本期「地工物理模型試驗與應用」專輯共收錄十篇內容豐富的文章。依據試驗時重力場的不同來區分，1 g 物理模型試驗共六篇，其中一篇則在現地土壤進行物理模型試驗；離心模型試驗則有四篇。就論文內容來區分，以邊坡穩定及擋土牆相關研究佔六篇，與目前工程界所關心的熱門議題一致，三篇液化相關研究及一篇樁載重行為的研究論文。由陳榮河及紀伯全（模型邊坡試驗之因次分析）、王國隆及林美聆（1-g條件下之大型邊坡模型受震行為）、黃景川（地工合成物加勁邊坡模型試驗－動態與擬靜態行為之比較）、李崇正等人（以離心模型振動台試驗模擬砂土層液化行為）所撰寫的文章，都對各自所進行的物理模型與原型的相似性，針對不同的考量重點，做相當詳盡的介紹。讀者可經由這些內容，推導自己所擬進行不同種類物理模型試驗的因次律。
The purpose of conducting model tests on slopes is to understand the mechanical behavior of prototype slopes from the test results on small scale physical models with reduced number of variables of the problem. Dimensional analysis plays an important role in the models test, for it assures the feasibility and reliability of test results. In this article, the principle and procedures of dimensional analysis are introduced at first. Then the use of dimensional analysis on important variables as well as different types of model tests subject to seismic forces, reinforcements, and water are discussed accordingly.
To investigate the earth pressure at-rest against retaining structures, an instrumented model retaining wall facility was developed at National Chiao Tung University. To achieve an at-rest condition, a stiff 1.6 m-high model wall and soil bin were made, so that the lateral wall deformation could be minimized. Sixteen soil-pressure transducers were attached to the center zone of the model wall to investigate the distribution of earth pressure at-rest. Air-dry Ottawa sand was used as backfill for the model wall experiments. Based on the experimental data, the following conclusions can be drawn about the build-up of horizontal stress with the filling of a loose cohesionless backfill. The distribution of earth pressure at-rest induced by the 1.5 m-thick loose backfill were approximately linear and in good agreement with Jaky’s theoretical solution. For the pluviated fill, the measured coefficient of earth pressure at-rest Ko would be independent of the thickness of backfill H. It is hoped that the NCTU non-yielding retaining-wall facility would provide a powerful tool to examine and justify the results obtained from theoretical and numerical studies based of many different assumptions.
The seismic slope model test is usually performed by using centrifuge modeling and shaking table test. Centrifuge model uses centrifuge acceleration as principal simulation law for prototype slope. The shaking table test under 1-g gravity force is based on the law of similarity to perform the test of the prototype slope. The shaking table model tests were used in this research.
Shaking table test for sandy slope specimen was performed. The material properties and boundary effects of the model were evaluated and verified before test. Results of shaking table test show that the slope remains in the state of elasticity when input acceleration is smaller than 0.4g, and the amplification effect increases with the input acceleration. However, slope behaves plastically when the input acceleration reaches 0.6g and the slope failure initiates. Dynamic soil properties obtained from the proposed method of Hardin and Drnevich (1972) can properly simulate seismic slope behavior under elastic mode. The effect of modul-us degradation and non-linear behavior should be taken into account when strain increased in slope specimens.
This paper presents briefly shaking table tests on model pile foundations in a liquefiable ground. It discusses the seismic behavior and failure mechanisms of pile foundations in liquefied sand and the recent dynamic physical model tests on pile foundation in liquefiable ground. The test setup, preliminary results and future plan of the physical model tests using a large biaxial laminar shear box (1.880 m×1.880 m×1.520 m) on the shaking table at the National Center for Research on Earthquake Engineering (NCREE) were also presented in this paper. It was found that the behavior of the model piles under shaking was affected by the soil specimen density, the dynamic characteristics of the piles and the surrounding soil, frequency content of earthquake shakings and the mass of the superstructure.
In order to overcome the difficulty for simulation of stress-strain behavior of real soil structures due to stress level limitations, this research use a geotechnical centrifuge modeling technique in studying slope stability. The slope models were prepared from sand, and sand mixed with 15 and 30% fines by weight, compacted at optimum water content. The validity of the modeling technique was confirmed using slope models of different heights, inclinations, and soil types. The soil behavior was studied through triaxial and plane strain tests, and the extended Mohr-Coulomb failure criterion was found relevant for expressing the strength of unsaturated compacted soil based on the angle of internal friction and apparent cohesion. The Bishop’s circular failure mechanism, together with the extended Mohr-Coulomb failure criterion, was able to simulate the slope failure reasonably well. The rainfall of different intensities was then induced on the 60 deg stable slopes of sand with 15% fines. It was found that the failure of slope under rainfall may be interpreted as a reduction in apparent cohesion. The centrifuge tests also allowed the accumulated rainfall threshold and the rainfall intensity-duration threshold curve to be generated for the test slopes, which is a successful application on slope stability investigation.
