure laboratory using grid technology, which includes the wind tunnel, the shaking table, and other scientific research equipment. Furthermore, Internet ISEE Earthquake Engineering Simulation System in Taiwan of China was the earthquake engineering research platform developed by National Earthquake Engineering Research Center of Taiwan, China, with the Internet. The platform not only allows several laboratories to interconnect each other to implement large-scale shaking table test but also permits different laboratory researchers around the world to observe the test simultaneously and synchronously.
In China, Hunan University firstly put forward the structure network synergy test research and cooperated with Vision Technology Co., Ltd in 2000 to develop the network structure laboratory (NetSLab is shown in Figure 8). The main module and interface are shown in Figure 8. Thereafter, Hunan University cooperated with Harbin Institute of Technology to accomplish secondary development to establish the network collaborative hybrid test system and conducted a structure remote collaborative test along with Tsinghua University, Harbin Institute of Technology. Three domestic universities firstly completed remote collaboration pseudodynamic test, which is shown in Figure 9.
This paper drew a conclusion of the construction, history, and status quo as well as application and research of shaking table and array. The main conclusions are as follows:(i)On account of factors of actual application demand and economy, the size of the shaking table is between 1m and Xm, among which 3m6m are the majority. For large span structures such as bridges and pipes many sets of small array mode of vibration table can be used.(ii)Shaking table mesa acceleration and speed are about and 80cm/s, respectively. Through statistics, the remarkable frequency of previous ground motion records is mainly within 0.1Hz30Hz, and the frequency range of medium shaking table should be in 0Hz50Hz according to the requirements of the similar rule. Moreover, tests with special requirements need to be above 100Hz.(iii)With the appearance of slender and shaped structures and the application of new materials in building engineering, the seismic test methods of structures are put forward with higher and higher requirements.
The authors acknowledge the support from the Science and Technology Breakthrough Project of the Science and Technology Department of Henan Province ( 9), the Key Scientific Research Projects in He Nan Province (No. 18B560009), and Nanhu Scholars Program for Young Scholars of XYNU in China. The authors thank Xin Yang Normal University School of Architecture and Civil Engineering Laboratory and would also like to thank teachers and students of the team for collecting data.
Copyright 2019 Chun-hua Gao and Xiao-bo Yuan. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
shaking table test - an overview | sciencedirect topics
Liquefaction-induced lateral spreads have been extensively investigated using shaking tables (e.g., Towhata et al., 1996). In these types of experiments, a reduced-scale model of soil deposits is subjected to short pulses or a continuous time history of acceleration simulating the earthquake ground motion. Figure 20 shows typical results of ground deformation obtained in shaking table tests. The acceleration, pore pressure, and displacement are measured at various locations in the model, which is 2 m long and 50 cm high. In the impact test [Figure 20(a)], the model is subjected to a very short acceleration pulse. The pore pressure rises very quickly, and the deformation takes place over an extended period of time after the impulse, in disagreement with centrifuge observations. In the shaking test [Figure 20(b)], the model is subjected to a sine-like base acceleration. The pore pressure gradually builds up until the soil liquefies. The ground deformations are progressive and stop with the base acceleration, as observed in centrifuge experiments.
In shaking table tests, Kokusho et al. (1998) reported the formation of a water film beneath a thin layer of silt sandwiched between sand layers. When such a film appeared, the soil mass above the silt layer was observed to glide in the downward direction not only during but also after the shaking. When the film did not form, the lateral flow took place mainly during the shaking. This shaking table experiment indicates the influence of drainage conditions on liquefaction-induced deformations.
Based on the shaking table test of the underground two-hole loess cave dwelling, the numerical model of three-dimensional ULC was established by using ABAQUS software. The numerical simulation of the numerical model was carried out, and the comparative analyzed between the numerical calculation results and the shaking table test results. The main conclusions are as follows:
The Hardin-Drnevich dynamic constitutive model was used to reflect the nonlinear characteristics of loess material, and the relevant parameters of the loess material of this model were determined according to geotechnical tests such as dynamic triaxial.
The IE artificial boundary established by the ABAQUS numerical analysis platform in this study was accuracy used to seismic analysis in underground loess structures. The equivalent node load on the artificial boundary was innovatively simplified to the equivalent surface load of the divided unit. The rationality and accuracy were verified, and this research provided a basis for the artificial boundary processing of the numerical model for the dynamic analysis of three-dimensional underground structures.
The ULC model was established by the IE-FE coupling method, and the seismic response of underground structures was analyzed. The results of the analysis showed that the modal of the numerical simulation was consistent with the test results, and the simulation of the first order and second order of natural vibration frequency was close to the experimental results.
According to the stress analysis, under the action of the earthquake, the large main tensile stress of the arch apex and its upper covering soil of the ULC structure appeared. The tensile stress level increased correspondingly with the increase of the PGA, which was consistent with the failure phenomenon by shaking table tests. The seismic acceleration and displacement of the numerical simulation in the ULC under earthquake action were agreed well with the experimental results. The establishment and analysis of this numerical model can be used to analyze the seismic performance of underground structures during earthquake action.
The size of the structure had a great influence on the seismic performance of the loess cave. Through parameter analysis, it could be obtained that: appropriately increasing the thickness of the covering loess was beneficial to improve the bearing capacity of the loess cave. When the thickness of the covering loess was greater than the height of the loess cave, it was beneficial to the stress of the loess cave and could significantly improve its seismic resistance; reasonable rise-span ratio could keep the loess body of loess cave under compression and effectively decrease the seismic response of the cave leg; with the increase of the width of the middle leg, the dynamic response of the cave leg portion decreased, which was beneficial to improve the seismic resistance of the loess cave structure.
Chinese government provided a shaking table test standard of architectural CW in 2001 . A testing frame is necessary to be designed for installation of the CW specimens, which should have the ability of producing the expected story drift. However, there are no provisions on how to design the frame. Designing a suitable testing frame is a challenging task prior to the test. The CW specimens should be tested at full scale and at least have two vertical story levels and three horizontal glass panels. Typical horizontal and vertical gaps between the neighboring panels and between the adjoining frames should be included. In order to determine the dynamic properties of the testing system, e.g., the vibration frequency, vibration mode, and damping ratio, white noise with peak acceleration of 0.070.1g should be input to the testing system before actual tests. The peak acceleration of the input motions should start from 0.07g and increase to 0.5 times the design peak ground motion (PGA) in the following testing scenarios. However, the required input motions, which are also known as the core part of the shaking table test, are not yet addressed. In the newest draft standard for reviewing , some natural ground motions are recommended if there are no proper motions available. Artificial motions are also acceptable and their selection method should follow the current seismic design code . The code specifications are helpful in shaking table testing of CW systems, but the issues of test frame design, and the selection and/or generation of the input motions are still critical problems to be solved.
Finally, it is worth to mention that the former two shaking table testing protocols (AC156  and FEMA 461) are very rarely used to investigate the seismic performance of the CW. In China, however, GB/T 18575 was used very often.
Igarashi et al.  performed a full-scale substructured RTHSTT of an idealised two DOF structural system with a tuned mass damper (TMD) providing structural control. The TMD formed the PS whilst the structural system formed the NS. Results showed that the control method using the TMD is feasible as long as the stability conditions of the test specimen and test parameters are satisfied. A similar substructured RTHT was performed by  of an active mass damper (AMD).
Carrion and Spencer  described a full-scale substructured RTHT of an MR damper for semi-active control of a three-storey steel framed structure. The MR damper formed the PS whilst the remainder of the structure formed the NS (see Fig. 17). Model-based feedforward compensation accounted for the variations on the actuator dynamics. The RTHT demonstrated the successful performance of the structural control algorithm. A number of other authors have also performed substructured RTHTs to investigate semi-active control of building structures using MR dampers, such as [104,185,186].