Four identical geosynthetic-reinforced slopes with a height of Ht = 480 mm, and a slope angle of = 60 were established and loaded to ultimate failure conditions via tilting box and shaking table test facilities. These reinforced slopes simulate idealized 2-D, round particle backfill, by using 150 mm-long, 1.96 mm-diameter stainless steel rods. Extensible non-woven geotextile reinforcement was used in conjunction with a 2 mm-thick aluminum facing plate which simulates a 51 mm-thick concrete facing for a 2 m-high reinforced soil slope. Shaking table tests were conducted under a uni-axial shaking condition, using sinusoidal single-cycle waves with frequencies of f = 3Hz, 6Hz, and 15Hz. For each series of tests with a specific wave frequency, input horizontal peak ground accelerations (HPGA) were step-wise intensified until the ultimate displacement state of the slope. It was found that the acceleration amplitude ratio (Am) at the crest of the slope is a function of HPGA (or Dmax/Ht), and wave frequency (f). They generally increase with the increase of f, while they decrease with the increase of HPGA (or Dmax/Ht). It was also found that a transition from an amplification state (Am > 1) into a de-amplification state (Am < 1) occurs when major sliding planes were extensively mobilized within the backfill. Test results generally support a hypothesis made earlier that the pseudo-static tilting load is a special case of dynamic seismic load with a significantly large period (or small frequency).
Implementations of laboratory testing results to practical projects are restricted due to small specimen size, sample disturbance, and representation of the state of stresses. Empirical correlations that correlated the cyclic stress resistance with in situ testing parameters are unreliable around the boundaries due to the assumptions and data used to develop the correlations. One of the alternative approaches is to develop an in situ soil liquefaction testing technique that measures the coupled shear strain-pore pressure responses in the field. This paper presents the development of an in situ liquefaction model test that cyclically shears the soil from the propagating of shallow stress waves induced from Vibroseis truck. Instrumentation techniques that can evaluate the induced shear strains and process the generated excess pore pressures are developed and deployed in liquefiable soil stratum. Cyclic strain approach that uses the shear strain amplitude as the key variable to describe the soil disturbance and the liquefaction resistance is adopted in data interpretations. The developed technique is used to quantitatively assess the effectiveness of vertical drains as a liquefaction countermeasure. The testing results show that the drainage provided by vertical drains can significantly reduce the generated excess pore pressure and associated settlement.
This study conducted a series of centrifuge model test to investigate bearing behaviors of a pile subjected to compressive and tensile loadings. Three different slender ratios of piles with the same embedded length were designed in the test. During the test, the model pile was first driven into the dry sand at 1g condition, and then an artificial gravity of 80g was applied to the model in the centrifuge to simulate a full-scale prototype situation. From the results, the smaller the slender ratio, the pile skin friction in compressive condition is significantly larger than the one in tensile condition.
Rare full scale or small scale experiments were performed to investigate the failure mechanism of superimposed geosynthetic reinforced earth wall (SGREW). Several empirical design rules were suggested in FHWA-NHI-00-043. The objective of this study is to investigate the effects of offset distance and reinforcement arrangement on the failure behavior of SGREWs. From the test results, several conclusions can be drawn. (1) If the offset distance is smaller than (H1+H2)/6.8, the SGREW should be designed as a single wall with reinforcement length of 0.7(H1+H2). (2) The failure wedges can be included by the plane failure surface with angle of . (3) If the reinforcement layers are between 8 and 16, the lateral earth pressure decreases with increasing reinforcement layers and decreasing equivalent wall inclination.
The principle of geotechnical centrifuge modeling and the scaling laws used in the modeling are introduced. A series of 1-D centrifuge shaking table test was conducted to investigate the seismic responses for the saturated sand deposits without or with a thin layer of silt at different depths. A 1-D shear beam theory was adopted to construct the shear stress-shear strain relation for the sand deposits during shaking. Visualization of the seismic response on the liquefied sand deposit is demonstrated as well.