Lee et al.  evaluated the vibration control effect of a scaled tuned liquid damper (TLD) for a building structure using the RTHSTT method. The TLD formed the PS and a numerical structural model of a single- and three-storey steel frame formed the NSs. Feedback from the shear force signal measured by a shear type load cell located between the shake table and TLD was used in the control loop as an interaction force between the TLD and NS as shown in Fig. 18. Comparison between the RHSTT method and a conventional shake table test showed good agreement. The test results showed that the TLD could effectively mitigate the seismic response of the structure investigated.
Stavridis and Shing  performed a series of RTHTs of a 1/3 scale three-storey suspended zipper steel frame. The bottom-storey formed the PS whilst the remainder of the structure formed the NS. The zipper struts were designed to transfer unbalanced forces up to the story above when the V-bracing in the storey below buckles. Results showed the top storey remained elastic and prevented collapse as designed.
Karavasilis et al.  evaluated the seismic performance of a full-scale two-storey four-bay steel moment resisting frame (MRF) structure with compressed elastomer dampers using a substructured RTHT. The tests were conducted to verify the performance-based seismic design of the structure. The experimental substructure consisted of two individual compressed elastomer dampers and the MRF formed the NS. Results showed that the steel MRF with elastomer dampers performed better than conventional special MRFs.
Experimental studies have shown non-uniformity of the soil gas and the unevenness of the penetration of the slurry material during the liquefaction mitigation [20,109]. Centrifuge and shaking table tests results reveal that the distribution of air bubbles is not completely uniform, and the higher the air injection pressure, the wider and more uniform the effective air-entrapped zone [103,122]. Sparging of air under high pressure might disturb the granular structure of sand. Thus, some measures can be taken to avoid high pressure sparging or alleviate structure disturbance .
However, the uniform distribution of gas cannot be efficiently examined, and the successful soil improvement by gas can not be confirmed. In some cases, the extent of the improvement is not reflected for in-situ test results until a period of time after the improvement has been completed. For the air injection method by sparging, a higher flow rate improves the air distribution. In other words, to be effective and uniformly distributed, air must be continuously or frequently introduced. As currently envisioned, the air injection method would be feasible for structures with limited access since air is injected, rather than grout [103,118].
On the other hand, a series of experiments studied the effects of groundwater flow on the flow patterns of injected air. The air flow patterns observed are found to depend significantly on the soil type and groundwater flow conditions . The shape of the injected air zone of influence is not affected by groundwater flow when the hydraulic gradient is less than or equal to 0.011 . Recent laboratory and pilot-scale research have shown that the effectiveness of air sparging is often limited by a number of factors in practice. One major constraint is the impact of channeling on air movement during sparging . Air injected into saturated porous soils frequently moves in discrete channels that comprise only a fraction of the entire cross-section of the zone, rather than passing through the whole medium as bubbles. This channeling phenomenon greatly reduces the stripping efficiency of air sparging. The physics of air bubbles movement under water are not widely understood and the movement is extremely sensitive to formation structure .
Although many in situ tests results reveal that the degree of saturation of soils after several years is noticeably, not significantly, higher as compared with that shortly after ground desaturation, some have survived for decades, and the longevity of air bubbles injected into the improved soil and evaluation of effectively desaturated zone is confirmed through field tests [100,129]. Anyway, the ground water flow may remove with time the injected gas from the target subsoil. There is no appropriate technology to capture the time-dependent loss of reduced saturation.
Study performed by Xue and Shinozuka  investigated damping, dynamic, and seismic behaviors of rubberized concrete for its potential application as structural material. Small scale rubberized concrete columns were fabricated and tested under free vibration to identify damping ratios, and on seismic shaking table tests to investigate the structural responses to earthquake ground motion. Increase of damping ratio and reduction of seismic response of concrete structures was reported. Moustafa et al.  performed first shake-table tests on two large scale cantilever reinforced concrete columns. One column was casted with conventional concrete and second with rubberized concrete in which 20% of fine aggregate volume was replaced with crumb rubber. Columns were subjected to a sequence of ground motions scaled to the certain design spectrum. Results are compared for both types of columns and presented in Table 6. It was reported that in rubberized concrete column lateral drift capacity and dissipated energy were increased.
Hassanli et al. [66,67] made four reinforced columns out of rubberized concrete mixtures with crumb rubber which were previously treated with NaOH solution which were tested under eccentrically monotonic axial load. This study reported that compressive strain capacity, viscous damping ratio and kinetic energy increased with increased rubber content, however adverse impact on the dissipated hysteretic energy was observed.
Following the extensive research efforts on connections and wall systems, full-scale tests were also performed on single and multi-storey CLT structures. The investigations have been firstly carried out on a one-, three- and seven-storey buildings as part of the SOFIE Project ; all structures were erected using narrow CLT panels, hold-downs and angle brackets similar to those adopted in lightweight structures, and vertical step joints with partially threaded screws.
Lauriola and Sandhaas  conducted pseudo-dynamic lateral tests on the one-storey building, while Ceccotti and Follesa  performed full-scale shaking table tests on the three-storey structure. Tests investigated the lateral deformability capacity caused by the presence of large openings at the ground floor. In particular, the experiments considered two symmetrical configurations parallel to the loading direction (with different layouts of openings) and a third one asymmetrical. Neither of the structures underwent any major damage in the CLT members; furthermore, because the deformability capacity was governed by the rocking of the single walls, tests exhibited lateral deflections proportional to the area of the openings.
Ceccotti et al.  carried out the shaking table tests on the seven-storey building, assembled using CLT elements and connection systems similar to those used in the one- and three-storey structures. However, because the hold-downs used for the three-storey building were not suitable for high slender structures, special high-strength hold-downs with 10mm thick plates were placed at the anchoring to the foundations. After several tests, carried out by varying the ground motion record, the building did not collapse and exhibited only local damages close to the connections location.
Flatscher and Schickhofer  conducted full-scale shaking table tests on a three-storey building, as part of the SERIES Project. Compared to the three-storey structure tested by Ceccotti and Follesa , the building was assembled using large monolithic walls rather than narrow panels with vertical step joints. Furthermore, fully threaded screws were primarily used as fasteners, rather than partially threaded screws. Consequently, the lateral deformability of the building was lower than the structures tested within the SOFIE Project, thus resulting in smaller inter-storey drifts.
Popovski and Gavric  tested a two-storey structure under quasi-static loading conditions, paying particular attention to the lateral strength and deformability capacities. Tests did not exhibit any global instability even after the attainment of the maximum load-carrying capacity; furthermore, only minor torsional effects were observed and the ultimate resistance was identical in both principal directions.
Finally, in Japan, several shaking table tests have been carried out on multi-storey CLT structures. Tsuchimoto et al.  focused on the static lateral capacity and seismic performance of a three-storey structure assembled with narrow CLT panels. Compared to the structures previously tested in Europe and Canada, the building was assembled using tension bolts and screwed steel-to-timber connections rather than with hold-downs and angle brackets. Kawai et al.  investigated the dynamic behaviour of a five-storey structure assembled using similar connections to those used by Tsuchimoto et al. . Furthermore, Kawai et al.  extended the analyses to a three-storey structure where the CLT panels were used as outside walls and solid timber frames were used in the inside. Finally, Yasumura et al.  tested two two-storey structures composed of monolithic and narrow CLT panels, respectively. In the structure with monolithic wall panels, some cracks were observed at the corners of the openings, which propagated both vertically along the grain of the panel surface and diagonally as the horizontal displacement increased. In the structure assembled with narrow wall panels, some gaps were observed between each wall and no cracks were visible at the corner of the openings; additionally, small gaps were observed at the floor joints above the openings and bending failure of the CLT floor panel was detected above the corners of the openings.
Moustafa et al.  studied the seismic behaviour of two large-scale reinforced concrete columns (NC and NRC with 20% rubber) using the shake-table tests. Although both columns passed the drift limit states, defined by Dutta and Mander , NC column did not pass the state of moderate damage whereas, NRC column passed it with spalling at 2.9% drift and 110% design earthquake. In addition, NC column did not pass the state of extensive damage, whereas NRC column passed the state of extensive damage with 5.4% drift at 190% design earthquake. Under 100% design earthquake, NRC column displayed 9% and 5% smaller drift and acceleration, respectively, compared with NC column. Also, the cumulative dissipated energy increased by 16.5% for NRC column compared with NC counterpart. Hence, NRC column outperformed NC column because of its high energy dissipation, high damping ratio, low residual drift and the delay in the rebar fracture. Xue and Shinozuka  investigated the dynamic behaviour of NRC small-scale columns (40mm square cross section and 500mm height) with 15% rubber using free vibration and shaking table test. In general, the NRC columns were exhibited fewer cracks compared to NC columns due to the higher ductility of the rubberised concrete. The acceleration decreased by 27% for the NRC columns compared to NC columns whereas the average damping ratio increased by 62%. The increase in the damping ratio is attributed to the high energy dissipation properties of NRC. In addition, the average natural frequency decreased by 28% compared to the NC columns.
As discussed above, in most cases and for different structural elements including columns, beams, beam-column joints, and walls, the incorporation of rubber particles as a partial replacement of sand or stone increases the energy dissipation capability, damping ratio, and ductility and reduces the brittleness index, which improves the cracking formation characteristics. Table 5 and Fig. 10 show the change in concrete material compressive strength, ultimate load carrying capacity, energy absorption, ductility, and damping for structural elements under cyclic and seismic loads. The data used in the table and figure are found in Refs. [22,24,64,65,7073,75,77,78] (only for crumb rubber replacing sand by volume). The percentage of these changes are not constant among different research as they depend on many factors such as the structural element being tested, rubber type and size, type of replacement, the control mixture components, the target strength, and the load patterns.
In general, these results highlight the potential improvements of the behaviour of structural elements under cyclic and seismic loads when rubber particles were employed in the concrete mixture. As shown in the figure and the table, Damping, ductility and energy absorption increase with the increase in rubber content. Also, the reduction in the peak loads was significantly smaller than the reduction in the concrete material strength when using rubberised instead of traditional concrete as shown in Table 5 and Fig. 10. As a summary, rubberised concrete has potential advantages in structural applications of columns, beams, walls, and beam-column joints in high seismic zones through providing good damping, ductility, and energy absorption properties.
This paper presents a literature review on wall-to-horizontal diaphragm connections in URM buildings, considering the typical connections found in traditional buildings and the main solutions to enhance the seismic behavior of such connections. Both experimental and numerical research studies aimed at mechanically characterizing these connections are reported and discussed. This study reveals the necessity of accounting for wall-to-floor or wall-to-roof connections in the seismic analysis of existing URM buildings and stresses the need of simple but clear guidelines for their assessment and recommendations on their numerical modelling.
Considering the peculiarity of the connections, the calibration of numerical models based on experimental data is necessary for a reliable evaluation of the seismic capacity of historical constructions. Despite pull-out monotonic and cyclic tests on detailed connections for the axial and shear capacity are appropriate for this scope, they are costly and complex to set-up. Failure mechanisms and ultimate dynamic capacity can be assessed through shaking table tests on building-scale and wall-scale specimens to understand the influence of the strengthening strategy on the original poor building.
Complex and detailed finite element models reveal to accurately simulate the global behavior of the URM building but they are hardly used from professionals who look for simpler procedure (equivalent frame or macro models). Uncertainties can be found on the definition of the mechanical characteristics of masonry elements and diaphragm are aether not-considered as structural elements or they are modelled as rigidly connected to the perimeter walls. From the literature review, it is clear how the structural nonlinearities are important, especially when dealing with connections. They affect the energy dissipation of the system, and therefore they influence the assessment of both global and local mechanisms. Pushover curves are widely used in the field of masonry constructions to evaluate the ultimate force/displacement capacity taking into account those nonlinearities. Non-linear time-history analysis is also considered to be a valid tool able to better simulate the structural response under real seismic input. Selection of appropriate input motions, essential for predicting reliable results is, however, a topic requiring clearly further research.
Many code standards still indicate force-based approaches for the assessment of local failure mechanism and only few refer to displacement-based approaches, in many cases more reliable in presence of local URM wall mechanisms. Currently, kinematic analysis proposed in the Italian code is the only one recognized as completely reliable in European Standards, and it can be performed in nonlinear range. Furthermore, the rocking approach considers the evolution of motion during time and dissipation properties at the plastic hinge around which the masonry wall rotates are not neglected. Limitations of such approach rely on the monolithic characteristic of the block, modelled as rigid, not always true for masonry walls or faades. However, improvements of the equation of motion have been done to account for the formation of an intermediate hinge, frequently developed during OOP bending of the walls horizontally restrained on top.
In literature, only few experimental data at connection level are available involving uncertainties on the seismic evaluation of the original or reinforced buildings. Common practise is to perform parametric analyses based on the calibrated model, in order to investigate a large value set of the interesting parameters. This study reveals the necessity of clear and simple ready-to-use research-based procedures grounded on experimental and numerical approaches useful for professionals to assess existing wall-to-horizontal diaphragm connections and design strengthening solutions in historical constructions.
Rock mass properties have been tested using rock samples and physical modeling in the laboratory, and in exploration tunnels and an underground laboratory. Tests on hard rocks under triaxial compressive stresses and true triaxial compressive stresses have been conducted by loading and unloading testing conditions. The size effect of unloaded rock mass has been studied (Li and Wang, 2003). Some true triaxial compressive testing machines have been developed to understand the properties of rock masses at high stresses.
The in situ stresses are measured using hydraulic fracturing and overcoring methods. The techniques for measuring in situ stresses at large overburden depth have been improved. A back analysis method has been developed to understand 3D stress distributions in deep valley regions by considering tectonic history of rock masses with brittle failure features. For example:
Risk factors are identified and discussed within the governing framework for identification, assessment and management of rock engineering risk developed by Hudson and Feng (2015) (see Fig.11). The potential failure risks of rock masses under high stress conditions such as collapse, rockburst, spalling, deep cracking, large deformation, and cracking of shotcrete are identified. The mechanisms for these geo-disasters have been investigated.
Geomechanics mechanism and characteristics of surrounding rock mass deformation failure in the construction phase for the underground powerhouse of the Jinping I hydroelectric station (Huang etal., 2011).
shaking table test of a multi-story subway station under pulse-like ground motions - sciencedirect
Shaking table tests were conducted on a four-story subway structure.Effects of pulse-like motions on seismic responses of underground structures were discussed.Rich low-frequency components of pulse-like motions result in severe responses.Lateral displacements of the side wall manifest as racking deformation.Central columns are vulnerable structures subjected to pulse-like ground motion.
A series of shaking table tests were conducted to investigate the effect of pulse-like ground motion on a multi-story subway station. Dynamic response data, including internal forces, column drift, and settlement and deformation of the soil were obtained and analyzed. Results show that the pulse-like ground motion increases dynamic responses of the subway station and surrounding soils mainly owing to its inherent rich low-frequency component and high energy. In terms of the structure, central columns, especially central columns on a floor with large story height, are vulnerable components of a multi-story subway station. Both the dynamic earth pressure and the deformation mode of the side wall were analyzed.